CN117932561B - Coupling torque data analysis method and system - Google Patents

Abstract

According to the method and the system for analyzing the torque data of the coupler, provided by the application, through the multidimensional example, the network pre-configuration is carried out by designing the indication of the association between the torque description content and the data description content of the data comprising various dimensions, and then the original information association network obtained through pre-configuration is debugged by using the torque data configuration example and the theme information configuration example through intelligent analysis, so that the obtained information association network can be compatible with the single-dimensional torque data and the multidimensional torque data, and therefore, in the association process of the torque data and the time information, the multidimensional data and the event details in the torque data can be obtained more accurately and efficiently, and the accuracy of the association result is improved, namely, the accuracy of the association between the theme information and the torque data is ensured.

Description

Coupling torque data analysis method and system

Technical Field

The present application relates to the technical field of data analysis, and in particular to a coupling torque data analysis method and system.

Background technique

At present, when determining the coupling torque, errors may cause the coupling torque data to be abnormal, which may lead to subsequent operational errors. Therefore, a coupling torque data analysis method is urgently needed to solve the above technical problems.

Summary of the invention

In order to improve the technical problems existing in the related technology, the present application provides a coupling torque data analysis method and system.

In a first aspect, a coupling torque data analysis method is provided, comprising:

Obtaining torque data to be associated and subject information to be associated, wherein the torque data to be associated includes one or more dimensional data;

Obtaining a feature of the subject information to be associated corresponding to the subject information to be associated, and obtaining a feature of the torque data to be associated corresponding to the torque data to be associated;

Combining the characteristics of the torque data to be associated with the characteristics of the subject information to be associated, obtaining an association result through an information association network, wherein the association result is used to characterize the degree of association between the torque data to be associated and the subject information to be associated, wherein the information association network is obtained by debugging a pre-configured original information association network based on each torque data configuration example and a subject information configuration example associated with each torque data configuration example, wherein the torque data configuration example includes one or more dimensional data, and the subject information configuration example is obtained by performing intelligent analysis in combination with the torque data configuration example;

In an independently implemented embodiment, the obtaining a feature of the subject information to be associated corresponding to the subject information to be associated includes:

In combination with the subject information to be associated, obtaining features of the subject information to be associated through the information association network;

The obtaining of the to-be-associated torque data feature corresponding to the to-be-associated torque data includes:

The event features to be associated in the torque data features to be associated are obtained through the information association network, and the torque data features to be associated are obtained from preset long short-term memory units, wherein the torque data features to be associated include at least one of key features to be associated and potential features to be associated.

In an independently implemented embodiment, the obtaining a feature of the torque data to be associated corresponding to the torque data to be associated includes:

If the torque data to be associated is event torque data, the characteristics of the torque data to be associated are obtained through the information association network in combination with the torque data to be associated, and the characteristics of the torque data to be associated are the characteristics of the event to be associated.

In an independently implemented embodiment, the obtaining a feature of the torque data to be associated corresponding to the torque data to be associated includes:

If the torque data to be associated is the first torque data, combining the torque data to be associated, obtaining the event feature to be associated through the information association network;

In combination with the to-be-associated torque data, the to-be-associated key features and the to-be-associated potential features are obtained from the long short-term memory unit;

Wherein, the torque data features to be associated include the event features to be associated, the key features to be associated and the potential features to be associated;

Among them, the long short-term memory unit includes: key features obtained by extracting key features from each first torque data in the first torque database through a key feature extraction network, and potential features obtained by extracting potential features from each first torque data through a potential feature extraction network.

In an independently implemented embodiment, the obtaining a feature of the torque data to be associated corresponding to the torque data to be associated includes:

If the torque data to be associated is attribute torque data, combining the torque data to be associated, and obtaining the event feature to be associated through the information association network;

In combination with the to-be-associated torque data, obtaining the to-be-associated key features from the long short-term memory unit;

Wherein, the torque data features to be associated include the event features to be associated and the key features to be associated;

Wherein, the long short-term memory unit includes: key features obtained by extracting key features of each attribute torque data in the attribute torque database through a key feature extraction network.

In an independently implemented embodiment, obtaining a multi-dimensional example set includes:

Obtaining the torque description content, the first data description content and the second data description content, wherein the correlation degree between the torque description content and the first data description content is greater than a correlation degree target value, and the correlation degree between the torque description content and the second data description content is less than the correlation degree target value;

Performing a hiding process on the torque description content to obtain the torque description content after the hiding process;

Performing a hiding process on the first data description content to obtain the first data description content after the hiding process;

Among them, the multidimensional example set includes at least: a first multidimensional example binary group including the torque description content and the first data description content, a second multidimensional example binary group including the torque description content and the second data description content, a third multidimensional example binary group including the torque description content after hidden processing and the first data description content, and a fourth multidimensional example binary group including the torque description content and the first data description content after hidden processing.

In an independently implemented embodiment, the obtaining of multi-dimensional data features corresponding to the torque description content in each multi-dimensional example binary group, and the subject data features corresponding to the data description content in each multi-dimensional example binary group, includes:

Obtaining a first multi-dimensional data feature corresponding to the torque description content in the first multi-dimensional example binary group, and a first subject data feature corresponding to the first data description content;

Obtaining a first multi-dimensional data feature corresponding to the torque description content in the second multi-dimensional example binary group, and a second subject data feature corresponding to the second data description content;

Obtaining a second multi-dimensional data feature corresponding to the torque description content after the hidden processing in the third multi-dimensional example binary group, and a first theme data feature corresponding to the first data description content;

The first multi-dimensional data feature corresponding to the torque description content in the fourth multi-dimensional example binary group and the third theme data feature corresponding to the first data description content after the hidden processing are obtained.

