CN117295008B

CN117295008B - Information pushing method and device

Info

Publication number: CN117295008B
Application number: CN202311576011.XA
Authority: CN
Inventors: 常生俊; 陈天辉
Original assignee: Honor Device Co Ltd
Current assignee: Honor Device Co Ltd
Priority date: 2023-11-24
Filing date: 2023-11-24
Publication date: 2024-04-05
Anticipated expiration: 2043-11-24
Also published as: CN118200845A; CN117295008A; CN118200845B

Abstract

The application relates to the technical field of terminal positioning, in particular to an information pushing method and equipment. An information pushing method, the method comprising: continuously acquiring cell information accessed by the terminal equipment according to a preset time interval in the moving process of the terminal equipment; when a cell accessed by the terminal equipment is switched from a first cell to a second cell and the second cell is a target cell, predicting a first time according to the continuously acquired cell information; and pushing service information aiming at the target behavior after the first time.

Description

Information pushing method and device

Technical Field

The application relates to the technical field of terminal positioning, in particular to an information pushing method and equipment.

Background

The terminal device may determine its own position by means of positioning techniques. In some application scenarios, when the terminal device arrives at a certain location, the relevant application or service may push location-related service information for the terminal device.

In a specific implementation, the terminal device may determine whether itself reaches the corresponding location through connected cell (cell) information, e.g., cell Identification (ID). That is, when the terminal device is connected to the target cell, the terminal device is considered to reach the corresponding location, and the service information can be triggered. However, the coverage area of some cells is larger, and the real position of the terminal equipment after being connected to the target cell may be further away from the corresponding position, so that the false triggering of the service information is caused, and the user experience is affected.

Disclosure of Invention

Aiming at the problem that in the prior art, terminal equipment cannot accurately judge whether the terminal equipment reaches the corresponding position, so that service information is triggered by mistake, the application provides an information pushing method and information pushing equipment.

In a first aspect, an embodiment of the present invention provides an information pushing method, where the method includes:

continuously acquiring cell information accessed by terminal equipment according to a preset time interval;

when a cell accessed by a terminal device is switched from a first cell to a second cell and the second cell is a target cell, predicting a first time according to continuously acquired cell information;

service information for the target behavior is pushed after a first time.

According to the embodiment of the invention, the route of the terminal equipment to the target area is determined by collecting the cell accessed by the terminal equipment, and the time consumed by the path from entering the target cell to actually reaching the target area is predicted by learning the travel habit and time rule on the corresponding route, so that the accuracy of arrival time prediction is improved, the service information is pushed at a proper time, the low power consumption and higher accuracy are simultaneously considered, and the user experience is improved.

Optionally, in order to improve accuracy of the predicted arrival time, the timing of performing the time prediction needs to be determined. Considering that the distance from the access target cell to the arrival target area is relatively fixed and is less likely to be interfered by other factors, when the second cell is determined as the target cell, the time for performing the time prediction is taken as the time for performing the time prediction, wherein the second cell is the target cell, and the method comprises the following steps:

The second cell is a cell in the first cell list;

the first cell list is used for recording cells covering the geographic positions of the users when the target behaviors are successfully generated, and the target behaviors are user operation instructions corresponding to the service information.

Optionally, in order to improve accuracy of predicting the arrival time, considering that different travel habits of the user may affect the arrival time, in order to eliminate the effect, comprehensive consideration needs to be performed on travel information of the user, where the method further includes:

collecting at least one of the following information:

traffic mode information, travel date information and travel time period information under the current travel date;

the acquired information is used for predicting the first time together with the cell information.

Optionally, the first time is predicted by training a time prediction model, and in order to learn the travel habits of the user going to the target area through different routes, respective time prediction models are respectively trained for different routes. In actual output, selecting a best-matching time prediction model to perform time prediction to improve accuracy of predicting arrival time, wherein predicting a first time according to continuously acquired cell information comprises:

Inputting continuously acquired cell information into a first time prediction model, wherein the first time prediction model is used for predicting duration according to the continuously acquired cell information, and taking the duration output by the first time prediction model as first time;

the first time prediction model is a time prediction model corresponding to a second cell list to which a second cell belongs, the cells recorded in the second cell list are cells covering geographic positions when the distance between geographic positions when the user successfully generates the target behavior is smaller than a first distance threshold, and a plurality of second cell lists form the first cell list.

Optionally, in order to ensure that the result of the model prediction does not have excessive errors, before inputting the continuously acquired cell information into the first time prediction model, the method further includes:

determining that the model error of the first temporal prediction model is less than a preset threshold.

Optionally, when there is a large error in the model, in order to prevent erroneous prediction information from degrading the user experience, the time of arrival is not predicted using the time prediction model, which further includes:

and pushing the service information when the model error of the first time prediction model is not smaller than a preset threshold value.

Optionally, to facilitate training of the time prediction model, continuously acquired cell information is input into the first time prediction model, including:

Coding the continuously acquired cell information into a route characteristic sequence according to the sequence of the acquisition time;

the route feature sequence is input into a first temporal prediction model.

Optionally, to perform training of the time prediction model, historical trip data needs to be collected, and a trip habit of the user on each route is learned according to the historical trip data, and the method further includes:

acquiring first historical trip data of a target behavior;

clustering according to the first historical trip data to obtain a position cluster;

and determining a second cell list corresponding to each position cluster according to the first historical trip data corresponding to each position cluster and training a time prediction model corresponding to each position cluster.

Optionally, the first historical trip data includes:

the method comprises the steps of enabling a terminal device to be in a geographic position when the terminal device generates target behaviors, enabling a third cell connected when the terminal device generates the target behaviors, enabling time of the target behaviors to be equal to a first time interval of first connection to the third cell, enabling the terminal device to be sequentially connected with cell information before the target behaviors occur, and enabling the terminal device to be connected with cell information of other cells accessible to the terminal device before the target behaviors occur.

Optionally, in order to determine a common route of the user to the target area for distinguishing, several geographic locations where the target behavior occurs and the distances are similar may be regarded as being reached through the same route, and the location cluster is obtained according to the first historical trip data cluster, including:

And clustering the geographic positions with the distance smaller than a first distance threshold according to the distance between the geographic positions when the target behavior occurs in the first historical trip data to obtain a plurality of position clusters.

Optionally, when training the model is specifically performed, determining a second cell list corresponding to each location cluster according to the first historical trip data corresponding to each location cluster and training a time prediction model corresponding to each location cluster, including:

coding a third cell in the first historical trip data contained in each position cluster, cell information which is connected with the terminal equipment before the target behavior occurs and cell information of other cells which can be accessed by the terminal equipment before the target behavior occurs, so as to obtain characteristic data;

taking a first time interval in the first historical trip data contained in each position cluster as a training label;

dividing first historical trip data containing characteristic data and training labels into training samples and evaluation samples;

and training according to the training sample and the evaluation sample to obtain a time prediction model corresponding to each position cluster.

Optionally, after model training is completed, updating the time prediction model may be further implemented according to the travel data collected later, so as to further improve accuracy of prediction of the time prediction model on the first time, where after service information for the target behavior is pushed after the first time, the method further includes:

When the target behavior is detected, determining the cell information continuously collected by the terminal equipment, the geographic position of the terminal equipment when the target behavior occurs, a third cell connected when the target behavior occurs, the time of the target behavior and a first time interval connected to the third cell for the first time, and the cell information of other cells accessible by the terminal equipment before the target behavior occurs as second historical trip data;

and retraining the existing time prediction model at a preset time according to the second historical trip data.

Optionally, when updating the training time prediction model, the newly collected historical trip data may include a new route to the target area, so updating the location clusters and updating the time prediction model corresponding to each location cluster need to be completed at the same time, and retraining the existing time prediction model according to the second historical trip data includes:

according to the geographic position of the terminal equipment when the target behavior occurs in the second historical trip data, distributing the geographic position of which the distance with other geographic positions when the target behavior occurs is smaller than a first distance threshold value into the existing position cluster;

dividing a new position cluster according to other geographic positions where the terminal equipment is located when the terminal equipment generates the target behavior in the second historical trip data;

Updating the second cell list corresponding to each position cluster, and updating and training the time prediction model corresponding to each second cell list according to the second historical trip data.

