CN116795628B - Power consumption processing method of terminal device, terminal device and readable storage medium - Google Patents

Abstract

The embodiment of the present application provides a power consumption processing method of a terminal device, a terminal device, and a readable storage medium. The method includes: the terminal device collects power consumption data of the terminal device, and the power consumption data includes power consumption data of a first application; according to the power consumption model of the first application, the user group to which the user corresponding to the terminal device belongs and the threshold of the main feature corresponding to the user group are determined; according to the threshold of the main feature and the value of the main feature in the power consumption data of the first application, whether the power consumption of the first application is abnormal is detected, and the power consumption detection result of the first application is obtained; according to the power consumption detection result of the first application, the corresponding operation is performed. In this way, the terminal device can determine the group to which the user belongs based on the user's habit of using the terminal device, and then use the corresponding threshold of the main feature to detect the power consumption, which is highly flexible.

Description

Power consumption processing method of terminal device, terminal device and readable storage medium

Technical Field

The present application relates to the field of terminal technology, and in particular to a power consumption processing method of a terminal device, a terminal device, and a readable storage medium.

Background technique

With the rapid development of terminal devices, more and more applications are installed on terminal devices. Users can open multiple applications at the same time. The operation of applications will increase the power consumption of terminal devices. Therefore, it is necessary to detect and manage the power consumption of applications to extend the battery life of terminal devices.

Currently, a power consumption threshold can be set, and the terminal device can detect the power consumption of the application running in the background. When the power consumption of the application running in the background is greater than the power consumption threshold, the terminal device can kill the application to reduce the power consumption of the terminal device.

Although the current method for processing the power consumption of terminal devices can reduce the power consumption of terminal devices, different user groups have different usage habits of terminal devices. Using the same method to process the power consumption of terminal devices for all user groups has poor flexibility.

Summary of the invention

The embodiments of the present application provide a method for processing power consumption of a terminal device, a terminal device, and a readable storage medium, which are applied to the field of terminal technology and can adaptively process the power consumption of the terminal device according to the user's usage habits of the terminal device, with high flexibility.

In the first aspect, the embodiment of the present application proposes a method for processing power consumption of a terminal device. The execution subject of the method may be a terminal device or a chip in the terminal device. The following description is made using a terminal device as an example. The method includes: the terminal device may collect power consumption data of the terminal device, and the power consumption data includes power consumption data of a first application. The power consumption data of the first application is the power consumption of each device in the terminal device when the first application is running.

The terminal device may store a power consumption model of the first application, and the terminal device may determine the user group to which the user corresponding to the terminal device belongs and the threshold of the main feature corresponding to the user group based on the power consumption model of the first application. The terminal device may detect whether the power consumption of the first application is abnormal based on the threshold of the main feature and the value of the main feature in the power consumption data of the first application, and obtain the power consumption detection result of the first application. The terminal device may perform corresponding operations based on the power consumption detection result of the first application, such as reducing the power consumption of the first application.

In the embodiment of the present application, different applications may correspond to different power consumption models, and the power consumption model has high accuracy. In addition, in the embodiment of the present application, the user group to which the user belongs can be determined based on the power consumption data, that is, based on the user's habits or characteristics of using the terminal device, the threshold of the main feature corresponding to the user can be used to detect whether the power consumption of the first application is abnormal. In the embodiment of the present application, for different user groups, different thresholds of the main feature can be used to detect whether the power consumption of the first application is abnormal, which has high flexibility and better matches the user's habits or characteristics, and can improve the user experience.

In a possible implementation, the power consumption model of the first application includes: an auxiliary feature condition, and a threshold of a main feature corresponding to at least one user group, wherein the auxiliary feature condition is used to determine the user group to which the user belongs, and at least one user group is a user group determined based on the auxiliary feature condition.

The terminal device may determine the user group to which the user corresponding to the terminal device belongs based on the auxiliary feature condition and the value of the auxiliary feature in the power consumption data of the first application. The terminal device may determine the threshold of the main feature corresponding to the user group to which the user corresponding to the terminal device belongs based on the threshold of the main feature corresponding to the at least one user group and the user group to which the user corresponding to the terminal device belongs.

In this implementation, the terminal device can determine the user group to which the user belongs and the threshold of the main feature corresponding to the user group based on the power consumption model of the first application, so that the threshold of the main feature adapted for the user group can be used to detect whether the power consumption of the first application is abnormal.

In a possible implementation, the power consumption model of the first application is specifically a power consumption model of a first application of a first product series, and the terminal device belongs to the first product series.

In this implementation, the terminal device can determine the user group to which the user corresponding to the terminal device belongs and the threshold of the main feature corresponding to the user group based on the power consumption model of the first application of the first product series, and then detect whether the power consumption of the first application is abnormal based on the threshold of the main feature corresponding to the user group.

In this implementation, because terminal devices of different product series use different devices, the power consumption of the devices is different when the first application runs at the same time. The embodiment of the present application is targeted at different product series and can obtain power consumption models of different applications of different product series, so that the power consumption model is more accurate and more suitable for terminal devices, which can improve the accuracy of power consumption processing.

In a possible implementation, when the value of the main feature is greater than or equal to the threshold of the main feature, the terminal device determines that the power consumption of the first application is abnormal, and the power consumption detection result of the first application is used to indicate that the power consumption of the first application is abnormal. When the value of the main feature is less than the threshold of the main feature, the terminal device determines that the power consumption of the first application is normal, and the power consumption detection result of the first application is used to indicate that the power consumption of the first application is normal.

In a possible implementation, the terminal device can not only determine whether the power consumption of the first application is abnormal, but also determine the abnormality level when the power consumption of the first application is abnormal. Exemplarily, the threshold of the main feature includes: a first threshold, a second threshold, and a third threshold, the first threshold is less than the second threshold, and the second threshold is less than the third threshold.

Wherein, when the value of the main feature is greater than or equal to the first threshold value and less than the second threshold value, the power consumption abnormality of the first application is determined to be the first abnormality level, and the power consumption detection result of the first application is used to indicate that the level of the power consumption abnormality of the first application is the first abnormality level. When the value of the main feature is greater than or equal to the second threshold value and less than the third threshold value, the power consumption abnormality of the first application is determined to be the second abnormality level, and the power consumption detection result of the first application is used to indicate that the level of the power consumption abnormality of the first application is the second abnormality level. When the value of the main feature is greater than or equal to the third threshold value, the power consumption abnormality of the first application is determined to be the third abnormality level, and the power consumption detection result of the first application is used to indicate that the level of the power consumption abnormality of the first application is the third abnormality level. When the value of the main feature is less than the first threshold value, the power consumption of the first application is determined to be normal, and the power consumption detection result of the first application is used to indicate that the power consumption of the first application is normal.

In a possible implementation, the power consumption abnormality level of the first application is different, and the corresponding operations performed by the terminal device are different.

In this implementation, the terminal device can process the abnormal power consumption level of the first application by adopting an operation corresponding to the abnormal level, which has high flexibility.

In a possible implementation, the terminal device may send a power consumption model acquisition request to the cloud to receive the power consumption model of the first application from the cloud.

In some embodiments, the terminal device may receive a power consumption model of at least one application from the cloud, and the at least one application may include the first application. Alternatively, the terminal device may receive a power consumption model of at least one application of at least one series from the cloud, and the at least one application may include the first application.

In a possible implementation, the terminal device may also send power consumption data of the terminal device to the data platform so that the cloud can train a power consumption model based on the power consumption data of the terminal device. Please refer to the description of the second aspect.

In the second aspect, the embodiment of the present application proposes a method for processing power consumption of a terminal device. The execution subject of the method can be the cloud or a chip in the cloud. The following description takes the cloud as an example. The method includes: the cloud can obtain power consumption data of at least one terminal device from the data platform, and train a power consumption model based on the power consumption data of the at least one terminal device. When the cloud receives a request to obtain a power consumption model from a terminal device, the power consumption model can be sent to the terminal device.

The power consumption model may include a power consumption model of at least one application, and the first application is included in the at least one application. Correspondingly, the power consumption data of each terminal device may include the power consumption data of the at least one application. In an embodiment of the present application, training the power consumption model in the cloud may include obtaining the power consumption model of the first application through cloud training. The following takes the power consumption model of the first application trained in the cloud as an example to illustrate the process of training the power consumption model in the cloud. Among them, the cloud can train the power consumption data of the first application based on the power consumption data of the first application within a preset time period.

In a possible implementation, the power consumption model of the first application is specifically a power consumption model of the first application of the first product series, and the power consumption model of the first application of the first product series is trained based on the power consumption data of the first application of the terminal device of the first product series within the preset time period.

The following specifically describes the training process of the power consumption model by taking the cloud as an example, in which the power consumption model of the first application of the first product series is obtained by training according to the power consumption data of the first application of the terminal device of the first product series within a preset time period:

The cloud can obtain the correlation coefficient between the main feature and each other feature based on the power consumption data of the first application of the terminal device of the first product series within the preset time period, and determine the auxiliary feature based on the correlation coefficient between the main feature and each other feature. After determining the auxiliary feature, the cloud can obtain the auxiliary feature condition based on the power consumption data of the first application of the terminal device of the first product series within the preset time period and the auxiliary feature, and the auxiliary feature condition is used to determine the user group to which the user belongs. The cloud can divide the users corresponding to the at least one terminal device into at least two user groups based on the auxiliary feature condition and the power consumption data of the first application of the terminal device of the first product series within the preset time period.

The cloud can determine the threshold of the main feature corresponding to each user group based on the power consumption data of the first application of the terminal device of the first product series within the preset time of each user group, and the threshold of the main feature is used to detect whether the power consumption of the first application is abnormal. The cloud generates a power consumption model of the first application of the first product series based on the auxiliary feature conditions and the threshold of the main feature corresponding to each user group. Among them, the power consumption model of the first application of the first product series may include: the auxiliary feature conditions, and the threshold of the main feature corresponding to at least one user group.

In a possible implementation, the correlation coefficient between the main feature and each of the other features includes a Pearson correlation coefficient and a Spearman correlation coefficient. Determining the auxiliary feature according to the correlation coefficient between the main feature and each of the other features includes: taking a feature greater than or equal to a Pearson correlation coefficient threshold and greater than or equal to a Spearman correlation coefficient threshold as a first candidate auxiliary feature; and determining the auxiliary feature according to the first candidate auxiliary feature.

In a possible implementation manner, the first candidate auxiliary feature may be used as the auxiliary feature.

In a possible implementation manner, determining the auxiliary feature according to the first candidate auxiliary feature includes:

Step A, obtaining the variance expansion factor VIF of each first candidate auxiliary feature and other first candidate auxiliary features;

Step B, for the first candidate auxiliary feature corresponding to the maximum VIF, if the maximum VIF is greater than the VIF threshold and the number of the first candidate auxiliary features is greater than the number threshold, executing step C; if the maximum VIF is greater than the VIF threshold and the number of the first candidate auxiliary features is less than or equal to the number threshold, or the maximum VIF is less than or equal to the VIF threshold, all the first candidate auxiliary features are used as second candidate auxiliary features, and the auxiliary feature is determined according to the second candidate auxiliary features;

Step C, delete the first candidate auxiliary feature corresponding to the maximum VIF, and return to execute steps A-B until the maximum VIF is less than or equal to the VIF threshold, or the number of remaining first candidate auxiliary features is less than or equal to the number threshold, and use the remaining first candidate auxiliary features as second candidate auxiliary features, and determine the auxiliary feature based on the second candidate auxiliary feature.