In an independently implemented embodiment, configuring the original preconfigured network based on the multi-dimensional data features and the subject data features in each multi-dimensional example binary group includes:

During configuration of the original pre-configured network:

Combining the first multi-dimensional data feature and the first subject data feature, obtaining an association result of the first multi-dimensional example binary group, wherein the association result is used to characterize the degree of association between the torque description content and the first data description content;

The network coefficients of the original preconfigured network are optimized in combination with the association results of the first multi-dimensional example bigrams.

In an independently implemented embodiment, the multi-dimensional example binary group specifically includes first content information and first description content, the first content information includes subject data, twist data and key data; the multi-dimensional data features specifically include event features, key features and potential features.

In an independently implemented embodiment, the original preconfigured network is a configured event dimension network, which is obtained after configuration based on event content information and description content, and the event dimension network is used to calculate the possibility that the event content information and description content belong to the same event torque data.

In an independently implemented embodiment, the information association network is obtained in the following manner:

Obtaining each torque data configuration example and a subject information configuration example associated with each torque data configuration example;

Based on each of the torque data configuration examples, a torque data feature corresponding to each of the torque data configuration examples is obtained, and based on each of the subject information configuration examples, a subject information feature corresponding to each of the subject information configuration examples is obtained;

Based on each of the torque data features and each of the subject information features, the regression analysis correlation degree between each of the torque data features and each of the subject information features is obtained, and based on each of the actual correlation degrees and each of the regression analysis correlation degrees, the original information correlation network is debugged to obtain the information correlation network.

In an independently implemented embodiment, the torque data configuration example is a first torque data configuration example, the first torque data configuration example is composed of subject data, torsion data and key data, and the torsion data is composed of a plurality of data nodes;

The step of obtaining torque data features corresponding to each of the torque data configuration examples based on each of the torque data configuration examples includes:

Performing event division on the subject data in each of the first torque data configuration examples to obtain an event queue corresponding to the subject data in each of the first torque data configuration examples, and generating, based on each of the event queues: event features corresponding to the subject data in each of the first torque data configuration examples, wherein the event queue includes a plurality of event analysis attributes;

Analyzing and processing the torsion data in each of the first torque data configuration examples to obtain key features corresponding to the torsion data in each of the first torque data configuration examples;

The key data in each of the first torque data configuration examples are parsed and processed to obtain potential features corresponding to the key data in each of the first torque data configuration examples; wherein the torque data features include the event features, the key features and the potential features.

In an independently implemented embodiment, the original information association network is obtained in the following manner:

Obtaining a multi-dimensional example set, wherein the multi-dimensional example set includes a plurality of multi-dimensional example 2-tuples, wherein the multi-dimensional example 2-tuples include torque description content and data description content, wherein the torque description content includes data of multiple dimensions;

Obtaining multi-dimensional data features corresponding to the torque description content in each of the multi-dimensional example binary groups, and subject data features corresponding to the data description content in each of the multi-dimensional example binary groups, wherein the multi-dimensional data features include multiple dimensional features;

Based on the multi-dimensional data features and the subject data features in each of the multi-dimensional example dyads, the original pre-configured network is configured to obtain a target pre-configured network, and the target pre-configured network is determined as the original information association network.

In a second aspect, a coupling torque data analysis system is provided, comprising a processor and a memory communicating with each other, wherein the processor is used to read a computer program from the memory and execute the computer program to implement the above method.

A coupling torque data analysis method and system provided in an embodiment of the present application obtains torque data to be associated and subject information to be associated, wherein the torque data to be associated includes one or more dimensional data, and then obtains subject information features to be associated corresponding to the subject information to be associated, and obtains torque data features to be associated corresponding to the torque data to be associated, and obtains association results through an information association network based on the torque data features to be associated and the subject information features to be associated, and the association results are used to characterize the degree of association between the torque data to be associated and the subject information to be associated, and the information association network is obtained after debugging the pre-configured original information association network based on each torque data configuration example and the subject information configuration example associated with each torque data configuration example, the torque data configuration example includes one or more dimensional data, and the subject information configuration example is obtained based on intelligent analysis of the torque data configuration example. Thus, a multidimensional example set is obtained, the multidimensional example set includes several multidimensional example tuples, the multidimensional example tuples include torque description content and data description content, the torque description content includes multi-dimensional data, and the multidimensional data features corresponding to the torque description content in each multidimensional example tuple, and the subject data features corresponding to the data description content in each multidimensional example tuple are obtained, the multidimensional data features include multi-dimensional features, and then based on the multidimensional data features and the subject data features in each multidimensional example tuple, the original preconfigured network is configured to obtain the target preconfigured network, and the target preconfigured network is determined as the original information association network. Through multi-dimensional examples, the torque description content including multi-dimensional data and the indication of associating the data description content are designed for network pre-configuration, and then the torque data configuration example and the subject information configuration example through intelligent analysis are used to debug the original information association network obtained by pre-configuration. The information association network thus obtained can be compatible with single-dimensional torque data and multi-dimensional torque data, so that in the process of associating torque data with time information, multi-dimensional data and event details in the torque data can be obtained more accurately and efficiently, thereby improving the accuracy of the association results, that is, ensuring the accuracy of the association between the subject information and the torque data.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

FIG1 is a flow chart of a coupling torque data analysis method provided in an embodiment of the present application.