Optionally, after updating the time prediction model is completed, error reduction of the model needs to be guaranteed, and accuracy is improved, and the method further includes:

if the model error of the retrained time prediction model is smaller than a preset threshold value, updating the time prediction model;

if the model error of the retrained time prediction model is greater than or equal to a preset threshold value, the time prediction model is kept unchanged.

Optionally, pushing service information after the first time includes:

and displaying the service information on a display interface of the terminal equipment.

In a second aspect, an embodiment of the invention provides an electronic device comprising a memory for storing computer program instructions and a processor for executing the computer program instructions, wherein the computer program instructions, when executed by the processor, trigger the electronic device to perform the method steps of any of the first aspects.

In a third aspect, embodiments of the present invention provide a computer-readable storage medium having a computer program stored therein, which when run on a computer, causes the computer to perform the method as in any of the first aspects.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

As shown in fig. 1A, a schematic diagram of an embodiment of information pushing provided by the present invention;

as shown in fig. 1B, a schematic diagram of an embodiment of information pushing provided by the present invention;

fig. 2 is a schematic diagram of an embodiment of information pushing provided by the present invention;

FIG. 3 is a flowchart of a data acquisition method according to an embodiment of the present invention;

fig. 4 is a flowchart of an information pushing method according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of an embodiment of a cluster of clustered positions according to an embodiment of the present invention;

FIG. 6 is a schematic diagram of a location cluster according to an embodiment of the present invention;

FIG. 7 is a flowchart of a model training method according to an embodiment of the present invention;

Fig. 8 is a flowchart of an information pushing method according to an embodiment of the present invention;

fig. 9 is a flowchart of an information pushing method according to an embodiment of the present invention;

fig. 10 is a schematic structural diagram of a terminal device according to the present invention;

FIG. 11 is a flowchart illustrating a specific embodiment of an information pushing method according to the present invention;

fig. 12 is a schematic hardware structure of an electronic device according to an embodiment of the present invention.

Detailed Description

For a better understanding of the technical solutions of the present application, embodiments of the present application are described in detail below with reference to the accompanying drawings.

It should be understood that the described embodiments are merely some, but not all, of the embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.

Prior to the description of the embodiments of the present application, description will be made first of related arts and technical problems.

The terminal equipment generally has a positioning function, and the terminal equipment can determine the real-time position of the terminal equipment by a positioning technology, so that the position information of a user holding the terminal equipment is determined, and various services are provided for the user holding the terminal equipment. For example, after the terminal device reaches some target areas, the terminal device can push service information related to the target areas through related applications or services for users to operate, so as to reduce operation steps of the users, or remind the users when the users forget to execute related operations.

The positioning function of the terminal device generally comprises GPS positioning, cell positioning and the like. GPS positioning accuracy is high, but frequent use of GPS positioning causes high power consumption for terminal devices. In order to avoid excessive power consumption losses for the terminal device, the location information of the terminal device is typically determined by cell positioning.

As shown in fig. 1A and 1B, taking the example of pushing the card punching service information for the user based on the cell positioning technology, it is assumed that a circular area with a radius of 500 meters is a card punching area with a work unit as a center. After the user holds the terminal equipment and enters the card punching area, the card can be successfully punched on the terminal equipment. In order to facilitate the operation of a user, the terminal equipment provides a card punching service for the user, after the user subscribes the card punching service, the terminal equipment determines whether the terminal equipment enters a card punching area or not through a cell positioning technology, pushes card punching service information after entering the card punching area, and prompts the user to execute card punching. Specifically, a target cell is recorded in the terminal equipment, the target cell is a cell connected with the terminal equipment when the history card punching is successful, and when the terminal equipment is connected to the target cell, the terminal equipment determines to enter a card punching area and pushes card punching service information for a user. The user may link to the punching application through the punching service information and perform punching within the punching application. For example, if the terminal device records the cell 1 to which the user is successfully connected, the terminal device pushes the card-punching service information for the user when the terminal device is connected to the cell 1.

However, the method for pushing the card punching service information for the user based on the cell positioning technology has the problem that the service information pushing time is unreasonable and the service information is pushed in error.

Through analysis, because the cell coverage area is larger, the cell coverage area is not necessarily completely located in the card punching area, and a method for judging whether to enter the card punching area by the terminal equipment through the cell connected with the terminal equipment when the historical card punching is successful is inaccurate. Taking the cell 1 in fig. 1A as an example, assuming that the user has performed a card punching operation when connecting to the cell 1, the cell 1 is recorded in the terminal device as a target cell, however, because the coverage area of the cell 1 is larger, only a part of the area of the cell 1 is located in the card punching area and coincides with the card punching area, so that the terminal device may not actually enter the card punching area when accessing to the cell 1. For example, as shown in fig. 1A, it is assumed that the user holds the terminal device to enter the coverage of the cell 1 from the location B and access the cell 1, and at this time, the terminal device decides to enter the card punching area and pushes the card punching service information. After observing the card punching service information pushed by the terminal equipment, the user can click to enter a card punching interface and execute card punching in the card punching interface. However, at this time, the terminal device is located at the position B and does not enter the card punching area, so that the user cannot successfully punch a card, and the card punching application displays the corresponding prompt information of the card punching failure. Since the terminal device has already pushed the card punching service information to the user, when the user actually enters the card punching area, for example, the position a, the terminal device does not push the card punching service again on the display interface, and the user still needs to open the card punching application by himself to finish card punching.

As can be seen, because the coverage area of the cell is larger, when the user is successful in punching a card, the coverage area of the cell accessed by the terminal device is not completely located in the punching area, and there may be a deviation between the position (for example, position B in fig. 1A) where the terminal device accesses the cell and the position (for example, position a in fig. 1A) where the terminal device enters the punching area, which may cause a deviation between the time when the terminal device determines to enter the punching area and the time when the terminal device actually enters the punching area, so that the timing of pushing the punching service information by the terminal device is unreasonable (for example, the position B in fig. 1A is pushed but the position a is not pushed), and a problem of false triggering of the punching service information occurs.

When service information of other behaviors is pushed based on the cell positioning technology, the problems of unreasonable service information pushing time and wrong pushing similar to the pushing of the card punching service information can also exist.

Based on the above analysis, it can be known that the time error of pushing the service information to the user is an error caused by the time spent by the terminal device accessing the location of the target cell (e.g., location B in fig. 1A) to the location of the target area (e.g., location a in fig. 1A). If the error caused by the time can be eliminated, the service information can be accurately pushed to the user when the user actually enters the target area.

The use of this path is relatively fixed, considering that the path of the user to the target area is relatively fixed, the location in the target area where the target action is performed is relatively fixed, and that the distance from the access cell (e.g., cell 1 in fig. 1A) to the path actually entering the target area (e.g., the punch-down area in fig. 1A) is relatively short in the path to the target area, and that there is a relatively small possibility of other interference in the shorter path. Based on the method, the embodiment of the invention can infer the time consumed by the user in the journey and eliminate errors caused by the time, thereby solving the problems of unreasonable pushing opportunity and wrong pushing of the service information.

In the information pushing method, the time from the terminal equipment accessing the target cell to the target area entering is predicted, and after the time passes, service information aiming at the target behavior is pushed. Therefore, when the service information aiming at the target behavior is pushed, the terminal equipment enters the target area, and the user can successfully execute the target behavior according to the pushed service information.