In a possible implementation manner, the second candidate auxiliary feature may be used as the auxiliary feature.

In a possible implementation, determining the auxiliary feature based on the second candidate auxiliary feature includes: obtaining the importance value of each second candidate auxiliary feature based on a minimum absolute shrinkage and selection model; and taking the second candidate auxiliary features ranked in the top N as the auxiliary feature, where N is equal to the quantity threshold.

In a possible implementation, the obtaining of auxiliary feature conditions based on the power consumption data of the first application of the terminal device of the first product series within the preset time period and the auxiliary feature includes: training a classification and regression tree CART based on the auxiliary feature and the power consumption data of the first application of the terminal device of the first product series within the preset time period; and using the leaf nodes in the CART as the auxiliary feature conditions.

In a possible implementation, the method of determining the threshold of the main feature corresponding to each user group based on the power consumption data of the first application of the terminal device of the first product series within the preset time of each user group includes: determining the threshold of the main feature corresponding to each user group based on the value of the main feature and a first preset ratio in the power consumption data of the first application of the terminal device of the first product series within the preset time of each user group.

In a possible implementation, the threshold of the main feature corresponding to each user group determined based on the first preset ratio is the first threshold. The method further includes: determining the second threshold of the main feature corresponding to each user group according to the value of the main feature and the second preset ratio in the power consumption data of the first application of the terminal device of the first product series within the preset time of each user group; determining the third threshold of the main feature corresponding to each user group according to the value of the main feature and the third preset ratio in the power consumption data of the first application of the terminal device of the first product series within the preset time of each user group, wherein the first preset ratio is less than the second preset ratio, the second preset ratio is less than the third preset ratio, and the first threshold, the second threshold, and the third threshold are used to determine the level of power consumption anomaly of the first application.

In a third aspect, an embodiment of the present application provides a terminal device, which may also be referred to as a terminal, user equipment (UE), mobile station (MS), mobile terminal (MT), etc. The terminal device may be a mobile phone, a smart TV, a wearable device, a tablet computer (Pad), a computer with wireless transceiver function, a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless terminal in industrial control, a wireless terminal in self-driving, a wireless terminal in remote medical surgery, a wireless terminal in smart grid, a wireless terminal in transportation safety, a wireless terminal in smart city, a wireless terminal in smart home, etc.

The terminal device comprises: a processor and a memory; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory, so that the terminal device executes the method of the first aspect.

In a fourth aspect, an embodiment of the present application provides an electronic device, which may be the cloud of the second aspect. The electronic device comprises: a processor and a memory; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory, so that the electronic device performs the method of the second aspect.

In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the method of the first aspect and the second aspect is implemented.

In a sixth aspect, an embodiment of the present application provides a computer program product, which includes a computer program. When the computer program is run, the computer executes the method of the first aspect or the second aspect.

In a seventh aspect, an embodiment of the present application provides a chip, the chip including a processor, the processor being used to call a computer program in a memory to execute the method described in the first aspect and the second aspect.

It should be understood that the third to seventh aspects of the present application correspond to the technical solutions of the first aspect of the present application, and the beneficial effects achieved by each aspect and the corresponding feasible implementation methods are similar and will not be repeated here.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a schematic diagram of a system architecture applicable to a method for processing power consumption of a terminal device provided in an embodiment of the present application;

FIG2A is another schematic diagram of a system architecture applicable to the power consumption processing method of a terminal device provided in an embodiment of the present application;

FIG2B is a schematic diagram of a flow chart of a method for processing power consumption of a terminal device provided in an embodiment of the present application;

FIG3 is a schematic diagram of a process for obtaining a power consumption model according to an embodiment of the present application;

FIG4 is a schematic diagram of a CART provided in an embodiment of the present application;

FIG5A is a schematic diagram of a screening auxiliary feature provided in an embodiment of the present application;

FIG5B is a schematic diagram of determining a threshold value of a main feature according to an embodiment of the present application;

FIG6 is a flow chart of another embodiment of a method for processing power consumption of a terminal device provided in an embodiment of the present application;

FIG7 is a flow chart of another embodiment of a method for processing power consumption of a terminal device provided in an embodiment of the present application;

FIG8 is a flow chart of another embodiment of a method for processing power consumption of a terminal device provided in an embodiment of the present application;

FIG. 9 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

Detailed ways

In order to facilitate the clear description of the technical solutions of the embodiments of the present application, in the embodiments of the present application, the words "exemplary" or "for example" or "such as" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "exemplary" or "for example" in the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

In the embodiment of the present application, "at least one" refers to one or more, and "multiple" refers to two or more. "And/or" describes the association relationship of the associated objects, indicating that there may be three relationships, for example, A and/or B, which can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The character "/" generally indicates that the associated objects before and after are in an "or" relationship. "At least one of the following items (individuals)" or similar expressions thereof refer to any combination of these items, including any combination of single items (individuals) or plural items (individuals). For example, at least one of a, b, or c can represent: a, b, c, a-b, a-c, b-c, or a-b-c, where a, b, c can be single or multiple. In addition, it should be understood that in the description of the present application, words such as "first" and "second" are only used to distinguish the purpose of description, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying order.

It should be noted that the "at..." or "when..." in the embodiments of the present application can be the instant when a certain situation occurs, or can be a period of time after a certain situation occurs, and the embodiments of the present application do not specifically limit this. In addition, the display interface provided in the embodiments of the present application is only an example, and the display interface can also include more or less content.

With the rapid development of terminal devices, more and more applications are installed on terminal devices. Users can open multiple applications at the same time. The operation of applications will increase the power consumption of terminal devices. Therefore, it is necessary to detect and manage the power consumption of applications to extend the battery life of terminal devices.

In some embodiments, compared with the applications running in the foreground of the terminal device, the user pays less attention to the applications running in the background. Therefore, a power consumption threshold can be set at present, such as the power consumption threshold can be the threshold of the background power consumption of the central processing unit (CPU) of the application, and the threshold of the CPU background power consumption can be understood as the power consumption generated by the application using the CPU when running in the background. The terminal device can detect the CPU power consumption of the application running in the background. When the CPU power consumption of the application running in the background is greater than the power consumption threshold, the terminal device can kill the application to reduce the power consumption of the terminal device. Alternatively, the terminal device will output a prompt message to prompt the user that the power consumption of the background application is high. When the CPU power consumption of the application running in the background is less than or equal to the power consumption threshold, the application can continue to run in the background, such as the terminal device can do no processing.

Although the current method for processing the power consumption of terminal devices can reduce the power consumption of terminal devices, different user groups have different usage habits of terminal devices. Using the same method (such as using the same power consumption threshold) to process the power consumption of terminal devices for all user groups has poor flexibility.

For example, user A likes to listen to music and listens to music for a long time every day, while user B does not like to listen to music and hardly listens to music every day. However, when both user A and user B open an audio playback application and return the audio playback application to the background, the audio playback application can play music. For user A and user B, if the terminal device follows the same processing logic, when the CPU background power consumption of the audio playback application is greater than the power consumption threshold, the terminal device will output a prompt message to prompt the user that the application has high power consumption. In this case, user A who likes music will receive a prompt message as soon as the audio playback application returns to the background, and may even receive prompt messages continuously, which disturbs the user too much and causes a poor user experience.

Based on the current problems, an embodiment of the present application provides a power consumption processing method for a terminal device, which can determine the user group to which a user belongs based on the user's habits or characteristics of using applications. Different user groups can correspond to different power consumption processing logics, so that for different users, the terminal device can adaptively adjust the power consumption processing method. It has high flexibility, and the power consumption processing logic is adapted to the user's habits or characteristics, which can improve the user experience.

Exemplarily, taking audio playback applications as an example, and taking CPU background power consumption as a detection condition as an example, user A likes to listen to music and listens to music for a long time every day, while user B does not like to listen to music and hardly listens to music every day. Among them, the power consumption threshold corresponding to user A can be a first power consumption threshold, and the power consumption threshold corresponding to user B can be a second power consumption threshold, and the first power consumption threshold is greater than the second power consumption threshold. For example, when both user A and user B open audio playback applications and return the audio playback applications to the background, user A's terminal device can compare the CPU background power consumption of the audio playback application with the first power consumption threshold. For example, when the CPU background power consumption of the audio playback application is greater than the first power consumption threshold, user A's terminal device can output a prompt message. User B's terminal device can compare the CPU background power consumption of the audio playback application with the second power consumption threshold. For example, when the CPU background power consumption of the audio playback application is greater than the second power consumption threshold, user B's terminal device can output a prompt message.

In this way, for user A who likes music, after the audio playback application is retired to the background, because the first power consumption threshold is large, user A will not receive prompt information frequently. For user B who does not like music, after the audio playback application is retired to the background, because the second power consumption threshold is small, user B's terminal device will promptly remind user B that the audio playback application is large, so that user B can timely understand the power consumption of the application in the terminal device.

Before introducing the power consumption processing method of the terminal device provided in the embodiment of the present application, the system architecture applicable to the power consumption processing method of the terminal device provided in the embodiment of the present application is first introduced. Figure 1 is a schematic diagram of the system architecture applicable to the power consumption processing method of the terminal device provided in the embodiment of the present application. Referring to Figure 1, the system architecture may include: at least one terminal device, a cloud, and a data platform. In Figure 1, the cloud and the data platform are represented by a server, and the embodiment of the present application does not limit the form of the cloud and the data platform. It should be understood that Figure 1 takes one terminal device as an example.

The terminal device may report the power consumption data of the terminal device to the data platform. In some embodiments, the terminal device may periodically report the power consumption data of the terminal device to the data platform, such as the terminal device may report the power consumption data of the terminal device to the data platform once every three days. In some embodiments, the terminal device may periodically report the power consumption data of the terminal device to the data platform, such as the terminal device may report the power consumption data of the terminal device to the data platform at 6 o'clock every day.

The power consumption data of the terminal device may include: the power consumption data of each application of the terminal device. When each application is running, different devices in the terminal device will be used. The devices used when the application is running can be called devices corresponding to the application. The power consumption data of each application may include: the power consumption data of the device corresponding to each application.

The device may include, but is not limited to: a CPU, a graphics processing unit (GPU), a global navigation satellite system (GNSS), a screen, a sensor, a camera, a flash, an audio module, a Bluetooth module, a modem, and a Wi-Fi module. For example, the Bluetooth module may be a Bluetooth chip, and the Wi-Fi module may be a Wi-Fi chip, which is not shown in the embodiment of the present application. Among them, the terminal device can realize Bluetooth communication through the Bluetooth module, and the terminal device can realize Wi-Fi communication through the Wi-Fi module.

The sensors in the terminal device may include but are not limited to: pressure sensor, gyroscope sensor, air pressure sensor, magnetic sensor, acceleration sensor, distance sensor, proximity light sensor, fingerprint sensor, temperature sensor, touch sensor, ambient light sensor, bone conduction sensor.