Detailed ways

In order to better understand the above technical scheme, the technical scheme of the present application is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical scheme of the present application, rather than limitations on the technical scheme of the present application. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

Please refer to FIG. 1 , which shows a coupling torque data analysis method, which may include the technical solutions described in the following steps 302 to 306 .

Step 302: Obtain torque data to be associated and subject information to be associated, where the torque data to be associated includes one or more dimensional data.

The torque data includes one or more dimensional data, that is, the torque data can be, but is not limited to, event torque data, first torque data, and attribute torque data. If the torque data is event torque data, the torque data includes subject data. Similarly, if the torque data is first torque data, the torque data includes subject data, torsion data, and key data. And if the torque data is attribute torque data, the torque data includes subject data and torsion data.

Step 304 , obtaining the to-be-associated subject information feature corresponding to the to-be-associated subject information, and obtaining the to-be-associated torque data feature corresponding to the to-be-associated torque data.

Specifically, the terminal performs event feature extraction processing on the to-be-associated element information in the to-be-associated subject information to obtain to-be-associated subject information features corresponding to each to-be-associated subject information, and obtains to-be-associated torque data features corresponding to the to-be-associated torque data.

Among them, since the element information to be associated is specifically event description information, the corresponding subject information characteristics to be associated are specifically event characteristics. Secondly, since the torque data can be any one of event torque data, first torque data and attribute torque data, or multi-dimensional torque data including subject data, first data and key data and other presentation forms, the torque data characteristics to be associated corresponding to the torque data to be associated can also be event characteristics, key characteristics, potential characteristics, or multi-dimensional characteristics including event characteristics, key characteristics and potential characteristics. The characteristics of the torque data to be associated need to be determined based on the torque data to be associated, so it is not limited here.

Step 306, based on the characteristics of the torque data to be associated and the characteristics of the subject information to be associated, an association result is obtained through an information association network, and the association result is used to characterize the degree of association between the torque data to be associated and the subject information to be associated. The information association network is based on each torque data configuration example and the subject information configuration example associated with each torque data configuration example, and is obtained after debugging the pre-configured original information association network. The torque data configuration example includes one or more dimensional data, and the subject information configuration example is obtained based on intelligent analysis of the torque data configuration example.

The above-mentioned correlation degree is specifically the correlation between the torque data to be correlated and the subject information to be correlated. The correlation result is used to characterize the correlation degree between the torque data to be correlated and the subject information to be correlated, that is, the correlation result can directly describe whether the torque data to be correlated is correlated with the subject information to be correlated, or not.

Secondly, the information association network is obtained by debugging the original information association network obtained by pre-configuration based on each torque data configuration example and the subject information configuration example associated with each torque data configuration example. The torque data configuration example includes one or more dimensional data, and the subject information configuration example is obtained by intelligent analysis based on the torque data configuration example. And the information association network is used to calculate the degree of association between the torque data and the subject information.

In order to make the obtained information association network compatible with multi-dimensional torque data and achieve a better configuration effect with as little intelligent analysis data as possible, before configuring the information association network, a network pre-configuration process based on multi-dimensional torque data is performed, and the pre-configured network is determined as the original information association network. Therefore, based on each torque data configuration example and the subject information configuration example associated with each torque data configuration example, the information association network can be obtained after debugging the pre-configured original information association network.

Among them, the method of obtaining the original information association network includes: obtaining a multi-dimensional example set, the multi-dimensional example set includes a number of multi-dimensional example tuples, the multi-dimensional example tuples include torque description content and data description content, and the torque description content includes multiple dimensional data.

Based on this, the multidimensional data features corresponding to the torque description content in each multidimensional example binary group and the subject data features corresponding to the data description content in each multidimensional example binary group are further obtained, and the multidimensional data features include multiple dimensional features. Specifically, when the torque description content is the first content information, the multidimensional data features include event features, key features and potential features. Similarly, when the torque description content is attribute content information, the multidimensional data features include event features and key features. Then, based on the multidimensional data features and subject data features in each multidimensional example binary group, the original preconfigured network is configured to obtain the target preconfigured network, and the target preconfigured network is determined as the original information association network. Among them, the original information association network is used to calculate the possibility that the torque description content and the data description content come from the same torque data.

Specifically, the terminal obtains the association result through the information association network, obtains the association result based on the regression analysis of the characteristics of the torque data to be associated and the characteristics of the subject information to be associated, and then determines the degree of association between the torque data to be associated and the subject information to be associated based on the association result. Based on this, the degree of association between the torque data to be associated and the subject information to be associated can determine whether the torque data to be associated and the subject information to be associated are associated, so that in actual applications, all torque data in the torque database are associated with the subject information to be associated one by one, so that different association results can be obtained, and then response operations are performed based on specific scenario requirements.