Taking the example that the target action is card punching, as shown in fig. 1B, when the terminal device accesses the cell 1 at the location B, the terminal device does not push the card punching service information on the interface, but pushes the card punching service information on the interface when the terminal device enters the target area, for example, when the terminal device actually reaches the location a. Because the terminal equipment enters the opening area, the user can successfully realize the card punching in the card punching application at the position A according to the pushed card punching service information.

In one embodiment, in order to improve accuracy of time prediction, time from the terminal device accessing the target cell to entering the target area on different paths can be predicted according to different paths of the user going to the target area, so that service information aiming at target behavior is pushed after the time.

Still taking the example that the target behavior is punch-card. Referring to fig. 2, it is assumed that the user has previously successfully punched a card using the terminal device at location B and location D, and the terminal device records cell 4 covering location B and cell 8 covering location D as target cells.

When the user goes to the card punching area through the path 1, the terminal device may record the cells which are sequentially accessed on the path 1, for example, the cell 1, the cell 2, the cell 3 and the cell 4 in fig. 2, and when the user arrives at the position a, the cell accessed by the terminal device is switched from the cell 3 to the cell 4, and the cell 4 is determined to be the target cell, then the terminal device may predict the time from the position a to the entering of the target area based on the cell 1, the cell 2, the cell 3 and the cell 4 accessed on the path 1, and push the card punching service information after the time.

When the user goes to the card punching area through the path 2, the terminal device may record the cells which are sequentially accessed on the path 2, for example, the cell 5, the cell 6, the cell 7 and the cell 8 in fig. 2, and when the user arrives at the position C, the cell accessed by the terminal device is switched from the cell 7 to the cell 8, and the cell 8 is determined to be the target cell, then the terminal device may predict the time from the position a to the entering of the target area based on the cell 5, the cell 6, the cell 7 and the cell 8 accessed on the path 2, and push the card punching service information after the time.

In order to predict the time from the terminal device accessing the target cell to entering the target area on different paths according to different paths of the user going to the target area, in one embodiment, the terminal device may train different time prediction models based on historical trip data of successfully executing the target behavior, so as to learn the trip rule of the user when going to the target area through the time prediction models, and predict the time from the terminal device accessing the target cell to entering the target area on one path based on the time prediction models obtained by training.

In the embodiment of the application, the terminal equipment can acquire travel data in the travel, and if the terminal equipment generates the target behavior in the travel, the travel data of the target behavior successfully executed at this time is recorded as historical travel data and used for model training; otherwise, the travel data collected in the current journey are deleted.

In this embodiment of the present application, the historical trip data stored after the target behavior is successfully executed may include: the method comprises the steps that cell information connected with the terminal equipment before the occurrence of target behavior, cell information of other cells which can be accessed by the terminal equipment before the occurrence of the target behavior, geographic positions of the terminal equipment when the terminal equipment generates the target behavior, a third cell connected when the terminal equipment generates the target behavior, time for generating the target behavior and a first time interval connected to the third cell for the first time are sequentially carried out.

Alternatively, training and updating the time prediction model based on the historical trip data may be performed periodically, and the specific value of the period is not limited in the embodiment of the present application, and may be, for example, 1 day, 1 week, or the like. When the model training is executed for the first time, executing the model training by using the first historical trip data acquired in a period of time after the function subscription; and when the subsequent model is updated, the travel data of the target behavior is collected from the last model training or updating to the current period of time and is used as second historical travel data, and the time prediction model is updated through the second historical travel data.

The realization of the terminal device in the embodiment of the present application to collect trip data of successful card punching is exemplified by the following working card punching scenario.

Fig. 3 is a flowchart of a data collection method according to an embodiment of the present invention.

Referring to fig. 3, a user subscribes to a card punching service on a terminal device, so that the terminal device can push the card punching service to the user when the terminal device reaches a card punching area later, and prompt the user to punch cards.

The user can subscribe the card punching service in the corresponding APP through Application programs (Application) such as an Application assistant or a voice assistant carried by the terminal equipment. The application assistant is exemplified as an a assistant. The user can subscribe various services provided by the A assistant for the user through a preset operation instruction, such as clicking a corresponding icon, in the interface of the A assistant.

Within days after subscription of the punching service is completed, the arrival time of the user to the punching area is not predicted, the punching service information is pushed by using a method in the related technology, and the travel data of the user in the time is collected as first historical travel data, and the travel rule of the user is learned according to the travel data. Optionally, in order to more closely fit the travel rule of the user for working and punching cards, the terminal device may only collect travel data of the user on a working day. For example, travel data of the user within 10 working days after subscribing to the card punching service may be used as historical travel data for learning a travel rule of the user.

Specifically, the terminal equipment determines that when the user leaves home on the working day, the terminal equipment starts to collect the cell information of the cell accessed by the terminal equipment. In the process of continuously collecting the cell information, the cell information is collected once after each fixed time interval, and the specific value of the fixed time interval is not limited in the embodiment of the present application, and may be, for example, 5 minutes. The acquired cell information includes cells accessed by the terminal equipment and cells which can be accessed but not accessed by the terminal equipment. The fixed time interval preset when acquiring the cell information may be, for example, 5 minutes, which may be set by default by the terminal device or may be set by a user interface provided by the user at the a assistant. The shorter the fixed time interval, the higher the acquisition accuracy is relatively.

Optionally, the terminal device may establish a user portrait according to a usage rule of the terminal device by the user, and determine a cell corresponding to an address where a home of the user is located according to the user portrait, so as to determine whether the user leaves the home on each working day, and start to collect cell information when the user leaves the home.

In the process of continuously collecting the cell information, the terminal equipment also judges whether the current journey is finished to punch a card or not so as to determine whether the collected cell information is effective or not and whether the collected cell information needs to be stored or not. When the card punching is completed, the geographical position of the card punching point is called and recorded, the cell information of the third cell which is accessed by the terminal equipment at present, the time interval from the access of the terminal equipment to the current time of the card punching is completed, and the cell information acquired before the successful card punching, namely the cell which is connected and connectable but not connected before the successful card punching, is used as the historical trip data. If the current trip does not reach the card punching position and does not finish card punching, the cell information acquired in the current trip is not used as data for learning the travel rule of the user, and the cell information acquired in the current trip is deleted.

When the terminal equipment identifies that the successful page of the card punching occurs, the stroke is determined to reach the card punching point and the card punching is completed.

The following describes the implementation process of training a time prediction model according to the collected historical trip data.

When the time prediction model is trained according to collected historical travel data, since more than one route is usually sent to the target area by the user, and travel rules of the user for the target area through different routes are different, errors are likely to occur if a single time prediction model is used for predicting the arrival time of the user on different routes. In order to solve the problems, when the time prediction model is trained, the embodiment of the invention determines different routes of the user to the target area, and performs differentiated learning on the different routes of the user to the target area to obtain a plurality of different time prediction models, and selects the model of the corresponding route for prediction during actual prediction so as to improve the accuracy of model prediction.

Corresponding to the above embodiment, as shown in fig. 4, a flowchart of an information pushing method provided by the embodiment of the present invention is provided, where the method includes specific steps:

s401, clustering the historical trip data to obtain a plurality of position clusters.

Specifically, after the function subscription is completed, first historical trip data of a period of time needs to be collected, and training of the time prediction model is completed based on the first historical trip data. After the time prediction model training is completed and before the next time of executing the time prediction model updating, second historical trip data are required to be collected, and updating of the time prediction model is executed based on the second historical trip data.

The historical trip data comprises the geographic position of the terminal equipment when the target behavior occurs. After the terminal equipment executes the target behavior, the terminal equipment executes accurate positioning through a GPS, beidou and the like so as to determine the position coordinates when the terminal equipment generates the target behavior. In a specific embodiment, when the terminal device completes the card punching in the card punching area, specific position coordinates of the completed card punching point are obtained.