The audio module in the terminal device may include but is not limited to: a speaker, a receiver, a microphone, etc. Exemplarily, the terminal device can play music, collect sound, etc. through the audio module.

In the embodiments of the present application, there is no restriction on the devices included in the terminal device, and there is no restriction on the division method of the devices in the terminal device. For example, in some embodiments, the Bluetooth module, modem, and Wi-Fi module can be divided into one module, such as a communication module. In some embodiments, the devices can also be divided into finer-grained or coarser-grained devices, and the embodiments of the present application do not limit this.

The terminal device may periodically collect statistics on the power consumption data of each application of the terminal device, and periodically report the power consumption data of the terminal device to the data platform.

For example, taking the power consumption data of an application counted once a day as an example, the power consumption data of each application in a day may include at least one of the following items: foreground usage time, foreground power consumption, background usage time, background power consumption, total foreground traffic, and total background traffic of the device corresponding to the application.

It should be understood that the power consumption data corresponding to different devices may be different. Exemplarily, such as a communication module, if the terminal device communicates through a modem and a Wi-Fi module, the traffic of the terminal device will be used, so the power consumption data of the modem and the Wi-Fi module may include: total foreground traffic, and total background traffic. Exemplarily, if the status of the screen includes screen on and screen off, the screen will not run in the background, so the power consumption data of the screen may include: screen power consumption when the screen is on, screen usage time, screen foreground power consumption, and screen foreground usage time. Exemplarily, whether the application is running in the foreground or the background, it will use the resources of the CPU, so the power consumption data of the CPU may include: CPU foreground power consumption, CPU foreground usage time, CPU background power consumption, and CPU background usage time.

For example, the terminal device reports the power consumption data of the terminal device to the data platform once a week. The terminal device may report the power consumption data of the terminal device to the data platform once every week, and the power consumption data may include power consumption data of the terminal device for 7 days.

Table 1 is an example diagram of the content included in the power consumption data of the terminal device and the meaning of the content. It should be understood that the data power consumption of the terminal device can be reported to the data platform in text form, table form, etc., and the embodiment of the present application does not limit the data form of the power consumption data.

Table I

In Table 1, FG represents the foreground, BG represents the background, and trfc represents the traffic.

In some embodiments, because the power consumption data of the terminal device may include the power consumption data of each application of the terminal device, in order to facilitate the power consumption data of different applications of the terminal device, the power consumption data of the terminal device may also include application information. Exemplarily, the application information may include the application name and the application version. In some embodiments, the application name may also be replaced with information that can distinguish different applications, such as the application number and the application icon.

In some embodiments, the power consumption data of the terminal device may also include: device information. For example, the device information may include the name of the terminal device (product name), the terminal model (product version), and the serial number (serial number, SN). In some embodiments, the product name may also be replaced with information such as the product number that can distinguish different products.

In some embodiments, the power consumption data of the terminal device may further include: a timestamp. The timestamp indicates the time when the terminal device reports the power consumption data of the terminal device to the data platform.

It should be noted that Table 1 also classifies the power consumption data of the terminal device and shows the characteristics corresponding to the power consumption data. For example, device information can represent product-related characteristics, application information can represent application characteristics, and screen power consumption, screen usage time, screen foreground power consumption, and screen foreground usage time can represent screen-related characteristics. Other relevant characteristics of the terminal device can be shown in Table 1.

It should be understood that, because at least one terminal device can report the power consumption data of the terminal device to the data platform, the data platform can store the power consumption data of the at least one terminal device.

For example, Table 2 shows the power consumption data of Application 1 of multiple terminal devices received by the data platform in one day. It should be understood that Table 2 takes the power consumption data of Application 1 including the foreground power consumption of the device corresponding to Application 1 as an example.

Table II

Exemplarily, the unit of power consumption involved in the embodiments of the present application may be milliampere-hour (mAh).

It should be understood that Table 1 shows terminal devices with different serial numbers of product 1, and shows the foreground power consumption of the device corresponding to application 1 of the terminal device on the same day (2023/5/15). It should be understood that the product names of different terminal devices may be the same or different. Exemplarily, when the terminal devices belong to the same product series, the product names of the terminal devices are the same, and when the terminal devices do not belong to the same product series, the product names of the terminal devices are different. Different terminal devices have different SNs, and different terminal devices can be distinguished by SN.

The cloud can obtain the power consumption data of the terminal device from the data platform, and train the power consumption model based on the power consumption data of the terminal device. In some embodiments, the cloud can periodically obtain the power consumption data of the terminal device from the data platform, or the cloud can periodically obtain the power consumption data of the terminal device from the data platform. The power consumption model can be used to determine the user group to which the user corresponding to the terminal device belongs, and the power consumption processing logic corresponding to the user group.

It should be noted that the cloud can periodically update the power consumption model. For example, taking one day as an example, the cloud can update the power consumption model every other day based on the power consumption data of the terminal device obtained from the data platform the previous day to ensure the accuracy and real-time performance of the power consumption model.

After the power consumption model is obtained through cloud training or updated on the cloud, the power consumption model can be sent to the terminal device.

The terminal device can store the power consumption model. During operation, the terminal device can collect the power consumption data of the terminal device in real time, and based on the power consumption model, obtain the user group to which the user corresponding to the terminal device belongs, and the power consumption processing logic corresponding to the user group. The terminal device can process the power consumption of the terminal device according to the power consumption processing logic corresponding to the user group. The processing logic of the terminal device can refer to the relevant descriptions in Figures 6 and 7.

In some embodiments, referring to FIG. 2A , the terminal device may include a computing engine and a power consumption processing module.

The power consumption processing module is used to collect the power consumption data of the terminal device in real time and send the power consumption data of the terminal device to the calculation engine.

The calculation engine is used to store the power consumption model and input the power consumption data of the terminal device from the power consumption processing module into the power consumption model to obtain the user group to which the user corresponding to the terminal device belongs and the power consumption processing logic corresponding to the user group. The calculation engine can process the power consumption of the terminal device according to the power consumption processing logic corresponding to the user group.

In some embodiments, the computing engine can detect whether the power consumption of the terminal device is abnormal according to the power consumption processing logic corresponding to the user group, and specifically detect whether the power consumption of the application of the terminal device is abnormal, and send the power consumption detection result to the power consumption processing module. The power consumption detection result may include but is not limited to: normal power consumption, abnormal power consumption.

Accordingly, the power consumption processing module, in response to the power consumption detection result from the computing engine, can perform corresponding operations based on the power consumption detection result to control the power consumption of the application of the terminal device.

In some embodiments, the computing engine may include a power consumption service module. The power consumption service module may include: a power consumption interface, a power consumption model calling unit, and a power consumption model storage unit.

Among them, the power consumption interface is used to transmit the power consumption data of the terminal device from the power consumption processing module, as well as the power consumption detection result.

The power consumption model storage unit is used to store the power consumption model. The power consumption model calling unit is used to call the power consumption model in the power consumption model storage unit to detect whether the power consumption of the terminal device is abnormal.

In some embodiments, the power consumption processing module may include: a power consumption processing process and a power consumption response unit. In some embodiments, the power consumption processing process and the power consumption response unit may be set separately, such as a terminal device may include a computing engine, a power consumption processing process, and a power consumption response unit. In some embodiments, the power consumption processing process and the power consumption response unit may be integrated into one, and FIG2A takes the power consumption processing process and the power consumption response unit as an example of being included in the power consumption processing module.

The power consumption processing process is used to collect the power consumption data of the terminal device and report the power consumption data of the terminal device to the data platform. The power consumption processing process is also used to collect the power consumption data of the terminal device in real time, send the power consumption data of the terminal device to the computing engine through the power consumption interface, and receive the power consumption detection results from the computing engine.

The power consumption processing process can instruct the power consumption response unit to perform corresponding operations based on the power consumption detection result to control the power consumption of the application.

FIG2B is a flow chart of a method for processing power consumption of a terminal device provided in an embodiment of the present application. FIG2B takes a terminal device as an example to illustrate the interaction process between the terminal device and the cloud and the data platform. Referring to FIG2B , the method for processing power consumption of a terminal device provided in an embodiment of the present application may include:

S201, the power consumption processing process collects power consumption data of the terminal device.

Exemplarily, the power consumption processing process may collect power consumption data of the terminal device every 10 minutes.

S202: The power consumption processing process reports the power consumption data of the terminal device to the data platform.

For example, taking the power consumption data of the terminal device reported once a day as an example, after the power consumption data of the terminal device is collected every 10 minutes, the power consumption data of the terminal device collected in the "day" can be counted every day, and the power consumption data of the terminal device can be reported to the data platform once a day.

S203: The data platform stores power consumption data of at least one terminal device.

S204, the cloud obtains power consumption data of at least one terminal device from the data platform.

S205: The cloud trains a power consumption model based on power consumption data of at least one terminal device.

In some embodiments, the power consumption data of different applications are different, and each application may correspond to a power consumption model. For a power consumption model of an application, the cloud can train the power consumption model of the application based on the power consumption data of the application of at least one terminal device. In this embodiment, the cloud can train the power consumption model corresponding to each application.

In this embodiment, the power consumption model obtained through cloud training may include: an identification of the application and a power consumption model of the application.

In some embodiments, terminal devices of different product series have different models of devices arranged therein, so when the same application uses the same device at the same time, the power consumption of the device is different. For example, if the terminal devices of product series 1 use Bluetooth chip 1, and the terminal devices of product series 2 use Bluetooth chip 2, at the same time, the same application in the terminal devices of product series 1 and the terminal devices of product series 2 use their respective Bluetooth chips for communication, and the power consumption of Bluetooth chip 1 and Bluetooth chip 2 is different. In this embodiment, for different product series, the cloud can train the power consumption model of the terminal devices of each product series based on the power consumption data of the terminal devices of different product series.

It should be understood that in the same product series, the cloud can train the power consumption model of each application in the product series based on the power consumption data of different applications in the product series. In this embodiment, the power consumption model trained by the cloud may include: the identification of the product series, the identification of the application, and the power consumption model of the application under the product series.

The specific training process of S205 may refer to the description in FIG. 3 .

S206, the computing engine obtains a power consumption model from the cloud.

In some embodiments, the computing engine may periodically or regularly obtain the power consumption model from the cloud. For example, the computing engine may send a power consumption model acquisition request to the cloud once a day, and the power consumption model acquisition request is used to request the latest power consumption model.

When the terminal device is powered on for the first time, the computing engine does not store the power consumption model. The computing engine may send a request to obtain the power consumption model to the cloud, and the cloud may send the latest power consumption model in the cloud to the computing engine in response to the request.

Because the cloud can update the power consumption model, in the embodiment of the present application, in order to facilitate the distinction between different versions of the power consumption model, the power consumption model can be version-identified or numbered. Taking the version identification as an example, when the terminal device is turned on for the first time, because the power consumption model is not stored in the computing engine, the power consumption model acquisition request sent by the computing engine may not include the version identification of the power consumption model, or the power consumption model acquisition request may include an identification indicating that the computing engine has not yet requested the power consumption model.