In the above-mentioned coupling torque data analysis method, through multi-dimensional examples, torque description contents including multi-dimensional data and instructions for associating data description contents are designed for network pre-configuration, and then the torque data configuration examples and subject information configuration examples obtained through intelligent analysis are used to debug the original information association network obtained by pre-configuration. The information association network thus obtained is compatible with single-dimensional torque data and multi-dimensional torque data, so that in the process of associating torque data with time information, multi-dimensional data and event details in the torque data can be obtained more accurately and efficiently, thereby improving the accuracy of the association results, that is, ensuring the accuracy of the association between subject information and torque data.

In one embodiment, obtaining the to-be-associated subject information feature corresponding to the to-be-associated subject information includes:

Step 502: Based on the subject information to be associated, characteristics of the subject information to be associated are obtained through the information association network.

Specifically, the terminal determines the subject information to be associated as the input of the information association network, and the derivative layer in the information association network extracts event features from the element information to be associated in the subject information to be associated to obtain the features of the subject information to be associated. Secondly, since the subject information to be associated is specifically event description information, the derivative layer in the information association network specifically divides the subject information to be associated into events to obtain a number of event analyses (attributes), and then obtains the event queue corresponding to the subject information to be associated based on the several attributes, and generates the features of the subject information to be associated corresponding to the subject information to be associated based on each event queue, and the features of the subject information to be associated are specifically event features.

Obtaining a to-be-associated torque data feature corresponding to the to-be-associated torque data, including:

Step 504, obtaining the event features to be associated in the torque data features to be associated through the information association network, and obtaining the torque data features to be associated from the preset long short-term memory unit, wherein the torque data features to be associated include at least one of the key features to be associated and the potential features to be associated.

Among them, the torque data to be associated includes data of multiple dimensions. Specifically, the terminal determines the torque data to be associated as the input of the information association network, and the derivative layer in the information association network extracts event features from the torque data to be associated to obtain the event features to be associated. Secondly, since in actual applications, the correlation calculation (i.e., correlation degree calculation) between the same torque data and different subject information may be involved, repeatedly extracting the torque data features of the same torque data will increase the association time and resource usage. Therefore, when the torque data is the first torque data or the attribute torque data, it involves downloading the first torque data and the attribute torque data, and the processing of the torsion data and the key data requires more feature extraction time.

Based on this, before associating the torque data with the subject information, each first torque data is obtained in advance, the key features of the first torque data are obtained using the key feature extraction network, and the potential features of the first torque data are obtained using the potential feature extraction network, and the key features and potential features of each first torque data are written into a preset long short-term memory unit for caching. Similarly, before associating the torque data with the subject information, each attribute torque data can also be obtained in advance, the key features of the attribute torque data are obtained using the key feature extraction network, and the key features of each attribute torque data are written into a preset long short-term memory unit for caching.

Furthermore, when the user needs to associate information, an information association request is initiated, and the information association request indicates that the torque data is associated with the subject information to be associated. After the terminal obtains the torque data to be associated and the subject information to be associated, it is necessary to further determine the specific data type of the torque data to be associated. If it is the first torque data or the attribute torque data, the key features to be associated, or the key features to be associated and the potential features to be associated are obtained from the preset long-term and short-term memory units, and the event features to be associated in the torque data features to be associated are obtained through the information association network, thereby completing the acquisition of the torque data features.

In actual applications, in order to facilitate the search for corresponding torque data features based on torque data, each torque data in this embodiment carries a torque data identifier. Therefore, when the key features, or key features and potential features corresponding to the torque data are written into the long-term and short-term memory unit for storage, the key features, or key features and potential features will all carry the torque data identifier corresponding to the torque data.

In the present embodiment, since the torque data features corresponding to the same torque data are usually unchanged, the extraction of event features is decoupled from the acquisition of key features to be associated and potential features to be associated. In a scenario where the torque data to be associated includes data of multiple dimensions, event features to be associated can be obtained through an information association network. However, there is no need to obtain key features to be associated and potential features to be associated through the information association network. Instead, at least one of the key features to be associated and potential features to be associated is obtained directly from preset long and short-term memory units. While ensuring the accuracy of the key features to be associated and the potential features to be associated, the time spent on processing the torsion data and key data to obtain key features and potential features is reduced, that is, the time spent on feature extraction of the key features to be associated and the potential features to be associated is reduced, thereby improving the efficiency of associating the torque data with the subject information while ensuring the reliability of the association between the torque data and the subject information.

In one embodiment, obtaining a to-be-associated torque data feature corresponding to the to-be-associated torque data includes:

Step 902 : If the torque data to be associated is event torque data, based on the torque data to be associated, a feature of the torque data to be associated is obtained through an information association network, and the feature of the torque data to be associated is a feature of the event to be associated.

In the above embodiments, the torque data to be associated may be event torque data, first torque data or attribute torque data. In this embodiment, when the torque data to be associated is event torque data, how to obtain the torque data features to be associated corresponding to the torque data to be associated is first described.

Specifically, after the terminal obtains the torque data to be associated, it is necessary to further determine the data type of the torque data to be associated. If the torque data to be associated is event torque data, the torque data to be associated is determined as the input of the information association network, and the derivative layer of the information association network processes the torque data to be associated to obtain the characteristics of the torque data to be associated, and the aforementioned characteristics of the torque data to be associated are the characteristics of the event to be associated.

Secondly, the derivative layer in the information association network specifically divides the torque data to be associated into events to obtain a number of event analyses (attributes), and then obtains the event queues corresponding to the torque data to be associated based on the several attributes, and generates the torque data features corresponding to the torque data to be associated based on each event queue.