In order to distinguish the routes of the user to the target area, the clustering may be performed according to the distance between the geographical locations where the terminal device is located when the target behavior occurs. In clustering, DBSCAN clustering is generally performed by applying longitude and latitude coordinates of each target site, and the stuck points with the distance smaller than a first distance threshold are clustered into a position cluster. In general, the preset sampling range is 5 meters during clustering, and the minimum sampling number is 3, namely, when the number of target behaviors occurring within the range of radius 5 meters is greater than or equal to three, the at least three position information are clustered into one position cluster.

Although different target positions can be regarded as different routes selected by a user when the user goes to a target area, different time prediction models need to be trained, because the positions of the user, at which the target behaviors occur, through the same route each time cannot be identical, the more approximate geographic positions can be regarded as target behaviors of the user at the same target position through the same route, only errors on a few routes exist, and the actual influence cannot be caused, and therefore, the similar geographic positions can be regarded as the same target position through clustering, and one time prediction model can be trained and used together.

After the clustering is completed, determining a cell set corresponding to the position cluster, and recording the cell set in a second cell list corresponding to the position cluster, wherein the cells are accessed by all the position information of the position cluster, which generates the target behavior, and the cells which are connectable but not accessed when the target behavior occurs. If a cell belongs to two or more position clusters at the same time, the cell is deleted from the position clusters with less occurrence times and is reserved in the position cluster with the most occurrence times. The second cell sets of all the position clusters are formed into a first cell list in a sharing mode, and the first cell list is pre-stored in the terminal equipment. The first cell list is used for determining whether a user enters a target cell, and the second cell list is used for determining a corresponding position cluster and a corresponding time prediction model.

Fig. 5 is a schematic diagram of a specific embodiment of a cluster location cluster according to an embodiment of the present invention, and an example of a shift card punching scenario is still described.

The first historical trip data comprises position information of target behaviors of the terminal equipment in the target area, namely specific positions of the terminal equipment for punching cards in the punching card area. Referring to fig. 5, the terminal device completes 7 times of punching in the punching area, and each time the punching is completed, the terminal device collects the position information of the current punching point and records the position information as the punching point. Clustering is carried out according to the position information of the 7-time punching points, and two position clusters, namely a position cluster 1 and a position cluster 2, are obtained by copolymerization. Wherein, the position cluster 1 comprises 3 punching points, and the position cluster 2 comprises four punching points. The 3 snap points in the position cluster 1 can be regarded as one target position, and the 4 snap points in the position cluster 2 can be regarded as one target position.

In training the temporal prediction model by the subsequent step, two temporal prediction models need to be trained for the position cluster 1 and the position cluster 2, respectively. The training methods of the two time prediction models are the same.

Fig. 6 is a schematic diagram of a location cluster according to an embodiment of the present invention. The position cluster is position cluster 1 as shown in fig. 5.

The position cluster 1 comprises a punched-card point 1, a punched-card point 2 and a punched-card point 3 which are obtained through clustering. And taking the cells accessed by the terminal equipment when the punching points 1, 2 and 3 execute punching and the cells which can be connected but not accessed when the terminal equipment is positioned in the punching points 1, 2 and 3 as a second target cell list of the position cluster 1.

Referring to fig. 6, the punch-card point 1 is located in the coverage area of the cell 1, and the terminal device accesses the cell 1 during punch-card, so that the cell 1 is configured in the cell list of the location cluster 1; the card punching point 2 is positioned in the coverage area of the cell 2, and the terminal equipment is accessed into the cell 2 during card punching, so that the cell 2 is configured in a cell list of the position cluster 1; the point 3 of punching card is located in the coverage of both the cell 2 and the cell 3, and although the terminal device accesses the cell 2 at the time of punching card, the cell 2 and the cell 3 are configured in the second cell list of the location cluster 1 in consideration of the possibility of accessing the cell 3 when the terminal device reaches the point 3 of punching card again later. The second cell list corresponding to the location cluster 1 includes cell 1, cell 2 and cell 3.

And counting the target cell list of each position cluster obtained by clustering to obtain a target cell list corresponding to the whole card punching area. For example, as shown in fig. 5, the second cell list corresponding to the location cluster 2 is cell 4, cell 5, cell 6 and cell 7, and the first cell list corresponding to the punching area is cell 1, cell 2, cell 3, cell 4, cell 5, cell 6 and cell 7.

When the position cluster is updated through the second historical trip data, determining the position information of the target behavior in the second historical trip data, and if the distance between the position information and the geographic position where the target behavior occurs in the existing position cluster is smaller than a first distance threshold, directly dividing the position information into the position cluster; if the location information and the location information belong to any existing location cluster, a new location cluster is generated according to the location information and the location information to update the existing location cluster. After the position cluster is updated, the second cell list corresponding to the position clusters and the first cell list are correspondingly updated.

S402, training a time prediction model corresponding to each position cluster through historical trip data.

Specifically, each history trip data is divided into corresponding position clusters according to the geographic position of the history trip data when the target action occurs. And training a time prediction model corresponding to each position cluster according to the historical trip data divided by each position cluster. The training modes of the time prediction models are the same, namely, feature data are determined from historical trip data, training labels are determined, and model training is performed based on sample data formed by the feature data and the training labels. The collected historical trip data comprises a first cell connected when the target behavior occurs, a first time interval between the time of the target behavior and the time of the first connection to the first cell, cell information which is sequentially connected by the terminal equipment in the first time before the target behavior occurs, and cell information of other cells which can be accessed by the terminal equipment in the first time before the target behavior occurs, besides the geographic position of the terminal equipment when the target behavior occurs.

And when the characteristic data are determined, coding the first cell in the history trip data contained in the current position cluster, cell information which is connected with the terminal equipment in sequence in the first time before the target action occurs, and cell information of other cells which can be accessed by the terminal equipment in the first time before the target action occurs to obtain the characteristic data. Namely, the route of the terminal device is encoded to obtain the characteristic data.

When coding, the cell sequence of the terminal equipment accessing each cell needs to be determined according to the time sequence of collecting the information of each cell, namely, a specific route is determined. Since the terminal device may be covered by a plurality of cells in most cases, and may be connected to other cells when it is next located in the same location, for example, the terminal device is simultaneously covered by cell 1 and cell 2 at the location a, the terminal device is connected to cell 1 in the present trip, and may be connected to cell 2 when it next passes through the location a. Therefore, in order to determine the whole representation method of the redirection line, when determining the cell sequence, the cells accessed by the terminal equipment and the cells which can be accessed but not accessed are arranged and combined, and then are coded.

The importance degree of different road sections on the route is different, for example, when a user just leaves home, the distance from the target area is far, the residual time is long, and even if the travel rule of the road section is learned, the follow-up time cannot be accurately predicted, so that in order to reduce the storage pressure and the calculation pressure of the terminal equipment, the cell information acquired when the user just leaves home can be properly deleted, and only the cell information acquired in the last first time is reserved. In a specific embodiment, the first time may be set to 30 minutes, i.e. the reservation of cell information acquired every 5 minutes, for the past 30 minutes is specified.

When determining the training tag, a first time interval in the historical trip data contained in the current position is used as the training tag, and the time when the target behavior is about to occur and the time interval when the target behavior is connected to the first cell for the first time are used as the training tag.

The feature data and the corresponding training labels are formed into sample data, the sample data are divided into training samples and evaluation samples, model training is carried out through the training samples, and error estimation and updating of the model are carried out through the evaluation samples. After the time prediction model reaches a certain accuracy, the time prediction model may be applied to perform time prediction.

When the model is trained, the time prediction model can return to a time point or a time interval. The model type may be a simple statistical model, a machine learning model, or a deep learning model. The simple statistical model is suitable for the situation that the sample size is less and only lower calculation complexity is allowed; the machine learning model is suitable for the condition that the sample size is moderate and the moderate calculation complexity can be allowed; the deep learning model is suitable for cases where the sample size is high and greater computational complexity may be allowed.