In some embodiments, when the computing engine sends a request to obtain a power consumption model to the cloud, the request may include a version identifier of the power consumption model stored in the computing engine. In this way, the cloud can determine the version of the power consumption model stored in the computing engine in response to the request to obtain a power consumption model from the computing engine. When the cloud stores the latest power consumption model, the cloud can send the latest power consumption model to the computing engine. When the version of the power consumption model stored in the cloud is the same as the version of the power consumption model of the computing engine, that is, the cloud has not updated the power consumption model, the cloud can feedback a response message of "power consumption model not updated" to the computing engine.

S207, the power consumption processing process sends the power consumption data of the terminal device to the calculation engine through the power consumption interface.

It should be understood that S207 and S202 can be executed simultaneously without distinction of order.

Exemplarily, the power consumption processing process may collect the power consumption data of the terminal device once every 10 minutes, and the power consumption processing process may send the power consumption data of the terminal device to the calculation engine once every 10 minutes through the power consumption interface.

S208, the calculation engine calls the power consumption model, and inputs the power consumption data of the terminal device into the power consumption model to obtain the power consumption detection result.

S209: The computing engine sends the power consumption detection result to the power consumption processing process.

S210, the power consumption processing process instructs the power consumption response unit to perform corresponding operations according to the power consumption detection result.

S208 - S210 may refer to the relevant descriptions in FIG. 6 and FIG. 7 .

FIG3 takes an application of a terminal device of a product series as an example to illustrate the process of obtaining a power consumption model in the cloud. It should be understood that the process of obtaining power consumption models of other product series and other applications can refer to the description in FIG3 .

FIG3 is a schematic diagram of a process for obtaining a power consumption model provided in an embodiment of the present application. Referring to FIG3 , a method for processing power consumption of a terminal device provided in an embodiment of the present application may include:

S301, the cloud determines a target application, where the target application is an application for which a power consumption model is to be trained.

In some embodiments, S301 is an optional step. In some embodiments, for each application, the cloud can train a power consumption model for each application. The target application can include each application of the terminal device.

In some embodiments, the target application may be a preset application. In some embodiments, the target application may be a popular application. In some embodiments, the target application may be an application customized by a developer.

In some embodiments, the cloud can determine the target application based on the power consumption data of at least one terminal device and the amount of power consumption data of the application. For example, if the terminal device collects power consumption data of an application once a day, referring to Table 2, a row of data in Table 2 can be used as one piece of power consumption data of application 1. The more users use the application, the more terminal devices report the power consumption data of the application, and the more the amount of power consumption data of the application.

Exemplarily, the cloud can sort the data volume of the power consumption data of all applications in order from most to least, and the cloud can select the top 150 applications as target applications. It should be understood that 150 is an example and the value can be customized. In some embodiments, the top 150 applications can be referred to as top150 applications.

In some embodiments, because some applications of the terminal device are resident applications, the failure of the resident application to run may cause a malfunction (bug) of the terminal device. Exemplarily, the resident application may be a system application, which may include but is not limited to: a desktop application, a lock screen application, and the like.

In an embodiment of the present application, the cloud can sort the data volume of the power consumption data of all applications in order from most to least, delete the resident applications, and then select the top 150 applications as target applications.

In some embodiments, the amount of power consumption data is different for different product lines and different applications. For example, for applications such as social applications and entertainment applications with a large number of users, the amount of power consumption data of the applications in one day can reach millions, while for some office applications with a small number of users, the amount of power consumption data of the applications in one day is relatively small.

In the embodiment of the present application, in order to avoid the problem of excessive amount of power consumption data of the first application of the terminal device of the first product series within a preset time period, resulting in a waste of training resources, and to avoid the problem of low accuracy of the power consumption model due to too little data, a first data amount threshold and a second data amount threshold can be preset, and the first data amount threshold is greater than the second data amount threshold. Exemplarily, the first amount threshold can be 1 million, and the second data amount threshold can be 100,000.

Among them, when the amount of power consumption data of the first application of the terminal device of the first product series within the preset time period is greater than 1 million (the first data amount threshold), the cloud can randomly extract 1 million power consumption data from the amount of power consumption data of the first application of the terminal device of the first product series within the preset time period as training data. Among them, when the amount of power consumption data of the first application of the terminal device of the first product series within the preset time period is less than or equal to 1 million (the first data amount threshold) and greater than or equal to 100,000 (the second data amount threshold), the cloud can use the power consumption data of the first application of the terminal device of the first product series within the preset time period as training data. When the amount of power consumption data of the first application of the terminal device of the first product series within the preset time period is less than 100,000 (the second data amount threshold), the cloud may not train the power consumption model, because the small amount of data will result in low accuracy of the power consumption model.

S302, the cloud pre-processes the power consumption data of the first application of the terminal device of the first product series within a preset time period.

In some embodiments, S302 is an optional step. In some embodiments, the cloud may not pre-process the power consumption data of the first application of the terminal device of the first product series within the preset time period, but directly train according to the power consumption data of the first application of the terminal device of the first product series within the preset time period to obtain the power consumption model.

It should be understood that the first product series can represent each product series, and the first application can represent each application in the target application. In the embodiment of the present application, the cloud can divide all power consumption data according to the product series and the application to obtain the power consumption data of the first application of the terminal device of the first product series. It is understandable that the same product series can include terminal devices of multiple models.

In some embodiments, the preset time period may be one day, one week, or one month, etc., which is not limited in the embodiments of the present application. Exemplarily, taking the preset time period of 30 days as an example, the cloud can obtain the power consumption data of the first application of the terminal device of the first product series in the last 30 days. In other words, the cloud can obtain the power consumption data of each application of the terminal device of each product series in the last 30 days, and obtain the power consumption model of each application of the terminal device of each product series based on the power consumption data of each application of the terminal device of each product series, and refer to the relevant description of "the first application of the terminal device of the first product series".

The preprocessing process is described below:

First, in some embodiments, due to network instability and other reasons, some characteristic data may be missing in the power consumption data reported by the terminal device. In the embodiment of the present application, the cloud can delete the power consumption data of the missing characteristic data. For example, referring to Table 2, if there is a power consumption data missing the screen foreground power consumption data, the cloud can delete the power consumption data.

In some embodiments, the power consumption data reported by the terminal device to the data platform may also store duplicates. For example, if there are multiple power consumption data for an application on the terminal device in one day, the cloud can perform deduplication processing and retain one power consumption data for an application on the terminal device. For example, the cloud can retain the power consumption data with the highest CPU background power consumption, which is not limited in the embodiments of the present application.

Secondly, because different applications use different devices when running, the power consumption values of the devices have different orders of magnitude. For example, as shown in Table 2, in the terminal devices of product series 1 and SN1, the foreground power consumption of the Wi-Fi module of application 1 is only 0.29, but in the terminal devices of product series 1 and SN2, the foreground power consumption of the screen of application 1 is 780.33, which is 7000 times that of the former. The order of magnitude difference of power consumption values is large. If the cloud directly uses power consumption data with large orders of magnitude differences, the features with larger values will appear more prominent and important in the power consumption model, while the features with smaller values will appear insignificant in the power consumption model, resulting in low accuracy of the power consumption model.

Therefore, in order to unify the comparison standards and ensure the accuracy of the power consumption model, the cloud can standardize the power consumption data so that the power consumption values of each device in the first application of the terminal device of the first product series within the preset time period are at the same order of magnitude to eliminate the impact of different features due to different orders of magnitude. Exemplarily, the cloud can perform Z-Score standardization on the power consumption data of the first application of the terminal device of the first product series within the preset time period. In some embodiments, the cloud can also use methods such as extreme value standardization or linear ratio standardization to standardize the power consumption data.

In an embodiment of the present application, after the cloud performs Z-Score normalization processing on the power consumption data of the first application of the terminal device of the first product series within a preset time period, the mean value between the power consumption data can be 0 and the standard deviation can be 1.

S303, the cloud obtains the correlation coefficient between each feature and the main feature in the power consumption data based on the power consumption data of the first application of the terminal device of the first product series within a preset time period.

It should be noted that the power consumption data used in S303 is pre-processed power consumption data.

The power consumption data of the first application of the terminal device of the first product series within the preset time period may include data of at least one feature. Referring to Table 1, the at least one feature may include but is not limited to: screen-related features, CPU-related features, GNSS-related features, sensor-related features, GPU-related features, camera-related features, flash-related features, audio-related features, Bluetooth-related features, modem-related features, and Wi-Fi-related features. Application features, time features, and product-related features do not participate in the calculation of the correlation coefficient with the main feature in S303.

The main feature is used to detect whether the power consumption of the first application is abnormal. In some embodiments, the main feature may be pre-set.

Exemplarily, the main feature may be the CPU background power consumption. Exemplarily, taking the CPU background power consumption as the main feature, for each power consumption data, the cloud can obtain the correlation coefficient between each feature and the CPU background power consumption in each power consumption data. Referring to Table 1, for the terminal devices of product 1 (first product series) and SN1, and the power consumption data of application 1 on 2023/5/15, the cloud can obtain the correlation coefficient between the screen power consumption and the CPU background power consumption, the correlation coefficient between the screen usage time and the CPU background power consumption, the correlation coefficient between the screen foreground power consumption and the CPU background power consumption, ..., and the correlation coefficient between the total background traffic of the Wi-Fi module and the CPU background power consumption.

In some embodiments, for each power consumption data, the cloud can obtain the Pearson correlation coefficient (Pearson correlation coefficient) between each feature and the main feature. The cloud can use the Pearson correlation coefficient as the correlation coefficient between each feature and the main feature. Among them, the Pearson correlation coefficient is used to describe the linear relationship between two features, and the value of the Pearson correlation coefficient can be between [-1, 1].

The stronger the linear relationship between two features, the closer the Pearson correlation coefficient will be to -1 or 1. The weaker the linear relationship between two features, the closer the Pearson correlation coefficient will be to 0. The cloud can use the following formulas 1 to 4 to calculate the Pearson correlation coefficient between each feature and the main feature:

Wherein, Y may represent a main feature, such as CPU background power consumption. X may represent each feature in the power consumption data, and each feature may include or exclude the main feature. Formula 1 is used to calculate the standard deviation σ _Y of the main feature in the power consumption data of the first application of the terminal device of the first product series within a preset time period, and Formula 2 is used to calculate the standard deviation σ _X of each feature in the power consumption data of the first application of the terminal device of the first product series within a preset time period.

Among them, i represents any power consumption data in the power consumption data of the first application of the terminal device of the first product series within the preset time period, and n represents the data volume of the power consumption data. _Yi represents the value of the main feature in the any power consumption data, and _μY represents the mean value of the main feature in the power consumption data of the first application of the terminal device of the first product series within the preset time period. Similarly, taking the example that each feature includes the first feature, _Xi represents the value of the first feature in the any power consumption data, and _μX represents the mean value of the first feature in the power consumption data of the first application of the terminal device of the first product series within the preset time period.

Among them, Formula 3 is used to calculate the covariance cov(X,Y) between each feature and the main feature.

Wherein, Formula 4 is used to calculate the Pearson correlation coefficient ρ(X,Y) between each feature and the main feature.

In some embodiments, for each power consumption data, the cloud can obtain the Spearman correlation coefficient between each feature and the main feature, and use the Spearman correlation coefficient as the correlation coefficient between each feature and the main feature. The Spearman correlation coefficient can represent the monotonic relationship between two features, and the value of the Spearman correlation coefficient is also between [-1, 1].