In this embodiment, the torque data to be associated is specifically divided into events through the derivative layer in the information association network to obtain an event queue corresponding to the torque data to be associated. The characteristics of the torque data to be associated thus obtained can accurately reflect the position of each attribute in the torque data to be associated, and the event characteristics corresponding to each attribute itself, thereby improving the reliability and characteristic integrity of the obtained characteristics of the torque data to be associated, so as to improve the reliability and accuracy of information skin association.

In one embodiment, obtaining a to-be-associated torque data feature corresponding to the to-be-associated torque data includes:

Step 1102: If the torque data to be associated is the first torque data, based on the torque data to be associated, a feature of the event to be associated is obtained through an information association network.

If the torque data to be associated is the first torque data, the obtained features of the torque data to be associated include event features to be associated, key features to be associated, and potential features to be associated.

Specifically, after the terminal obtains the torque data to be associated, it is necessary to further determine the data type of the torque data to be associated. If the torque data to be associated is the first torque data, the torque data to be associated is determined as the input of the information association network, and the derivative layer of the information association network processes the torque data to be associated to obtain the characteristics of the torque data to be associated, and the aforementioned characteristics of the torque data to be associated are the characteristics of the events to be associated. Secondly, the derivative layer in the information association network specifically divides the torque data to be associated into events to obtain a number of attributes, and then obtains the event queues corresponding to the torque data to be associated based on the several attributes, and generates the event characteristics to be associated corresponding to the torque data to be associated based on each event queue. The specific implementation method is similar to the aforementioned embodiment and will not be repeated here.

Step 1104 , based on the torque data to be associated, obtain the key features to be associated and the potential features to be associated from the long short-term memory unit.

The long short-term memory unit includes: key features obtained by extracting key features from each first torque data in the first torque database through a key feature extraction network, and potential features obtained by extracting potential features from each first torque data through a potential feature extraction network.

Secondly, in order to facilitate the search for corresponding torque data features based on torque data, each first torque data in this embodiment carries a torque data identifier. Therefore, when the key features and potential features corresponding to the first torque data are written into the long short-term memory unit for storage, the key features and potential features will carry the torque data identifier corresponding to the first torque data.

Specifically, after the terminal determines that the torque data to be associated is the first torque data, and determines the torque data to be associated as the input of the information association network, it further obtains the key features and potential features corresponding to the torque data identifier from the long short-term memory unit based on the torque data identifier corresponding to the torque data to be associated, and determines the key features and potential features corresponding to the torque data identifier as the key features to be associated and the potential features to be associated.

Therefore, through steps 1102 and 1104 , the torque data features to be associated including the event features to be associated, the key features to be associated and the potential features to be associated can be obtained.

In this embodiment, the extraction of event features is decoupled from the acquisition of key features and potential features. While ensuring the accuracy of torque data features, it can reduce the time spent on feature extraction of associated key features and associated potential features in actual applications, thereby improving the efficiency of the association between torque data and subject information while ensuring the reliability of the association between torque data and subject information.

In one embodiment, obtaining a to-be-associated torque data feature corresponding to the to-be-associated torque data includes:

Step 1202: If the torque data to be associated is attribute torque data, based on the torque data to be associated, the feature of the event to be associated is obtained through the information association network.

If the torque data to be associated is attribute torque data, the obtained features of the torque data to be associated include event features to be associated and key features to be associated.

Specifically, the torque data to be associated is determined as the input of the information association network, and the derivative layer of the information association network processes the torque data to be associated to obtain the characteristics of the torque data to be associated, and the aforementioned characteristics of the torque data to be associated are the characteristics of the events to be associated. Secondly, the derivative layer in the information association network specifically divides the torque data to be associated into events to obtain a number of attributes, and then obtains the event queues corresponding to the torque data to be associated based on the several attributes, and generates the event characteristics to be associated corresponding to the torque data to be associated based on each event queue.

Step 1204 , based on the torque data to be associated, obtain the key features to be associated from the long short-term memory unit.

The long short-term memory unit includes: key features obtained by extracting key features of each attribute torque data in the attribute torque database through a key feature extraction network.

Secondly, in order to facilitate the search for corresponding torque data features based on torque data, each attribute torque data in this embodiment carries a torque data identifier. Therefore, when the key features corresponding to the attribute torque data are written into the long short-term memory unit for storage, the key features will carry the torque data identifier corresponding to the torque data.

Specifically, after the terminal determines that the torque data to be associated is attribute torque data, and determines the torque data to be associated as the input of the information association network, it further obtains the key features corresponding to the torque data identifier from the long short-term memory unit based on the torque data identifier corresponding to the torque data to be associated, and determines the key features corresponding to the torque data identifier as the key features to be associated.

Therefore, through steps 1202 and 1204 , the torque data features to be associated including the event features to be associated and the key features to be associated can be obtained.

In this embodiment, the extraction of event features is decoupled from the acquisition of key features. While ensuring the accuracy of torque data features, it is possible to reduce the time spent on feature extraction of associated key features in actual applications, thereby further improving the efficiency of the association between torque data and subject information while ensuring the reliability of the association between torque data and subject information.

In one embodiment, in the process of obtaining the original information association network, each multi-dimensional example binary is obtained as follows: obtaining a multi-dimensional example set includes:

Step 1302, obtaining torque description content, first data description content and second data description content, the correlation degree between the torque description content and the first data description content is greater than the correlation target value, and the correlation degree between the torque description content and the second data description content is less than the correlation target value.