The model training performed by the first historical trip data is the same as the model updating method performed by the second historical trip data. When model training is performed, the error of the model needs to be ensured to be smaller than a certain threshold value, and then the model training can be considered as completion. When the model is updated, the error of the model needs to be ensured to be smaller than a preset threshold value, the updated model is kept, and otherwise, the model is kept unchanged. The error threshold set for model updating should be lower than the error threshold set for model training. Alternatively, it may be determined whether to retain the updated model by whether the error is reduced.

According to the embodiment of the invention, the positions of the target behaviors in the target area are classified, the corresponding time prediction models are trained, and the time spent for the last distance on different routes is respectively predicted, so that the accuracy of predicting the first time is improved.

The model training will be described by way of example with the goal of behavior being punch cards.

Fig. 7 is a flowchart of a model training method according to an embodiment of the present invention. And the corresponding terminal equipment learns the travel rule of the user according to the historical travel data and trains a time prediction model.

Referring to fig. 7, the terminal device sets a daily trigger for training or updating the time prediction model. When model training is not completed, training of a time prediction model is performed through the first historical trip data; and after the training of the time prediction model is completed, updating the time prediction model through the second historical trip data. The terminal device may trigger training or updating of the time prediction model at a preset time, e.g., 3 late night, or an idle period when the terminal device is off-screen during charging.

After training or updating the time prediction model is triggered, clustering is carried out based on the geographical positions of the stuck points in the collected historical trip data, and a plurality of position clusters are obtained. The above clustering method may be, for example, a density-based clustering algorithm (DBSCAN). And distributing the historical trip data to the corresponding position clusters according to the cell information accessed by the geographic position during the card punching, and executing the training of the time prediction model corresponding to each position cluster. The training modes aiming at each time prediction model are the same, and the time prediction models corresponding to different position clusters are used for learning travel rules of users reaching different punching points.

And when the time prediction model is specifically trained, performing feature processing on the historical trip data, namely performing feature encoding on the cell information connected during punching and the cell information acquired within 30 minutes before punching recorded in the historical trip data, and taking the cell information as the feature data of the sample data. And executing label processing, namely taking the time interval between the terminal equipment and the time of completing the card punching from accessing the cell as a training label of sample data. Sample data is divided into training samples and test samples, for example, 80% training samples and 20% test samples, and a time prediction model is trained by the training samples. And evaluating the error of the time prediction model through the test sample, and keeping the time prediction model unchanged when the error of the time prediction model is not reduced; and when the error of the time prediction model is reduced, updating the time prediction model until the error of the time prediction model reaches the requirement.

A method of performing time prediction according to the trained time prediction model and pushing service information will be described,

as shown in fig. 8, a flowchart of an information pushing method provided by an embodiment of the present invention is used for predicting a time to reach a target area, and the method may be applied in various scenarios including a work card-punching scenario, and the specific steps of the method include:

S801, continuously collecting cell information accessed by terminal equipment according to a preset time interval.

Specifically, the terminal equipment collects the cell information of the cell each time, including the cell information of the cell accessed by the terminal equipment when collecting the cell information.

The position of the terminal equipment is determined to move according to whether the cell accessed by the terminal equipment changes or not. The terminal equipment determines the user portrait through the use record of the user, determines the cell corresponding to the address of the user according to the user portrait, and determines that the position of the terminal equipment moves and starts to collect the cell information when the cell connected with the terminal equipment changes with the cell corresponding to the address of the user.

S802, when a cell accessed by the terminal equipment is switched from a first cell to a second cell, and the second cell is a target cell, predicting first time according to continuously acquired cell information.

Specifically, a first cell list is pre-stored in the terminal device, and cell information of a plurality of target cells is recorded in the first cell list, so as to determine whether the cell accessed by the terminal device is the target cell.

And when the cell accessed by the terminal equipment changes from the first cell to the second cell, determining whether the second cell is in a preset first cell list. And when the second cell is a cell in the first cell list, determining the second cell as a target cell. Wherein the first cell list is composed of the second cell list of each location cluster.

And taking the time when the cell currently connected with the terminal equipment is determined to be the target cell as a prediction opportunity, and predicting the first time when the terminal equipment reaches the target area according to the continuously acquired cell information.

When the terminal equipment reaches the first time of the target area according to the continuously collected cell information, a first time prediction model corresponding to the current connection target cell can be determined through the second cell list, and the continuously collected cell information is input into the first time prediction model so as to predict the first time of the terminal equipment reaching the target area.

Optionally, when continuously acquired cell information is input into the first time prediction model, feature encoding needs to be performed on the acquired cell information. Specifically, according to the sequence of the terminal equipment for collecting the cell information, the cells accessed by the continuously collected terminal equipment are encoded into a route characteristic sequence, and the route characteristic sequence is input into a first time prediction model to execute time prediction.

The coding mode includes but is not limited to integer coding, single-hot coding, word vector coding, longitude and latitude approximate vector coding, sequence feature library coding and the like.

In a specific embodiment, the terminal device may encode the route feature sequence 0102030405 before 20 minutes, before 15 minutes, before 10 minutes, before 5 minutes, and the currently connected cells are cell 1, cell 2, cell 3, cell 4, and cell 5, respectively. After the cell information is collected once every update, the route characteristic sequence can be updated. For example, when updating the acquired cell information, it is determined that the terminal device accesses the cell 6 again after connecting to the cell 5, and then the route feature sequence is updated to 010203040506.

S803, service information aiming at target behaviors is pushed after the first time.

Specifically, after the first time, the terminal device determines that the terminal device reaches the target area, and triggers service information corresponding to the target area for operation of a user. For example, when the target area is a punch point, after the first time, determining that the user enters the punch point, pushing punch information to the user for the user to execute punch.

And detecting whether the user can execute corresponding target behaviors according to the pushed service information within a preset time period after the service information is displayed on the display interface. When the corresponding target behavior is detected, the data acquired in the current journey are used as historical trip data for updating a prediction model of the subsequent execution time.

According to the embodiment of the invention, the route of the terminal equipment to the target area is determined by collecting the cell accessed by the terminal equipment, and the time required by the user from entering the target cell to finally reaching the target area is predicted by learning the travel habit and the time rule on the corresponding route, so that the accuracy of arrival time prediction is improved, the service information is ensured to be pushed in proper practice, the low power consumption and the higher accuracy are simultaneously considered, and the user experience is improved.

Optionally, in some embodiments, in order to improve accuracy of time prediction, when performing S802 to predict the first time according to the continuously acquired cell information, other characteristics such as trip mode on the trip are also considered, so as to improve accuracy of the predicted arrival time. So when executing S801 to continuously collect the cell information accessed by the terminal device according to the preset time interval, other relevant travel information needs to be collected.

Specifically, when the terminal equipment collects the cell information, the terminal equipment also collects the travel information of the travel. The travel information includes, but is not limited to, traffic mode information, travel date information, and travel time period information under the current travel date. The travel time information and the travel date information can be obtained by directly retrieving the system information by the terminal equipment. The traffic mode information comprises the behavior walking, riding, driving, taking buses, taking subways and the like, and can be identified by the terminal equipment through sensor data such as an acceleration sensor, a gyroscope sensor and the like. Optionally, the traffic mode information may also be obtained by other APP when the data right of other APP is provided. For example, after the terminal equipment unlocks the shared bicycle through the shared bicycle APP code scanning, determining that the traffic mode of the trip is riding; after the terminal equipment enters a subway station through the subway APP code scanning, the traffic mode of the travel is determined to be subway riding.

In a specific embodiment, the user selects a walking travel mode at 9 am on monday to go to the target area, and the terminal equipment on the road going to the target area is sequentially connected to cell 1, cell 2, cell 3, cell 4 and cell 5, and codes the route to obtain a route feature sequence 0109010102030405.

Optionally, in some embodiments, before performing S802 to predict the first time based on the continuously acquired cell information, an error of the first time prediction model needs to be verified to determine whether the first time prediction model is valid. Only when the first time prediction model is valid will the first time of arrival of the terminal device at the target area be predicted using the first time prediction model.