In some embodiments, for each power consumption data, the cloud can obtain the Pearson correlation coefficient between each feature and the main feature, and the Spearman correlation coefficient between each feature and the main feature. The correlation coefficient between each feature and the main feature may include: the Pearson correlation coefficient between each feature and the main feature, and the Spearman correlation coefficient between each feature and the main feature.

S304, the cloud determines the auxiliary features according to the correlation coefficient between each feature and the main feature.

In some embodiments, a Pearson correlation coefficient threshold and a Spearman correlation coefficient threshold may be preset. The Pearson correlation coefficient threshold and the Spearman correlation coefficient threshold may be equal or unequal. For example, the Pearson correlation coefficient threshold may be 0.3, and the Spearman correlation coefficient threshold may be 0.3.

Among them, when the cloud obtains the Pearson correlation coefficient between each feature and the main feature, the cloud can use features greater than or equal to the Pearson correlation coefficient threshold as auxiliary features. When the cloud obtains the Spearman correlation coefficient between each feature and the main feature, the cloud can use features greater than or equal to the Spearman correlation coefficient threshold as auxiliary features. When the cloud obtains the Pearson correlation coefficient between each feature and the main feature, and the Spearman correlation coefficient between each feature and the main feature, the cloud can use features greater than or equal to the Pearson correlation coefficient threshold and greater than or equal to the Spearman correlation coefficient threshold as auxiliary features.

In some embodiments, taking the cloud obtaining the Pearson correlation coefficient between each feature and the main feature, and the Spearman correlation coefficient between each feature and the main feature as an example, the cloud can use features that are greater than or equal to the Pearson correlation coefficient threshold and greater than or equal to the Spearman correlation coefficient threshold as the first candidate auxiliary features. The cloud can screen the first candidate auxiliary features to obtain auxiliary features.

In some embodiments, there will be multicollinearity between some first candidate auxiliary features, that is, the first candidate auxiliary features are not independent of each other, and one first candidate auxiliary feature may be a linear combination of one or several other first candidate auxiliary features. Exemplarily, taking a navigation application as an example, when a user uses a navigation application, the screen can be in a bright screen state, then there is multicollinearity in the screen foreground usage time, the screen bright screen usage time, and the screen bright screen power consumption. In this case, if the first candidate auxiliary features with multicollinearity are used as auxiliary features to train the power consumption model, the first candidate auxiliary features will affect each other, which is not conducive to the power consumption model outputting correct power consumption detection results.

In some embodiments, the variance inflation factor (VIF) can measure the collinearity between the first candidate auxiliary features. The larger the VIF value, the more obvious the problem of collinearity between the features. In some embodiments, the first VIF threshold and the second VIF threshold can be preset. Exemplarily, the first VIF threshold can be 100, and the second VIF threshold can be 10.

Among them, when the VIF between the first candidate auxiliary features is greater than or equal to 100, it means that there is serious multicollinearity between the first candidate auxiliary features. When the VIF between the first candidate auxiliary features is less than 10, it means that there is no collinearity between the first candidate auxiliary features. When the VIF between the first candidate auxiliary features is greater than or equal to 10 and less than 100, it means that there is strong multicollinearity between the first candidate auxiliary features.

Therefore, in the embodiment of the present application, the cloud can perform the following steps:

Step A: taking any first candidate auxiliary feature as the dependent variable and each other first candidate auxiliary feature as the independent variable, and performing linear regression processing on the first candidate auxiliary feature and the other first candidate auxiliary features, the following formula 5-1 can be referred to:

Wherein, X represents the first candidate auxiliary feature, and the number of the first candidate auxiliary features is p. _Xj represents any first candidate auxiliary feature, and _β0 , _β1 , ..., _βj represent coefficients of linear regression processing. Represents the value of X _j after linear regression processing.

The cloud can use the following formula 5-2 and formula 6 to calculate the VIF between the first candidate auxiliary feature and other first candidate auxiliary features:

Wherein, i represents any one of the power consumption data, n represents the total number of power consumption data, that is, the data volume of the power consumption data. X _ji represents the value of any first candidate auxiliary feature in any one of the power consumption data, Indicates the value of any first candidate auxiliary feature in any power consumption data after linear regression processing, /> represents the mean value of any first candidate auxiliary feature in the n power consumption data. _{R j} ² represents the determination coefficient corresponding to any first candidate auxiliary feature. VIF _j represents the VIF of any first candidate auxiliary feature.

Among them, R _j ² is the determination coefficient in the linear regression processing model, which can measure how much of the change in the dependent variable can be explained by the independent variable. The larger the R _j ² , the larger the VIF _j , the better the model explanatory power, the stronger the linear relationship between the independent and dependent variables, and the more likely it is that multicollinearity exists.

For example, it is assumed that there are 10 first candidate auxiliary features, namely, first candidate auxiliary feature 1, first candidate auxiliary feature 2, ..., and first candidate auxiliary feature 10. The cloud can calculate the VIF of the first candidate auxiliary feature 1 using Formula 5-1, Formula 5-2, and Formula 6.

Step B: repeat step A, taking each other first candidate auxiliary feature as a dependent variable in turn, and calculating the VIF of each other first candidate auxiliary feature.

Exemplarily, the cloud can also calculate the VIF of the first candidate auxiliary feature 2 , the VIF of the first candidate auxiliary feature 3 , . . . , and the VIF of the first candidate auxiliary feature 10 .

In some embodiments, step B and step A may be combined into one step, that is, the cloud obtains the variance inflation coefficient VIF of each first candidate auxiliary feature and other first candidate auxiliary features.

Step C: For the first candidate auxiliary feature corresponding to the maximum VIF, if the maximum VIF is greater than 10 and the number of the first candidate auxiliary features is greater than 3, the first candidate auxiliary feature corresponding to the maximum VIF is deleted, and step D is performed. If the VIF of all first candidate auxiliary features is less than or equal to 10, or the maximum VIF is greater than 10 and the number of the first candidate auxiliary features is less than or equal to 3, all first candidate auxiliary features are used as second candidate auxiliary features. It should be understood that 3 is an example and can be replaced by other values. 3 can be called a quantity threshold. In some embodiments, 10 (second VIF threshold) can be used as the VIF threshold.

Exemplarily, if the first candidate auxiliary feature corresponding to the maximum VIF is the first candidate auxiliary feature 1, and the VIF of the first candidate auxiliary feature 1 is greater than 10, and the number of the first candidate auxiliary features is 10, and the number of the first candidate auxiliary features is greater than 3, then the first candidate auxiliary feature 1 is deleted. If the VIF (maximum VIF) of the first candidate auxiliary feature 1 is less than or equal to 10, then the first candidate auxiliary feature 1, the first candidate auxiliary feature 2, ..., and the first candidate auxiliary feature 10 are all used as the second candidate auxiliary features. Alternatively, the maximum VIF is greater than 10, and the number of the first candidate auxiliary features is less than or equal to 3, such as the VIF of the first candidate auxiliary feature 1 is greater than 10, but there are only three first candidate auxiliary features, namely, the first candidate auxiliary feature 1, the first candidate auxiliary feature 2, and the first candidate auxiliary feature 3, so the first candidate auxiliary feature 1, the first candidate auxiliary feature 2, and the first candidate auxiliary feature 3 can be used as the second candidate auxiliary features.

Step D: For other first candidate auxiliary features after deleting the first candidate auxiliary feature corresponding to the maximum VIF, repeat steps A to C until the maximum VIF is less than or equal to 10, or the number of remaining first candidate auxiliary features is less than or equal to a quantity threshold (such as 3), and the remaining first candidate auxiliary features are used as second candidate auxiliary features.

Exemplarily, if after deleting the first candidate auxiliary feature corresponding to the maximum VIF as the first candidate auxiliary feature 1, there are still 9 first candidate auxiliary features remaining, for these 9 first candidate auxiliary features, steps A to C can be re-executed until the maximum VIF is less than or equal to 10, or the number of remaining first candidate auxiliary features is less than or equal to the quantity threshold, the cloud can use the last remaining first candidate auxiliary feature as the second candidate auxiliary feature.

In some embodiments, the cloud may use the second candidate auxiliary feature as the auxiliary feature.

In some embodiments, the cloud can also screen the second candidate auxiliary features to obtain the auxiliary features. Exemplarily, the cloud can use a least absolute shrinkage and selection operator (LASSO) model to screen the second candidate auxiliary features to obtain the auxiliary features.

The LASSO model is essentially a linear regression model, but it is different from the traditional linear model in that it adds an L1 regularized penalty function. Therefore, the LASSO model has the following two characteristics:

1. Compress some regression coefficients, that is, force the sum of the absolute values of the coefficients to be less than a fixed value.

2. Make some regression coefficients zero to form sparse results, which is conducive to screening out important features.

In the power consumption data of the first application of the terminal device of the first product series within a preset time period, the cloud can obtain the value of the main feature and the value of the second candidate auxiliary feature. The cloud can input the value of the main feature and the value of the second candidate auxiliary feature into the LASSO model to obtain the importance value of each second candidate auxiliary feature.

In the embodiment of the present application, the cloud can use the second candidate auxiliary features corresponding to the three largest importance values as auxiliary features. The three are examples and can be replaced by other values.

In some embodiments, the LASSO model may be as shown in Formula 7:

Wherein, P represents the number of second auxiliary features, j represents any second auxiliary feature. λ is a hyperparameter and can be a preset value. β ₀ is a constant. _Yi represents the value of the main feature in any one of the n power consumption data, and _Xji represents the value of any second auxiliary feature in any one of the n power consumption data.

In the embodiment of the present application, the cloud can input the value of the main feature and the value of the second candidate auxiliary feature into the LASSO model to obtain β _j of any second auxiliary feature. The cloud takes the absolute value of β _j , and the absolute value of β _j can represent the importance of the second candidate auxiliary feature. The larger the absolute value of β _j , the greater the importance value of the second candidate auxiliary feature.

FIG5A is a schematic diagram of a screening auxiliary feature provided by an embodiment of the present application. FIG5A a shows all the features in the power consumption data. After calculating the Pearson correlation coefficient and the Spearman correlation coefficient, the first candidate auxiliary feature is screened out. Referring to FIG5A b, the first candidate auxiliary feature may include: camera foreground power consumption, camera foreground usage time, modem foreground power consumption, modem background power consumption, Wi-Fi module foreground power consumption, Wi-Fi module foreground total flow, and Wi-Fi module background power consumption. After VIF calculation and screening, the second candidate auxiliary feature can be obtained. Referring to FIG5A c, the second candidate auxiliary feature may include: camera foreground usage time, modem background power consumption, Wi-Fi module foreground power consumption, Wi-Fi module foreground total flow, and Wi-Fi module background power consumption. After LASSO model processing, auxiliary features can be obtained. Referring to FIG5A d, the auxiliary feature may include: modem background power consumption, Wi-Fi module foreground total flow, and Wi-Fi module background power consumption.

S305, the cloud obtains auxiliary feature conditions, and the auxiliary feature conditions are used to determine the user group to which the user belongs.

In the embodiment of the present application, after determining the auxiliary feature, the cloud can also obtain the auxiliary feature condition. The auxiliary feature condition is used to determine the user group to which the user belongs.