Since most of the data content has a strong correlation with the data title, based on this, when obtaining the multi-dimensional example set, the torque description content, the first data description content with a strong correlation with the torque description content, and the first data description content with a weak correlation with the torque description content are first obtained. The aforementioned subject data is specifically the data title corresponding to the torque description content.

Step 1304: hide the torque description content to obtain the torque description content after hiding.

Step 1306: perform hiding processing on the first data description content to obtain the first data description content after hiding processing.

It can be seen that the multidimensional example set includes at least: a first multidimensional example tuple including torque description content and first data description content, a second multidimensional example tuple including torque description content and second data description content, a third multidimensional example tuple including torque description content and first data description content after hidden processing, and a fourth multidimensional example tuple including torque description content and first data description content after hidden processing.

Based on this, the multidimensional example set specifically includes several positive and negative multidimensional examples. After completing the construction of the multidimensional examples, an indication of the association between the torque description content and the data description content is designed to perform network preconfiguration. The network preconfiguration indication is to input the torque description content and the data description content, use the network to determine whether the torque description content and the data description content come from the same torque data, and perform reverse gradient propagation to correct the network coefficients of the original preconfigured network.

In this embodiment, the network is pre-configured by designing an indication of the association between the torque description content and the data description content, so that the information association network can learn the multi-dimensional feature information of the pre-configured torque description content during the association process, further improve the extraction of multi-dimensional details such as multi-dimensional data and events, and further improve the accuracy of the association results.

In one embodiment, in the process of obtaining the original information association network, the multi-dimensional data features and the subject data features are obtained as follows: the multi-dimensional data features corresponding to the torque description content in each multi-dimensional example binary group and the subject data features corresponding to the data description content in each multi-dimensional example binary group are obtained, including:

Step 1402: Obtain a first multi-dimensional data feature corresponding to the torque description content in the first multi-dimensional example binary group, and a first theme data feature corresponding to the first data description content.

The torque description content includes multiple dimensional data, and when the torque description content is specifically the first content information, the data dimensions involved are the most, that is, when the torque description content is the first content information, it specifically includes subject data, torsion data and key data, and the torsion data is composed of several data nodes. Based on this, when obtaining the multi-dimensional data features and subject data features in each multi-dimensional example binary, the torque description content in the first multi-dimensional example binary is input into the derivative layer in the original pre-configured network, and the subject data, torsion data and key data are derived through the derivative layer to obtain the first multi-dimensional data features corresponding to the torque description content.

It should be understood that in actual applications, if the torque description content after hidden processing does not include any of the subject data, torsion data and key data, the derivative layer in the original pre-configured network can set the feature part corresponding to the data as a placeholder special symbol. And if the actual application also includes other types of torque data dimensions, the torque data dimension can be directly added after the current feature input, and the corresponding feature processing flow and feature part position can be set.

Secondly, the first data description content in the first multi-dimensional example tuple is determined as the input of the original information association network, and the derivative layer in the original information association network performs event division on the first data description content to obtain a number of attributes of the first data description content, and then an event queue corresponding to the first data description content is obtained based on the several attributes of the first data description content, and based on the event queue corresponding to the first data description content, a first topic data feature corresponding to the first data description content is generated.

Based on this, the first multi-dimensional data features corresponding to the torque description content in the first multi-dimensional example binary group and the first theme data features corresponding to the first data description content can be obtained.

Step 1404: Obtain the first multi-dimensional data features corresponding to the torque description content in the second multi-dimensional example binary group and the second subject data features corresponding to the second data description content.

Similar to step 1402, when the torque description content is the first content information, it specifically includes subject data, torsion data and key data, and the torsion data is composed of several data nodes. Based on this, the torque description content (i.e., the first content information) in the second multi-dimensional example binary is input into the derivative layer in the original pre-configured network, and the subject data, torsion data and key data are derived through the derivative layer to obtain the first multi-dimensional data feature corresponding to the torque description content. Secondly, the second data description content in the second multi-dimensional example binary is determined as the input of the original information association network, and the derivative layer in the original information association network performs event division on the second data description content to obtain several attributes of the second data description content, and then obtains the event queue corresponding to the second data description content based on the several attributes of the second data description content, and based on the event queue corresponding to the second data description content, generates the second subject data feature corresponding to the second data description content.

Based on this, the first multi-dimensional data features corresponding to the torque description content in the second multi-dimensional example binary group and the second subject data features corresponding to the second data description content can be obtained.

Step 1406, obtaining the second multi-dimensional data features corresponding to the torque description content after the hidden processing in the third multi-dimensional example binary group, and the first theme data features corresponding to the first data description content.

Similar to step 1402, when the torque description content is the first content information, it specifically includes subject data, torsion data and key data, and the torsion data is composed of several data nodes. Based on this, the torque description content (i.e., the first content information) after hidden processing in the third multi-dimensional example binary is input into the derivative layer in the original pre-configured network, and the subject data, torsion data and key data are derived through the derivative layer to obtain the second multi-dimensional data features corresponding to the torque description content after hidden processing. It should be understood that since any one of the subject data, torsion data and key data may be hidden during the hidden processing process, the second multi-dimensional data features include at least one of the event features, key features and potential features.