Specifically, determining a model error of a first time prediction model, when the model error of the first time prediction model is smaller than a preset threshold value, determining that the first time prediction model is effective, and inputting the coded route feature sequence into the first time prediction model to execute time prediction; when the error of the first time prediction model is larger than or equal to a preset threshold value, determining that the first time prediction model fails, namely that the first time predicted by using the first time prediction model has larger error, and determining the first time of the terminal equipment reaching the target area at the current moment without being predicted by the first time prediction model, and immediately pushing service information to the user.

The evaluation may be performed by the mean absolute error MAE of the first temporal prediction model or the root mean square error RMSE when evaluating the model error.

Alternatively, the feature coding may be performed by a plurality of methods when coding the cell to which the terminal device is connected. Specifically, the integer code allocates a single integer ID for each cell information, and has the advantages of simple processing and larger error when considering the cell information as fixed-distance data; the single-hot coding has the advantages of simple processing, and the disadvantage that the coded vector is too sparse, so that the calculation and storage efficiency is not high, and the distance information between the cell information is not utilized; the word vector coding uses the sequence information of the historical cell information, and obtains the low-dimensional word vector of each cell information based on a training mode of a continuous word bag model or a Skip-gram model, so that the method has the advantages of effectively measuring the distance between cells, having lower dimension and having more complex processing; the longitude and latitude approximate vector coding uses operator public data, each cell information is mapped to the longitude and latitude of the base station, and the longitude and latitude are used as the vectors of the cell information, so that the method has the advantages of being capable of effectively measuring the distance between the cell information and being very low in dimension, and the method has the defect of needing to use the operator public data; the sequence feature library code is clustered based on the card punching cell information sequences to form a feature library, the similarity measurement adopts a dynamic time warping (Dynamic Time Warping, DTW) mode, the cell information sequences with the same route but different lengths and different rhythms of a user can be clustered into the same cluster, the feature sequences of each route are attributed to each cluster, and then the clusters are subjected to simple codes such as independent-heat codes, so that the method has the advantages that the longer route feature sequences can be compressed into the feature library code, and the defect that the cell information which does not appear is difficult to encode.

The implementation of the information pushing method according to the embodiment of the present application is described below in conjunction with a specific embodiment.

Fig. 9 is a flowchart of an information pushing method according to an embodiment of the present invention. And the corresponding terminal equipment executes the process of time prediction and pushing the service information.

Referring to fig. 9, a user of a terminal device leaves home on a workday to start to collect cell information of a cell to which the terminal device is connected. In the process of continuously acquiring cell information, the acquisition of cell information is performed every 5 minutes. And determining whether the cell connected with the terminal equipment changes in the process of collecting the cell information, and determining whether the cell newly accessed by the terminal equipment is a target cell when the cell connected with the terminal equipment changes.

When the newly accessed cell of the terminal equipment is the target cell, determining a second cell list affiliated to the target cell, and determining a time prediction model corresponding to the second cell list as a first time prediction model for actually executing prediction. Determining whether the error of the first time prediction model is smaller than a preset threshold value, when the error is smaller than the preset threshold value, encoding the acquired cell information, inputting the encoding result into the first time prediction model, and predicting the time of the terminal equipment reaching the corresponding stuck point. After the predicted time passes, the A assistant pushes the card punching service to the user at the main interface of the terminal equipment for the user to execute card punching. When the error of the first time prediction model is larger than or equal to a preset threshold value, determining that the model is unavailable, and immediately pushing a card punching service to a user through an A assistant for the user to execute card punching.

The card punching service comprises an icon of a third party card punching application as a link, a user can quickly enter a card punching interface of the card punching application through clicking the icon of the card punching application, and preset operation instructions such as card punching and the like are completed in the card punching application.

Fig. 10 is a schematic structural diagram of a terminal device according to the present invention.

Referring to fig. 10, an application layer and an application framework layer are included in the terminal device. The application layer comprises an A assistant, and the application framework layer comprises a processing module, a perception module, a data middle stage and a learning middle stage.

The assistant A is generally an APP carried in the terminal equipment and is used for pushing service information.

The sensing module specifically comprises a geofence sensing module, a data acquisition sensing module and a page sensing module, and is used for sensing and acquiring various data and sending the acquired data to a data center for processing. Specifically, the geofence sensing module is used for determining a cell connected with the terminal equipment, the data acquisition module is used for continuously acquiring cell information, and the page sensing module is used for identifying an interface of the terminal equipment.

The processing module is used for indicating the learning middle station to train the time prediction model according to the data in the data middle station and predicting the first time according to the time prediction model.

Fig. 11 is a flowchart of a specific embodiment of an information pushing method according to the present invention. Specifically, the interaction timing diagram of each functional component of the terminal device is shown in fig. 11.

Referring to fig. 11, after a user subscribes to a card punching service on a corresponding page, the page sensing module identifies an interface with successful subscription, sends a result of successful subscription to the processing module, and when the result reaches a card punching point, the processing module pushes card punching information to the user, and when the card punching is successful, stores the acquired cell information as historical trip data. And subscribes to the out-of-home fence to perceive whether the cell accessed by the terminal equipment changes. After subscribing the fence away from home, the geofence module perceives that the accessed cell and the cell to which the address should be accessed change, so that the user holding the terminal equipment is determined to leave home, the information of the accessed cell is acquired every 5 minutes through the data acquisition perceiving module, and the acquired data is stored in the data center station.

The user subscribes to the punch-through fence, and the processing module triggers execution of time prediction when reaching a target cell connected to the target area. The geofence module performs positioning, determines that a stuck point is reached, namely, a target cell is reached, determines a corresponding position cluster according to the target cell which is accessed currently, and determines a first time prediction model of the position cluster from the learning platform. The learning center determines a model error of the first temporal prediction model to determine whether the first temporal prediction model is valid. When the model fails, a card punching prompt is immediately pushed to a user through a pushing module. When the model is effective, the first time prediction model is used for predicting the card punching time according to the acquired cell information, and a card punching reminder is pushed to a user through the pushing module after the predicted card punching time is passed.

After the user successfully punches a card through the corresponding interface, the page sensing module recognizes the corresponding page, the data acquisition module acquires the cell information connected with the terminal equipment when the card is successfully punched, longitude and latitude coordinates of the terminal equipment when the card is successfully punched, and the time interval from the connection of the terminal equipment to the target cell to the completion of the card punching is calculated. And storing the acquired data into a data center table, and performing data bipartition, wherein the data is divided into a training sample and an evaluation sample.

The learning platform trains and updates the time prediction model. Clustering the stuck points through longitude and latitude coordinates to obtain a plurality of position clusters. And dividing the collected historical trip data into corresponding position clusters, and training and updating the time prediction models.

The embodiment can be generalized to other scenes, the punching area can be generalized to any target area that the user wants to reach, the punching action performed in the punching area can be any target action performed in the target area by the user and related to the target area, and the pushed punching service can be service information corresponding to the target action.

In a specific embodiment, the information pushing method can be applied when the user holds the terminal device and walks to the subway station. The target area can be a subway station which is frequently ridden by a user, the target behavior corresponding to the target area is to show a subway riding code, and the terminal equipment pushes the subway APP when reaching a gate of the subway station so as to help the user to quickly show the subway riding code.

Specifically, after a user subscribes to the riding code service through the a assistant interface, the terminal device can determine a subway station where the user frequently goes to ride through the geographical position acquired when the user passes to ride on a subway, and takes the subway station or a gate in the station as a target place. When a user leaves home, cell information accessed in the travel path of the user and cell information which can be accessed but not accessed are acquired, when the user opens a subway APP through terminal equipment and invokes a subway riding code, the user is determined to have target behaviors, the geographic position of the terminal equipment at the moment is acquired, and a first cell connected at the moment is connected to the first cell for the first time and a first time interval between the first cell and the subway riding code is invoked. The collected data is used as a history for training of the temporal prediction model.