In some embodiments, the cloud can train a classification and regression tree (CART) based on the auxiliary features and the power consumption data of the first application of the terminal device of the first product series within a preset time period, and use the leaf nodes in the CART as auxiliary feature conditions. It should be understood that the habits of the user groups in each leaf node are similar. It should be understood that the power consumption data of the first application of the terminal device of the first product series within the preset time period used in S305-S307 is data that has not been normalized by Z-Score.

In some embodiments, in order to prevent CART from overfitting and ensure that each leaf node has sufficient data, the following hyperparameters can be set when training CART in the embodiments of the present application:

1. The loss function is mean square error.

2. The maximum depth is 3. Among them, the root node can be used as the 0th layer.

3. The maximum number of leaf nodes is 8.

4. The minimum amount of data for a leaf node is 5,000.

FIG4 is a schematic diagram of a CART provided in an embodiment of the present application. FIG4 takes the CART including 4 leaf nodes as an example, and with reference to FIG4, the auxiliary feature conditions may include: "The background power consumption of the Wi-Fi module is less than or equal to 5.732, and the background power consumption of the modem is less than or equal to 0.942", and "The background power consumption of the Wi-Fi module is less than or equal to 5.732, and the background power consumption of the modem is greater than 0.942", and "The background power consumption of the Wi-Fi module is greater than 5.732, and the background power consumption of the modem is less than or equal to 24186.737", and "The background power consumption of the Wi-Fi module is greater than 5.732, and the background power consumption of the modem is greater than 24186.737".

S306, the cloud divides the users corresponding to the terminal device into at least two user groups according to the auxiliary feature condition and the power consumption data of the first application of the terminal device of the first product series within a preset time period.

Exemplarily, using the auxiliary feature conditions in FIG. 4, users can be divided into four user groups, namely user group 1, user group 2, user group 3 and user group 4. For example, the power consumption data of the terminal device corresponding to user group 1 satisfies "the background power consumption of the Wi-Fi module is less than or equal to 5.732, and the background power consumption of the modem is less than or equal to 0.942", the power consumption data of the terminal device corresponding to user group 1 satisfies "the background power consumption of the Wi-Fi module is less than or equal to 5.732, and the background power consumption of the modem is greater than 0.942", the power consumption data of the terminal device corresponding to user group 3 satisfies "the background power consumption of the Wi-Fi module is greater than 5.732, and the background power consumption of the modem is less than or equal to 24186.737", and the power consumption data of the terminal device corresponding to user group 4 satisfies "the background power consumption of the Wi-Fi module is greater than 5.732, and the background power consumption of the modem is greater than 24186.737".

S307, the cloud determines the threshold of the main feature corresponding to each user group based on the power consumption data of the first application of the terminal device of the first product series within a preset time of each user group, and the threshold of the main feature is used to detect whether the power consumption of the first application is abnormal.

After the cloud divides the user groups based on the auxiliary feature conditions, the cloud can determine the threshold of the main feature corresponding to each user group based on the power consumption data of the first application of the terminal device of the first product series within the preset time of each user group (referred to as power consumption data). The threshold of the main feature is used to detect whether the power consumption of the first application is abnormal. Exemplarily, if the main feature is CPU background power consumption, the cloud can determine the threshold of the CPU background power consumption corresponding to each user group based on the power consumption data of each user group. The threshold of the CPU background power consumption corresponding to each user group is used to detect whether the CPU background power consumption in the power consumption data of the user group is abnormal. The CPU background power consumption can indicate that the power consumption of the first application is abnormal.

In some embodiments, a first preset ratio may be pre-set, and the first preset ratio is used to determine the threshold of the main feature. Taking the main feature as CPU background power consumption as an example, the cloud can sort the CPU background power consumption values in the power consumption data of each user group in order of size, and use the first value of the CPU background power consumption value that is greater than or equal to 90% as the threshold of the CPU background power consumption.

Exemplarily, taking user group 1 as an example, if the first preset ratio is 90%, a in FIG5B represents the distribution of the value of the CPU background power consumption in the power consumption data corresponding to user group 1 with a rectangular box. In the value of the CPU background power consumption in the power consumption data corresponding to user group 1, the cloud can use the first value of the CPU background power consumption greater than or equal to 90% as the threshold of the CPU background power consumption. Exemplarily, if the 100 power consumption data corresponding to user group 1 are arranged in order from small to large according to the value of the CPU background power consumption, the first value greater than or equal to the value of the CPU background power consumption in the 90th power consumption data is used as the threshold of the CPU background power consumption. Exemplarily, the value of the CPU background power consumption in the 90th power consumption data is A, and the value of the CPU background power consumption in the 91st power consumption data is B, then any value of A and B can be used as the first value, or A or B can be used as the first value, that is, the threshold of the CPU background power consumption.

The threshold of CPU background power consumption is used to detect whether the power consumption of the first application is abnormal. When the value of the CPU background power consumption in the power consumption data of the first application is greater than or equal to the threshold of the CPU background power consumption, it can be determined that the power consumption of the first application is abnormal; when the value of the CPU background power consumption in the power consumption data of the first application is less than the threshold of the CPU background power consumption, it can be determined that the power consumption of the first application is normal.

In some embodiments, the cloud can also divide the abnormality level. In the embodiment of the present application, a second preset ratio and a third preset ratio can also be preset. Exemplarily, the second preset ratio can be 98%, and the third preset ratio can be 99.5%. Referring to b in Figure 5B, b in Figure 5B represents the numerical distribution of CPU background power consumption in the power consumption data corresponding to user group 1 with a rectangular box. The cloud can calculate a second value of the numerical value of the CPU background power consumption that is greater than or equal to the second preset ratio (98%), and a third value of the numerical value of the CPU background power consumption that is greater than or equal to the third preset ratio (99.5%).

In this embodiment, the cloud can use the first value as the first threshold of the CPU background power consumption during an abnormality, the second value as the second threshold of the CPU background power consumption during a super abnormality, and the third value as the third threshold of the CPU background power consumption during an extreme abnormality.

In some embodiments, the threshold of CPU background power consumption may include: a first threshold of CPU background power consumption for detecting whether the power consumption of the first application is abnormal, a second threshold of CPU background power consumption for detecting whether the power consumption of the first application is extremely abnormal, and a third threshold of CPU background power consumption for detecting whether the power consumption of the first application is extremely abnormal. When the value of the CPU background power consumption in the power consumption data of the first application is greater than or equal to the first threshold and less than the second threshold, it can be determined that the power consumption of the first application is abnormal, when the value of the CPU background power consumption in the power consumption data of the first application is greater than or equal to the second threshold and less than the third threshold, it can be determined that the power consumption of the first application is extremely abnormal, and when the value of the CPU background power consumption in the power consumption data of the first application is greater than or equal to the third threshold, it can be determined that the power consumption of the first application is extremely abnormal.

In some embodiments, "abnormal, super abnormal, and extremely abnormal" can be respectively used as three abnormal levels of power consumption abnormality of the first application. Exemplarily, when the value of the CPU background power consumption in the power consumption data of the first application is greater than or equal to the first threshold and less than the second threshold, the power consumption abnormality of the first application belongs to the first abnormal level, when the value of the CPU background power consumption in the power consumption data of the first application is greater than or equal to the second threshold and less than the third threshold, the power consumption abnormality of the first application belongs to the second abnormal level, and when the value of the CPU background power consumption in the power consumption data of the first application is greater than or equal to the third threshold, the power consumption abnormality of the first application belongs to the third abnormal level. Among them, the abnormality degree of the first abnormality level is less than the abnormality degree of the second abnormality level, and the abnormality degree of the second abnormality level is less than the abnormality degree of the third abnormality level.

In some embodiments, for applications with a large amount of data, the distribution of the values of the main features is relatively dense, and there will be a problem that the thresholds corresponding to different abnormal levels are slightly different. In the embodiment of the present application, the cloud can adjust the second threshold and the third threshold so that based on the adjusted second threshold and the adjusted third threshold, the abnormal level of the first application can be clearly distinguished.

For the second threshold, the cloud's adjustment rules are as follows:

The cloud can use the following formula 8 to obtain the normal distribution Z value of the second threshold to detect whether the second threshold needs to be adjusted:

Wherein, Q _0.98 represents the second threshold, μ represents the mean value of the CPU background power consumption corresponding to each user group, and σ represents the standard deviation of the CPU background power consumption corresponding to each user group.

Wherein, based on Formula 8, if the obtained normal distribution Z value is less than the preset Z value, the cloud can increase the second threshold. The cloud uses the increased second threshold and recalculates using Formula 8, and the obtained normal distribution Z value is equal to the preset Z value. For example, the preset Z value can be 3.

The adjusted second threshold may be a maximum value between the original second threshold and the increased second threshold. For example, the adjusted second threshold may be max(Q _0.98 , 3σ+μ).

For the third threshold, the cloud's adjustment rules are as follows:

The cloud can use the following formula 9 to obtain the normal distribution Z value of the third threshold to detect whether the third threshold needs to be adjusted:

Wherein, Q _0.995 represents the third threshold. Based on Formula 9, if the obtained normal distribution Z value is less than the preset Z value, the cloud can increase the third threshold. The cloud uses the increased third threshold and recalculates using Formula 9, and the obtained normal distribution Z value is equal to the preset Z value. Exemplarily, the increased third threshold can be referred to as the first candidate third threshold.

In addition, the adjusted third threshold value needs to be greater than the outlier of the box plot. The outlier of the box plot can be calculated using the following formula 10:

outlier = Q _0.75 + 1.5 (Q _0.75 - Q _0.25 ) Formula 10

Among them, Q _0.75 represents a value greater than or equal to 75% of the values of the CPU background power consumption in the power consumption data corresponding to the user group, and Q _0.25 represents a value greater than or equal to 25% of the values of the CPU background power consumption in the power consumption data corresponding to the user group. Exemplarily, the adjusted third threshold, i.e., the third threshold greater than the outlier of the box plot, can be called the second candidate third threshold.

It is understandable that in the embodiment of the present application, the adjusted third threshold may be the maximum value among the original third threshold, the first candidate third threshold, and the second candidate third threshold. For example, the adjusted third threshold may be max(Q _0.995 , 3σ+μ, outlier).

S308, the cloud generates a power consumption model of the first application of the terminal device of the first product series according to the auxiliary feature conditions and the threshold of the main feature corresponding to each user group.

In some embodiments, the power consumption model of the first application of the terminal device of the first product series may be referred to as the power consumption model of the first application of the first product series. The power consumption model of the first application of the first product series may include: auxiliary feature conditions for dividing user groups, and thresholds of main features corresponding to each user group. In some embodiments, the threshold of the main feature may include a first threshold. In some embodiments, the threshold of the main feature may include: a first threshold, a second threshold, and a third threshold. In some embodiments, the second threshold in the threshold of the main feature may be an adjusted second threshold, and the third threshold in the main feature may be an adjusted third threshold.

The threshold of the main feature corresponding to each user group is used to detect whether the power consumption of the first application is abnormal and to detect the abnormal level of the power consumption of the first application. In some embodiments, the threshold of the main feature corresponding to each user group can be referred to as the processing logic of the power consumption of the first application.

In some embodiments, the cloud can store the power consumption model of the first application of the first product series in a table form, a text form, etc., and the embodiment of the present application does not limit the form of the power consumption model. Table 3 shows the power consumption model of the first application of the first product series in a table form.