Secondly, the first data description content in the third multi-dimensional example tuple is determined as the input of the original information association network, and the derivative layer in the original information association network performs event division on the first data description content to obtain several attributes of the first data description content, and then obtains the event queue corresponding to the first data description content based on the several attributes of the first data description content, and based on the event queue corresponding to the first data description content, generates a first topic data feature corresponding to the first data description content.

Based on this, the second multi-dimensional data features corresponding to the torque description content after hidden processing in the third multi-dimensional example binary group and the first theme data features corresponding to the first data description content can be obtained.

Step 1408: Obtain the first multi-dimensional data features corresponding to the torque description content in the fourth multi-dimensional example binary group, and the third theme data features corresponding to the first data description content after the hidden processing.

Similar to step 1402, when the torque description content is the first content information, it specifically includes subject data, torsion data and key data, and the torsion data is composed of several data nodes. Based on this, the torque description content (i.e., the first content information) in the fourth multi-dimensional example binary is input into the derivative layer in the original pre-configured network, and the subject data, torsion data and key data are derived through the derivative layer to obtain the first multi-dimensional data feature corresponding to the torque description content. Secondly, the first data description content after hidden processing in the fourth multi-dimensional example binary is determined as the input of the original information association network, and the derivative layer in the original information association network performs event division on the first data description content after hidden processing to obtain several attributes of the first data description content information after hidden processing, and then obtains the event queue corresponding to the first data description content after hidden processing based on the several attributes of the first data description content after hidden processing, and generates the third subject data feature corresponding to the first data description content after hidden processing based on the event queue corresponding to the first data description content after hidden processing.

Based on this, the first multi-dimensional data features corresponding to the torque description content in the fourth multi-dimensional example binary group and the third theme data features corresponding to the first data description content after hidden processing can be obtained.

In this embodiment, by obtaining the multi-dimensional data features of the torque description content and the torque description content after the hidden processing, the feature extraction capability of the positive and negative examples in the pre-configuration process is improved, thereby ensuring that the features of various dimensions can be obtained more accurately in the subsequent network configuration. Secondly, the extraction of the features of this article can also further ensure the acquisition of event features in the feature decoupling process in practical applications, thereby improving the feasibility of this solution and the reliability of information association.

In one embodiment, the original pre-configured network is configured based on the multi-dimensional data features and the subject data features in each multi-dimensional example binary, including:

During configuration of the original pre-configured network:

Step 1602: Based on the first multi-dimensional data feature and the first subject data feature, obtain the association result of the first multi-dimensional example binary group, where the association result is used to characterize the degree of association between the torque description content and the first data description content.

Specifically, the first multi-dimensional data feature and the first subject data feature are input to the original pre-configured network, and the original pre-configured network outputs the association result of the first multi-dimensional example binary group, and the association result of the first multi-dimensional example binary group is used to characterize the degree of association between the torque description content and the first data description content, that is, the association result of the first multi-dimensional example binary group can be used to determine whether the torque description content and the first data description content come from the same first. If the association result of the first multi-dimensional example binary group tends to the real association result, that is, the association result of the first multi-dimensional example binary group should describe that the degree of association between the torque description content and the first data description content is high.

Step 1604 : Optimize the network coefficients of the original pre-configured network based on the association result of the first multi-dimensional example binary group.

Specifically, the association result of the first multi-dimensional example binary group is compared with the actual association degree between the first torque data and the first data description content, so as to perform reverse gradient propagation to correct the network coefficients of the original pre-configured network.

It should be understood that step 1602 and step 1604 only describe the process of adjusting the network coefficient of the original preconfigured network once during the configuration process of the original preconfigured network. In actual applications, the configuration process of the original preconfigured network at least includes: inputting the first multidimensional data features and the second subject data features in the second multidimensional example binary group into the original preconfigured network, and the original preconfigured network outputs the association result of the second multidimensional example binary group, and the association result of the second multidimensional example binary group is used to characterize the degree of association between the torque description content and the second data description content, and then based on the association result of the second multidimensional example binary group, it is compared with the actual degree of association between the first torque data and the second data description content, so as to perform reverse gradient propagation to correct the network coefficient of the original preconfigured network.

Similarly, the configuration process of the original preconfigured network may also include: inputting the second multidimensional data features and the first subject data features in the third multidimensional example binary group into the original preconfigured network, the original preconfigured network outputting the association results of the third multidimensional example binary group, the association results of the third multidimensional example binary group are used to characterize the degree of association between the torque description content after the hidden processing and the first data description content, and then based on the association results of the third multidimensional example binary group, the actual degree of association between the torque description content after the hidden processing and the first data description content is compared, so as to perform reverse gradient propagation to correct the network coefficients of the original preconfigured network.

Furthermore, the configuration process of the original preconfigured network may also include: inputting the first multidimensional data feature and the third subject data feature in the fourth multidimensional example binary group into the original preconfigured network, the original preconfigured network outputting the association result of the fourth multidimensional example binary group, the association result of the fourth multidimensional example binary group is used to characterize the degree of association between the torque description content and the first data description content after the hidden processing, and then based on the association result of the fourth multidimensional example binary group, it is compared with the actual degree of association between the torque description content and the first data description content after the hidden processing, so as to perform reverse gradient propagation to correct the network coefficients of the original preconfigured network.

In this embodiment, the network is used to determine whether the first torque data and the event torque data come from the same first source, and compared with the actual result, thereby performing reverse gradient propagation to correct the network coefficients of the original pre-configured network to ensure the reliability of the pre-configuration.