After model training is completed, when the user leaves home, determining a travel route of the user by collecting cell information accessed by the terminal equipment, and predicting the first time from entering a target cell to reaching a subway gate based on a time prediction model. When the first time is reached, subway service is displayed on a display interface of the terminal equipment, and a user can enter a riding code interface of the subway APP by clicking a subway service link to quickly display the riding code.

In another specific embodiment, the information pushing method can be applied when the user holds the terminal device to go to the express post to take the express. The target area may be a courier station. The target behavior corresponding to the target area is to show the express delivery and pickup code, and the terminal equipment pushes the short message APP when reaching the express delivery post station so as to help the user to quickly search the express delivery and pickup code.

Specifically, after a user subscribes to the express delivery and taking piece through the A assistant interface, the terminal device can determine the express delivery post which the user frequently goes to through the geographical position acquired by the user in the past interval, and the express delivery post is taken as a target place. When a user leaves home, cell information accessed in the travel path of the user and cell information which can be accessed but not accessed are acquired, and when the user opens a short message APP through terminal equipment and invokes an express delivery pickup code, the user is determined to have target behaviors, the geographic position of the terminal equipment at the moment is acquired, and a first cell connected at the moment is connected to a first time interval between the first cell and the invoking of the express delivery pickup code for the first time. The collected data is used as a history for training of the temporal prediction model.

After model training is completed, when the user leaves home, the travel route of the user is determined by collecting the cell information accessed by the terminal equipment, and the first time from entering the target cell to reaching the express post is predicted based on the time prediction model. When the first time is reached, express delivery and pickup service is displayed on a display interface of the terminal equipment, and a user can enter a pickup code interface of the short message APP by clicking an express delivery and pickup service link to realize quick inquiry of pickup codes.

Fig. 12 is a schematic hardware structure of an electronic device according to an embodiment of the present invention, where the electronic device may be implemented as a terminal device provided in an embodiment of the present application. As shown in fig. 12, the electronic device 1200 may include a processor 1201, an internal memory 1202, an antenna 1, an antenna 2, a mobile communication module 1203, a wireless communication module 1204, a display 1205, and the like.

It is to be understood that the structure illustrated in the embodiments of the present application does not constitute a specific limitation on the electronic device 1200. In other embodiments of the present application, electronic device 1200 may include more or fewer components than shown, or may combine certain components, or split certain components, or a different arrangement of components. For example, in a computer device, there may be no antenna, no mobile communication module, and no wireless communication module. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

The processor 1201 of the electronic device 1200 may be a device-on-chip SOC, which may include a central processing unit (Central Processing Unit, CPU) therein, and may further include other types of processors. For example, the processor 1201 may include an application processor (application processor, AP) and/or a neural Network Processor (NPU) or the like.

The processor 1201 may include one or more processing units. Wherein the different processing units may be separate components or may be integrated in one or more processors. In some embodiments, the electronic device 1200 may also include one or more processors 1201. The controller can generate operation control signals according to the instruction operation codes and the time sequence signals to finish the control of instruction fetching and instruction execution.

The NPU is a neural-network (NN) computing processor, and can rapidly process input information by referencing a biological neural network structure, for example, referencing a transmission mode between human brain neurons, and can also continuously perform self-learning. Applications such as intelligent cognition of electronic devices can be realized through the NPU, for example: image recognition, face recognition, speech recognition, text understanding, etc.

The internal memory 1202 of the electronic device 1200 can be used to store one or more computer programs, including instructions. The processor 1201 can cause the electronic device 1200 to perform methods provided in some embodiments of the present application, as well as various applications, data processing, and the like, by executing the above-described instructions stored in the internal memory 1202. The internal memory 1202 may include a code storage area and a data storage area. Wherein the code storage area may store an operating system. The data storage area may store data created during use of the electronic device 1200, etc. In addition, the internal memory 1202 may include high-speed random access memory, and may also include non-volatile memory, such as one or more disk storage units, flash memory units, universal flash memory (universal flash storage, UFS), and the like.

The internal memory 1202 may be a read-only memory (ROM), other type of static storage device that can store static information and instructions, a random access memory (random access memory, RAM) or other type of dynamic storage device that can store information and instructions, an electrically erasable programmable read-only memory (electrically erasable programmable read-only memory, EEPROM), a compact disc read-only memory (compact disc read-only memory) or other optical disk storage, optical disk storage (including compact disc, laser disc, optical disc, digital versatile disc, blu-ray disc, etc.), magnetic disk storage media, or other magnetic storage devices, or any computer readable medium that can be utilized to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer.

The processor 1201 and the internal memory 1202 may be combined into one processing device, more commonly separate components, and the processor 1201 is configured to execute program code stored in the internal memory 1202 to implement the methods described in the embodiments of the present application. In particular implementations, the internal memory 1202 may also be integrated into the processor or separate from the processor.

The wireless communication function of the electronic device 1200 can be implemented by the antenna 1, the antenna 2, the mobile communication module 1203, the wireless communication module 1204, a modem processor, a baseband processor, and the like.

The antennas 1 and 2 are used for transmitting and receiving electromagnetic wave signals. Each antenna in the electronic device 1200 may be used to cover a single or multiple communication bands. Different antennas may also be multiplexed to improve the utilization of the antennas. For example: the antenna 1 may be multiplexed into a diversity antenna of a wireless local area network. In other embodiments, the antenna may be used in conjunction with a tuning switch.

The mobile communication module 1203 may provide a solution for wireless communication including 2G/3G/4G/5G etc. applied on the electronic device 1200. The mobile communication module 1203 may include at least one filter, switch, power amplifier, low noise amplifier (low noise amplifier, LNA), etc. The mobile communication module 1203 may receive electromagnetic waves from the antenna 1, perform processing such as filtering and amplifying the received electromagnetic waves, and transmit the processed electromagnetic waves to a modem processor for demodulation. The mobile communication module 1203 may amplify the signal modulated by the modem processor, and convert the signal into electromagnetic waves through the antenna 1 to radiate. In some embodiments, at least some of the functional modules of the mobile communication module 1203 may be provided in the processor 1201.

The modem processor may include a modulator and a demodulator. The modulator is used for modulating the low-frequency baseband signal to be transmitted into a medium-high frequency signal. The demodulator is used for demodulating the received electromagnetic wave signal into a low-frequency baseband signal. The demodulator then transmits the demodulated low frequency baseband signal to the baseband processor for processing. The low frequency baseband signal is processed by the baseband processor and then transferred to the application processor. The application processor displays images or video through a display screen 1205. In some embodiments, the modem processor may be a stand-alone device. In other embodiments, the modem processor may be provided in the same device as the mobile communication module 1203 or other functional module, independent of the processor 1201.

The wireless communication module 1204 may provide solutions for wireless communication including wireless local area networks (wireless local area networks, WLAN) (e.g., wireless fidelity (wireless fidelity, wi-Fi) networks), bluetooth (BT), global navigation satellite systems (global navigation satellite system, GNSS), frequency modulation (frequency modulation, FM), near field wireless communication technology (near field communication, NFC), infrared technology (IR), etc., as applied on the electronic device 1200. The wireless communication module 1204 may be one or more devices that integrate at least one communication processing module. The wireless communication module 1204 receives electromagnetic waves via the antenna 2, modulates the electromagnetic wave signals, filters the electromagnetic wave signals, and transmits the processed signals to the processor 1201. The wireless communication module 1204 may also receive a signal to be transmitted from the processor 1201, frequency modulate it, amplify it, and convert it into electromagnetic waves to radiate through the antenna 2. The communication connection between the anchor device and the sub-device in the embodiment of the present application may be a Wi-Fi network provided by the wireless communication module 1204.

In some embodiments, antenna 1 and mobile communication module 1203 of electronic device 1200 are coupled, and antenna 2 and wireless communication module 1204 are coupled, such that electronic device 1200 may communicate with a network and other devices through wireless communication techniques.