Table 3

In Table 3, inf means positive infinity.

In the embodiment of the present application, the cloud can generate and update the power consumption model based on the latest power consumption data of the terminal device, and the power consumption model has high real-time and accuracy. In addition, for different product series and for different applications, power consumption models for different product series and for different applications can be generated respectively, with finer granularity and higher accuracy.

After the cloud generates a power consumption model or updates the power consumption model, it can be sent to the terminal device. The sending method can refer to the relevant description in the above embodiments. In some embodiments, the terminal device can use the power consumption model to detect whether the power consumption of the application of the terminal device is abnormal, and the abnormal level, and then perform corresponding operations to reduce the power consumption of the application.

FIG6 is a flow chart of another embodiment of a method for processing power consumption of a terminal device provided in an embodiment of the present application. Referring to FIG6 , the method for processing power consumption of a terminal device provided in an embodiment of the present application may include:

S601, a terminal device collects power consumption data of the terminal device, where the power consumption data of the terminal device includes power consumption data of a first application, and the terminal device belongs to a first product series.

It should be understood that the embodiment of the present application takes the terminal device belonging to the first product series as an example to illustrate the process of obtaining the power consumption detection result of the first application by the terminal device using the power consumption model of the first application of the first product series and the power consumption data of the first application. In some embodiments, the first application can be an application running in the foreground of the terminal device and/or an application running in the background.

In an embodiment of the present application, the terminal device may collect power consumption data of the terminal device, and the power consumption data of the terminal device may include power consumption data of the first application.

It should be understood that when the power consumption model sent down from the cloud is the power consumption model of different applications and does not distinguish between product series, the terminal device can use the power consumption model of the first application to detect whether the power consumption of the first application is abnormal, and can refer to the relevant description of the power consumption model of the first application in the first product series.

S602: The terminal device determines the user group to which the user belongs and the threshold of the main feature corresponding to the user group according to the power consumption model of the first application of the first product series and the power consumption data of the first application.

The terminal device can determine the user group to which the user belongs and the threshold of the main feature corresponding to the user group based on the power consumption data of the first application and the auxiliary feature condition in the power consumption model of the first application of the first product series. The user can be understood as the user corresponding to the terminal device.

For example, in the power consumption data of the first application, the background power consumption of the Wi-Fi module is 6, and the background power consumption of the modem is 20. Based on the power consumption model in Table 3, the terminal device can determine that the user belongs to user group 3. The thresholds of the main features corresponding to user group 3 include: the first threshold (90.83), the second threshold (146.64), and the third threshold (203.57).

S603: The terminal device obtains a power consumption detection result of the first application according to the threshold of the main feature corresponding to the user group and the power consumption data of the first application.

The power consumption data of the first application may include the value of the main feature. Exemplarily, if the main feature is CPU background power consumption, the power consumption data of the first application may include the value of CPU background power consumption. In some embodiments, the power consumption detection result of the first application is used to indicate whether the power consumption of the first application is abnormal. When the value of the main feature of the power consumption data of the first application is greater than or equal to the threshold of the main feature corresponding to the user group, it can be determined that the power consumption of the first application is abnormal, and when the value of the main feature of the power consumption data of the first application is less than the threshold of the main feature corresponding to the user group, it can be determined that the power consumption of the first application is normal.

In some embodiments, when the threshold of the main feature in the power consumption model includes abnormal thresholds of different levels, the terminal device can detect whether the power consumption of the first application is abnormal based on the threshold of the main feature corresponding to the user group and the power consumption data of the first application, and can determine the abnormal level of the power consumption of the first application. In this embodiment, the power consumption detection result of the first application is used to indicate whether the power consumption of the first application is abnormal and the abnormal level of the power consumption of the first application.

For example, if the user belongs to user group 3, referring to Table 3, the thresholds of the main features corresponding to user group 3 include: a first threshold (90.83), a second threshold (146.64), and a third threshold (203.57). For example, in the power consumption data of the first application, the value of the CPU background power consumption is 100. Based on the threshold of the main feature, the cloud can determine that the value of the CPU background power consumption is greater than the first threshold and less than the second threshold, and can determine that the power consumption of the first application is abnormal, and the power consumption abnormality of the first application is the first abnormality level.

S604: When the power consumption of the first application is abnormal, the terminal device performs a corresponding operation.

In the embodiment of the present application, when the power consumption of the first application is abnormal, the terminal device performs a corresponding operation. Exemplarily, the terminal device can output a prompt message, which is used to prompt that the power consumption of the first application is abnormal.

In some embodiments, when the power consumption of the first application is abnormal, and the terminal device can determine the abnormal level of the power consumption of the first application, the terminal device can perform corresponding operations according to the abnormal level of the power consumption of the first application. The operations performed by the terminal device may be different for different abnormal levels.

Exemplarily, when the abnormal level of the power consumption of the first application is the first abnormal level, the terminal device can reduce the frequency of the CPU used by the first application to the first frequency. When the abnormal level of the power consumption of the first application is the second abnormal level, the terminal device can reduce the frequency of the CPU used by the first application to the second frequency. When the abnormal level of the power consumption of the first application is the third abnormal level, the terminal device can reduce the frequency of the CPU used by the first application to the third frequency. Among them, the first frequency is greater than the second frequency, and the second frequency is greater than the third frequency. In other words, the more severe or higher the abnormal level of the power consumption of the first application is, the more the CPU frequency decreases.

In the embodiment of the present application, there is no restriction on the operations of terminal devices corresponding to different abnormality levels.

In the embodiment of the present application, the terminal device does not manage the power consumption of all terminal devices based on a fixed power consumption threshold, but for different users, the user group to which the user belongs can be determined, and the thresholds of the main features corresponding to different user groups can be different. The terminal device can process the power consumption of the terminal device according to the threshold of the main feature corresponding to the user group to which the user belongs, so that the terminal device can flexibly control the power consumption of the terminal device based on the user's characteristics or usage habits. In addition, when the power consumption of the terminal device is abnormal, the terminal device can also determine the abnormal level. For different abnormal levels, the terminal device can perform different response operations, which is highly flexible.

The following describes the power consumption processing method of the terminal device provided by the embodiment of the present application from the perspective of module interaction in the terminal device. FIG7 is a flow chart of another embodiment of the power consumption processing method of the terminal device provided by the embodiment of the present application. Referring to FIG7, the power consumption processing method of the terminal device provided by the embodiment of the present application may include:

S701, the power consumption processing process collects power consumption data of a terminal device, the power consumption data of the terminal device includes power consumption data of a first application, and the terminal device belongs to a first product series.

S701 may refer to the description of S201 in FIG. 2B .

S702, the power consumption processing process sends the power consumption data of the terminal device to the calculation engine through the power consumption interface.

S703: The computing engine calls the power consumption model of the first application of the first product series, and determines the user group to which the user belongs and the threshold of the main feature corresponding to the user group according to the power consumption data of the first application.

S704: The calculation engine obtains a power consumption detection result of the first application according to the threshold of the main feature corresponding to the user group and the power consumption data of the first application.

S703 - S704 may refer to the description in S602 - S603 .

S705: The computing engine sends the power consumption detection result of the first application to the power consumption processing process.

S706, the power consumption processing process sends a control instruction to the power consumption response unit based on the power consumption detection result of the first application, when determining that the power consumption of the first application is abnormal, wherein the control instruction is used to instruct the power consumption response unit to perform a corresponding operation.

S707, the power consumption response unit executes a corresponding operation in response to the control instruction.

S707 may refer to the description in S604.

The power consumption processing method of the terminal device provided in the embodiment of the present application has the same technical principle and technical effect as the embodiment shown in Figure 6, and can refer to the relevant description in the above embodiment.

FIG8 is a flow chart of another embodiment of a method for processing power consumption of a terminal device provided in an embodiment of the present application. FIG8 combines the steps in FIG2B, FIG3, and FIG6 and FIG7. Referring to FIG8, the method for processing power consumption of a terminal device provided in an embodiment of the present application may include:

S801, the cloud determines the top 150 applications as target applications based on the power consumption data of at least one terminal device.

S802, deleting resident applications in the cloud.

S803, the cloud obtains the power consumption data of the top 150 applications of terminal devices of different product series in the past 30 days.

S804, the cloud determines the data volume of the power consumption data.

S801-S804 may refer to the relevant description in S301.

S805, the cloud pre-processes the power consumption data.

S805 may refer to the relevant description in S302.

S806, cloud-based screening of auxiliary features.

S806 may refer to the description in S303-S304.

S807, the cloud performs data decomposition on the power consumption data based on the auxiliary features.

The cloud performs data decomposition on the power consumption data, which can be understood as: the cloud obtains auxiliary feature conditions, which can refer to the description in S305.

S808, the cloud obtains the threshold of the main feature corresponding to each user group.

S808 may refer to the description in S306-S307.

In some embodiments, the cloud may also adjust the second threshold and the third threshold in the threshold of the main feature, and reference may be made to the relevant description in S307.

S809, cloud output power consumption model.

S810: The terminal device processes the power consumption of the terminal device according to the power consumption model.

S810 may refer to the description in the embodiments shown in FIG. 6 and FIG. 7 .

It should be understood that in FIG. 8 , solid lines are used to divide the steps in the cloud and the steps in the terminal device, and dashed lines are used to divide the different steps performed in the cloud.

The power consumption processing method of the terminal device provided in the embodiments of the present application has the same technical principles and technical effects as the embodiments shown in Figures 2B, 3, 6 and 7, and reference may be made to the relevant descriptions in the above embodiments.

Those skilled in the art can understand that the system architecture and the power consumption processing method of the terminal device in the embodiment of the present application can be combined and referenced with each other, and the system architecture provided in the embodiment of the present application can execute the steps in the power consumption processing method of the above-mentioned terminal device.

It should be noted that the power consumption data involved in this application (including but not limited to data used for analysis, storage, display, etc.) are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

The power consumption processing method of the terminal device provided in the embodiment of the present application can be applied to electronic devices with communication functions. The electronic device includes the terminal device, and the specific device form of the terminal device can refer to the above related description, which will not be repeated here.

An embodiment of the present application provides a terminal device, which includes: a processor and a memory; the memory stores computer-executable instructions; the processor executes the computer-executable instructions stored in the memory, so that the terminal device executes the above method.

The embodiment of the present application also provides an electronic device, which may be the terminal device or the cloud described in the above embodiment. Referring to Figure 9, the electronic device may include: a processor 901 (e.g., a CPU) and a memory 902. The memory 902 may include a high-speed random access memory (RAM), and may also include a non-volatile memory (NVM), such as at least one disk memory. Various instructions may be stored in the memory 902 to complete various processing functions and implement the method steps of the present application.

Optionally, the electronic device involved in the present application may further include: a power supply 903, a communication bus 904 and a communication port 905. The above-mentioned communication port 905 is used to realize the connection and communication between the electronic device and other peripherals. In the embodiment of the present application, the memory 902 is used to store computer executable program code, and the program code includes instructions; when the processor 901 executes the instruction, the instruction causes the processor 901 of the electronic device to perform the action in the above-mentioned method embodiment, and its implementation principle and technical effect are similar, which will not be repeated here.