The electronic device 1200 may implement touch input through the display 1205, touch sensor, processor 110, and so on. For example, the touch sensor and the display 1205 are integrated into a touch screen, and a clicking operation of the user on the touch screen is collected as a touch signal by the touch sensor, the touch signal is collected and converted by the sensor module 180 and then transmitted to the processor 110, and the processor 110 analyzes the touch operation behavior of the user through the recognition of the touch signal.

Further, the devices, apparatuses, modules illustrated in the embodiments of the present application may be implemented by a computer chip or entity, or by a product having a certain function.

It will be apparent to those skilled in the art that embodiments of the present application may be provided as a method, apparatus, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media having computer-usable program code embodied therein.

In several embodiments provided herein, any of the functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer-readable storage medium. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the embodiments of the present application.

Specifically, in an embodiment of the present application, there is further provided a computer readable storage medium, where a computer program is stored, when the computer program is executed on a computer, to cause the computer to perform the method provided in the embodiment of the present application.

An embodiment of the present application also provides a computer program product comprising a computer program which, when run on a computer, causes the computer to perform the method provided by the embodiments of the present application.

The description of embodiments herein is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (devices), and computer program products according to embodiments herein. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In the embodiments of the present application, the term "at least one" refers to one or more, and the term "a plurality" refers to two or more. "and/or", describes an association relation of association objects, and indicates that there may be three kinds of relations, for example, a and/or B, and may indicate that a alone exists, a and B together, and B alone exists. Wherein A, B may be singular or plural. The character "/" generally indicates that the context-dependent object is an "or" relationship. "at least one of the following" and the like means any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c may represent: a, b, c, a and b, a and c, b and c or a and b and c, wherein a, b and c can be single or multiple.

In the present embodiments, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.

The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.

All embodiments in the application are described in a progressive manner, and identical and similar parts of all embodiments are mutually referred, so that each embodiment mainly describes differences from other embodiments. In particular, for the device embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and reference is made to the description of the method embodiments in part.

Those of ordinary skill in the art will appreciate that the various elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as a combination of electronic hardware, computer software, and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.

It will be clearly understood by those skilled in the art that, for convenience and brevity of description, specific working procedures of the apparatus, the apparatus and the units described above may refer to corresponding procedures in the foregoing method embodiments, which are not repeated herein.

The foregoing is merely a specific embodiment of the present application, and any person skilled in the art may easily think of changes or substitutions within the technical scope of the present application, and should be covered in the scope of the present application. The protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A method for pushing information, characterized in that the method comprises:

Continuously collect cell information accessed by the terminal device at preset time intervals;

When the cell accessed by the terminal device is switched from the first cell to the second cell, and the second cell is a target cell, predicting a first time according to the continuously collected cell information;

Pushing service information for the target behavior after the first time;

The predicting the first time according to the continuously collected cell information includes:

Inputting the continuously collected cell information into a first time prediction model, the first time prediction model is used to predict the duration according to the continuously collected cell information, and taking the duration output by the first time prediction model as the first time;

The first time prediction model is a time prediction model corresponding to a second cell list to which the second cell belongs, the cells recorded in the second cell list are cells covering the geographical locations where the distance between the geographical locations where the user successfully performs the target behavior is less than a first distance threshold, and a plurality of second cell lists constitute the first cell list;

Among them, obtaining the first historical travel data in which the target behavior occurs;

Clustering the first historical travel data to obtain a location cluster;

A second cell list corresponding to each location cluster is determined according to the first historical travel data corresponding to each location cluster, and a time prediction model corresponding to each location cluster is trained.

2. The method according to claim 1, wherein the second cell is a target cell, comprising:

The second cell is a cell in the first cell list;

The first cell list is used to record cells covering the geographical location where the user is located when the target behavior successfully occurs, and the target behavior is a user operation instruction corresponding to the service information.

3. The method according to claim 1, characterized in that the method further comprises:

Collect at least one of the following information:

Transportation information, travel date information, and travel time period information under the current travel date;

The collected information is used to predict the first time together with the cell information.

4. The method according to claim 1, characterized in that before inputting the continuously collected cell information into the first time prediction model, the method further comprises:

Determine whether a model error of the first time prediction model is less than a preset threshold.

5. The method according to claim 4, further comprising:

When it is determined that the model error of the first time prediction model is not less than a preset threshold, the service information is pushed.

6. The method according to claim 1, characterized in that the step of inputting the continuously collected cell information into a first time prediction model comprises:

According to the order of collection time, the continuously collected cell information is encoded into a route feature sequence;

The route feature sequence is input into the first time prediction model.

7. The method according to claim 1, characterized in that the first historical travel data comprises:

The geographical location of the terminal device when the target behavior occurs, the third cell connected when the target behavior occurs, the first time interval between the time when the target behavior occurs and the first connection to the third cell, the cell information to which the terminal device has been connected in sequence before the target behavior occurs, and the cell information of other cells that the terminal device can access before the target behavior occurs.

8. The method according to claim 1, characterized in that the step of clustering the first historical travel data to obtain a location cluster comprises:

According to the distances between the geographical locations where the target behavior occurred in the first historical travel data, the geographical locations whose distances are less than the first distance threshold are clustered to obtain a plurality of location clusters.

9. The method according to claim 7, characterized in that the determining the second cell list corresponding to each location cluster according to the first historical travel data corresponding to each location cluster and training the time prediction model corresponding to each location cluster comprises:

Encode the third cell in the first historical travel data contained in each of the location clusters, the cell information that the terminal device has sequentially connected to before the target behavior occurs, and the cell information of other cells that the terminal device can access before the target behavior occurs to obtain feature data;

Using the first time interval in the first historical travel data included in each of the location clusters as a training label;

Dividing the first historical travel data including the feature data and the training labels into training samples and evaluation samples;

A time prediction model corresponding to each of the position clusters is obtained by training according to the training samples and the evaluation samples.

10. The method according to claim 1, characterized in that after pushing the service information for the target behavior after the first time, the method further comprises:

When the target behavior is detected, the cell information continuously collected by the terminal device, the geographical location of the terminal device when the target behavior occurs, the third cell connected when the target behavior occurs, the first time interval between the time when the target behavior occurs and the first connection to the third cell, and the cell information of other cells accessible to the terminal device before the target behavior occurs are determined as the second historical travel data;

The existing time prediction model is retrained at a preset time according to the second historical travel data.

11. The method according to claim 10, characterized in that the retraining of the existing time prediction model according to the second historical travel data comprises: updating the time prediction model corresponding to each second cell list, comprising:

According to the geographical location where the terminal device was located when the target behavior occurred in the second historical travel data, geographical locations whose distances from other geographical locations where the terminal device was located when the target behavior occurred are less than a first distance threshold are allocated to the existing location cluster;

Dividing a new location cluster according to other geographical locations where the terminal device was located when the target behavior occurred in the second historical travel data;

The second cell lists corresponding to each of the location clusters are updated, and the time prediction models corresponding to each of the second cell lists are updated and trained according to the second historical travel data.

12. The method according to claim 11, characterized in that the method further comprises:

If the model error of the retrained time prediction model is less than a preset threshold, updating the time prediction model;

If the model error of the retrained time prediction model is greater than or equal to a preset threshold, the time prediction model is kept unchanged.

13. The method according to claim 1, wherein the pushing service information after the first time comprises:

The service information is displayed on a display interface of a terminal device.

14. An electronic device, characterized in that the electronic device comprises a memory for storing computer program instructions and a processor for executing computer program instructions, wherein when the computer program instructions are executed by the processor, the electronic device is triggered to execute the method steps as described in any one of claims 1-13.

15. A computer-readable storage medium, characterized in that a computer program is stored in the computer-readable storage medium, and when the computer-readable storage medium is run on a computer, the computer is enabled to execute the method according to any one of claims 1 to 13.