The embodiment of the present application provides a chip. The chip includes a processor, and the processor is used to call a computer program in a memory to execute the technical solution in the above embodiment. Its implementation principle and technical effect are similar to those of the above related embodiments, and will not be repeated here.

The embodiments of the present application also provide a computer-readable storage medium. The computer-readable storage medium stores a computer program. The above method is implemented when the computer program is executed by the processor. The method described in the above embodiment can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. If implemented in software, the function can be stored as one or more instructions or codes on a computer-readable medium or transmitted on a computer-readable medium. Computer-readable media can include computer storage media and communication media, and can also include any medium that can transfer a computer program from one place to another. The storage medium can be any target medium that can be accessed by a computer.

In one possible implementation, a computer-readable medium may include RAM, ROM, compact disc read-only memory (CD-ROM) or other optical disk storage, disk storage or other magnetic storage devices, or any other medium that is intended to carry or store the required program code in the form of instructions or data structures and can be accessed by a computer. Moreover, any connection is appropriately referred to as a computer-readable medium. For example, if a coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL) or wireless technology (such as infrared, radio and microwave) is used to transmit software from a website, server or other remote source, the coaxial cable, fiber optic cable, twisted pair, DSL or wireless technology such as infrared, radio and microwave are included in the definition of medium. Disks and optical disks as used herein include optical disks, laser disks, optical disks, digital versatile disks (DVD), floppy disks and Blu-ray disks, where disks usually reproduce data magnetically, while optical disks reproduce data optically using lasers. Combinations of the above should also be included in the scope of computer-readable media.

An embodiment of the present application provides a computer program product, which includes a computer program. When the computer program is executed, the computer executes the above method.

The present application embodiment is described with reference to the flowchart and/or block diagram of the method, device (system) and computer program product according to the present application embodiment. It should be understood that each flow process and/or box in the flow chart and/or block diagram and the combination of the flow chart and/or box in the flow chart and/or block diagram can be realized by computer program instructions. These computer program instructions can be provided to the processing unit of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable device to produce a machine, so that the instructions executed by the processing unit of the computer or other programmable data processing device produce a device for realizing the function specified in one flow chart or multiple flows and/or one box or multiple boxes of the block chart.

The above specific implementation methods further illustrate the purpose, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above are only specific implementation methods of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solutions of the present invention should be included in the protection scope of the present invention.

Claims (17)

1. A method for processing power consumption of a terminal device, characterized in that it is applied to the terminal device, and the method comprises: Collecting power consumption data of the terminal device, wherein the power consumption data includes power consumption data of the first application; Determine, according to the power consumption model of the first application, a user group to which the user corresponding to the terminal device belongs, and a threshold value of a main feature corresponding to the user group; detecting whether the power consumption of the first application is abnormal according to the threshold of the main feature and the value of the main feature in the power consumption data of the first application, and obtaining a power consumption detection result of the first application; Execute a corresponding operation according to the power consumption detection result of the first application; The power consumption model of the first application includes: an auxiliary feature condition and a threshold value of a main feature corresponding to at least one user group, wherein the auxiliary feature condition is used to determine the user group to which the user belongs; The determining, according to the power consumption model of the first application, a user group to which the user corresponding to the terminal device belongs and a threshold value of a main feature corresponding to the user group includes: Determine the user group to which the user corresponding to the terminal device belongs according to the auxiliary feature condition and the value of the auxiliary feature in the power consumption data of the first application; The threshold of the main feature corresponding to the user group to which the user corresponding to the terminal device belongs is determined according to the threshold of the main feature corresponding to the at least one user group and the user group to which the user corresponding to the terminal device belongs. 2. The method according to claim 1 is characterized in that the power consumption model of the first application is specifically the power consumption model of the first application of the first product series, and the terminal device belongs to the first product series. 3. The method according to claim 1, characterized in that the detecting whether the power consumption of the first application is abnormal according to the threshold of the main feature and the value of the main feature in the power consumption data of the first application to obtain the power consumption detection result of the first application comprises: When the value of the main feature is greater than or equal to the threshold of the main feature, determining that the power consumption of the first application is abnormal, and the power consumption detection result of the first application is used to indicate that the power consumption of the first application is abnormal; When the value of the main feature is less than the threshold of the main feature, it is determined that the power consumption of the first application is normal, and the power consumption detection result of the first application is used to indicate that the power consumption of the first application is normal. 4. The method according to claim 1, characterized in that the threshold of the main feature includes: a first threshold, a second threshold, and a third threshold, the first threshold is smaller than the second threshold, and the second threshold is smaller than the third threshold; The detecting whether the power consumption of the first application is abnormal according to the threshold of the main feature and the value of the main feature in the power consumption data of the first application, and obtaining the power consumption detection result of the first application includes: When the value of the main feature is greater than or equal to the first threshold and less than the second threshold, it is determined that the power consumption abnormality of the first application is at a first abnormality level, and the power consumption detection result of the first application is used to indicate that the level of the power consumption abnormality of the first application is at the first abnormality level; When the value of the main feature is greater than or equal to the second threshold and less than the third threshold, it is determined that the power consumption abnormality of the first application is at a second abnormality level, and the power consumption detection result of the first application is used to indicate that the level of the power consumption abnormality of the first application is at the second abnormality level; When the value of the main feature is greater than or equal to the third threshold, determining that the power consumption abnormality of the first application is at a third abnormality level, and the power consumption detection result of the first application is used to indicate that the level of the power consumption abnormality of the first application is at the third abnormality level; When the value of the main feature is less than the first threshold, it is determined that the power consumption of the first application is normal, and the power consumption detection result of the first application is used to indicate that the power consumption of the first application is normal. 5. The method according to claim 4 is characterized in that the power consumption abnormality level of the first application is different, and the corresponding operations performed by the terminal device are different. 6. The method according to any one of claims 1 to 5, characterized in that the method further comprises: Send a request to the cloud for obtaining the power consumption model; A power consumption model of the first application is received from the cloud. 7. The method according to any one of claims 1 to 5, characterized in that after collecting the power consumption data of the terminal device, it also includes: The power consumption data of the terminal device is sent to the data platform. 8. A method for processing power consumption of a terminal device, characterized in that it is applied to the cloud, and the method comprises: Acquire power consumption data of at least one terminal device from a data platform; the power consumption data includes power consumption data of at least one application, a first application is included in the at least one application, and the power consumption model includes a power consumption model of the first application; the power consumption model of the first application is specifically a power consumption model of a first application of a first product series; Training a power consumption model according to the power consumption data of the at least one terminal device; Receiving a request for obtaining a power consumption model from a terminal device; Sending the power consumption model to the terminal device; The power consumption model of the first application of the first product series is obtained by training, including: Obtaining a correlation coefficient between a main feature and each other feature according to power consumption data of a first application of a terminal device of the first product series within a preset time period; Determine the auxiliary feature according to the correlation coefficient between the main feature and each other feature; Acquire an auxiliary feature condition according to the power consumption data of the first application of the terminal device of the first product series within the preset time period and the auxiliary feature, wherein the auxiliary feature condition is used to determine the user group to which the user belongs; According to the auxiliary feature condition and the power consumption data of the first application of the terminal device of the first product series within the preset time period, users corresponding to the at least one terminal device are divided into at least two user groups; Determine, according to the power consumption data of the first application of the terminal device of the first product series within a preset time of each user group, a threshold value of the main feature corresponding to each user group, wherein the threshold value of the main feature is used to detect whether the power consumption of the first application is abnormal; A power consumption model of the first application of the first product series is generated according to the auxiliary feature condition and the threshold value of the main feature corresponding to each user group. 9. The method according to claim 8, characterized in that the correlation coefficient between the main feature and each of the other features comprises a Pearson correlation coefficient and a Spearman correlation coefficient; Determining the auxiliary feature according to the correlation coefficient between the main feature and each other feature includes: The features that are greater than or equal to the Pearson correlation coefficient threshold and greater than or equal to the Spearman correlation coefficient threshold are taken as the first candidate auxiliary features; The auxiliary feature is determined according to the first candidate auxiliary feature. 10. The method according to claim 9, characterized in that determining the auxiliary feature according to the first candidate auxiliary feature comprises: Step A, obtaining the variance expansion factor VIF of each first candidate auxiliary feature and other first candidate auxiliary features; Step B, for the first candidate auxiliary feature corresponding to the maximum VIF, if the maximum VIF is greater than the VIF threshold and the number of the first candidate auxiliary features is greater than the number threshold, executing step C; if the maximum VIF is greater than the VIF threshold and the number of the first candidate auxiliary features is less than or equal to the number threshold, or the maximum VIF is less than or equal to the VIF threshold, all the first candidate auxiliary features are used as second candidate auxiliary features, and the auxiliary feature is determined according to the second candidate auxiliary features; Step C, delete the first candidate auxiliary feature corresponding to the maximum VIF, and return to execute steps A-B until the maximum VIF is less than or equal to the VIF threshold, or the number of remaining first candidate auxiliary features is less than or equal to the number threshold, and use the remaining first candidate auxiliary features as second candidate auxiliary features, and determine the auxiliary feature based on the second candidate auxiliary feature. 11. The method according to claim 10, characterized in that determining the auxiliary feature according to the second candidate auxiliary feature comprises: Obtaining the importance value of each second candidate auxiliary feature according to the minimum absolute shrinkage and selection model; The second candidate auxiliary features ranked in the top N are used as the auxiliary features, where N is equal to the quantity threshold. 12. The method according to any one of claims 8 to 11, characterized in that the acquiring of the auxiliary feature condition according to the power consumption data of the first application of the terminal device of the first product series within the preset time period and the auxiliary feature comprises: Training a classification and regression tree CART according to the auxiliary feature and the power consumption data of the first application of the terminal device of the first product series within the preset time period; The leaf nodes in the CART are used as the auxiliary feature conditions. 13. The method according to any one of claims 8 to 11, characterized in that the step of determining the threshold of the main feature corresponding to each user group according to the power consumption data of the first application of the terminal device of the first product series within a preset time of each user group comprises: In the power consumption data of the first application of the terminal device of the first product series within the preset time of each user group, the threshold of the main feature corresponding to each user group is determined according to the value of the main feature and the first preset ratio. 14. The method according to claim 13, characterized in that the threshold of the main feature corresponding to each user group determined based on the first preset ratio is a first threshold; the method further comprises: Determine, in the power consumption data of the first application of the terminal device of the first product series within the preset time of each user group, a second threshold value of the main feature corresponding to each user group according to the value of the main feature and a second preset ratio; In the power consumption data of the first application of the terminal device of the first product series within the preset time of each user group, the third threshold of the main feature corresponding to each user group is determined according to the value of the main feature and the third preset ratio, the first preset ratio is smaller than the second preset ratio, the second preset ratio is smaller than the third preset ratio, and the first threshold, the second threshold, and the third threshold are used to determine the level of power consumption anomaly of the first application. 15. A terminal device, comprising: a processor and a memory; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the terminal device executes the method according to any one of claims 1 to 7. 16. An electronic device, comprising: a processor and a memory; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the electronic device performs the method according to any one of claims 8 to 14. 17. A computer-readable storage medium storing a computer program, wherein the computer program implements the method according to any one of claims 1 to 14 when executed by a processor.