In one embodiment, in order to reduce the cost of data annotation of torque data configuration examples and subject information configuration examples associated with each torque data configuration example, the original pre-configuration network adopts a configured event dimension network, which is obtained after configuration based on event content information and description content, and the event dimension network is used to calculate the possibility that the event content information and description content belong to the same event torque data. Based on this, the configured event dimension network is loaded as the original pre-configuration network during the pre-configuration process, and the original pre-configuration network is further pre-configured based on a multi-dimensional example set. Therefore, in the process of pre-configuring the configured event dimension network, event features can be learned more quickly, so that a small number of torque data configuration examples and subject information configuration examples can be used in the information association network to learn multi-dimensional details such as events and data. On the basis of ensuring the configuration effect of the information association network, the configuration efficiency of the information association network can also be improved.

In one embodiment, the information association network is obtained by:

Step 1702: Obtain each torque data configuration example and a subject information configuration example associated with each torque data configuration example.

In order to make the information association network compatible with the input of three different types of torque data, namely event torque data, attribute torque data and first torque data, and to ensure the effect of the information association network, in the process of configuring the network to obtain the information association network, the torque data configuration examples include event torque data configuration examples, attribute torque data configuration examples and first torque data configuration examples.

Based on this, torque data configuration examples including multiple torque data types and subject information configuration examples associated with each torque data configuration example are obtained. The actual degree of association between the torque data configuration example and the subject information configuration example can be greater than the association target value, and the actual degree of association between the torque data configuration example and the subject information configuration example can also be less than the association target value, that is, there are positive and negative examples.

Step 1704: based on each torque data configuration example, obtain the torque data feature corresponding to each torque data configuration example, and based on each topic information configuration example, obtain the topic information feature corresponding to each topic information configuration example.

Specifically, the method described in the above embodiment is used to obtain the torque data features corresponding to each torque data configuration example based on each torque data configuration example, and to obtain the theme information features corresponding to each theme information configuration example based on each theme information configuration example.

Step 1706, based on each torque data feature and each subject information feature, obtain the regression analysis correlation degree between each torque data feature and each subject information feature, and debug the original information correlation network based on each actual correlation degree and each regression analysis correlation degree to obtain the information correlation network.

Specifically, the torque data features and the subject information features are determined as the input of the original information association network, and then the original information association network outputs the regression analysis correlation degree between the torque data features and the subject information features. The coefficients of the original information association network are adjusted based on the actual correlation degrees and the regression analysis correlation degrees to obtain the information association network.

In one embodiment, the torque data configuration example is a first torque data configuration example, and the first torque data configuration example is composed of subject data, torsion data, and key data, and the torsion data is composed of a plurality of data nodes. Based on this, based on each torque data configuration example, the torque data features corresponding to each torque data configuration example are obtained, including:

Step 2102, perform event division on the subject data in each first torque data configuration example, obtain the event queue corresponding to the subject data in each first torque data configuration example, and generate based on each event queue: event features corresponding to the subject data in each first torque data configuration example, the event queue includes several event analysis attributes.

Among them, the event features include: first position embedding features, first expert embedding features and event analysis features. The first position embedding features are used to characterize the positions of each attribute in the event queue, the first expert embedding features are used to characterize the types of event features, and the event analysis features include event features corresponding to each attribute.

Step 2104: parse and process the torsion data in each first torque data configuration example to obtain key features corresponding to the torsion data in each first torque data configuration example.

Step 2106: parse and process the key data in each first torque data configuration example to obtain potential features corresponding to the key data in each first torque data configuration example.

And, the torque data features finally obtained include event features, key features and potential features.

In this embodiment, in the process of performing network configuration to obtain the information association network, configuration is performed through the event torque data configuration example, the attribute torque data configuration example and the first torque data configuration example. On the basis of ensuring the reliability of network association, the information association network can also be made compatible with multi-dimensional torque data.

Based on the above embodiment, the complete process of associating torque data with subject information is described below. In this embodiment, the method includes the following steps:

Step 2301, obtain a multi-dimensional example set.

Based on this, the torque description content, the first data description content whose correlation with the torque description content is greater than the correlation target value, and the second data description content whose correlation with the torque description content is less than the correlation target value are obtained.

It can be seen that the multidimensional example set includes at least: a first multidimensional example tuple including torque description content and first data description content, a second multidimensional example tuple including torque description content and second data description content, a third multidimensional example tuple including torque description content and first data description content after hidden processing, and a fourth multidimensional example tuple including torque description content and first data description content after hidden processing.

Based on this, the multidimensional example set specifically includes several positive and negative multidimensional examples. After completing the construction of the multidimensional examples, an indication of the association between the torque description content and the data description content is designed to perform network preconfiguration. The network preconfiguration indication is to input the torque description content and the data description content, use the network to determine whether the torque description content and the data description content come from the same torque data, and perform reverse gradient propagation to correct the network coefficients of the original preconfigured network.

Step 2302: Obtain multi-dimensional data features corresponding to the torque description content in each multi-dimensional example binary group, and subject data features corresponding to the data description content in each multi-dimensional example binary group.

The torque description content includes multiple dimensional data, and when the torque description content is specifically the first content information, the data dimensions involved are the most, that is, when the torque description content is the first content information, it specifically includes subject data, torsion data and key data, and the torsion data is composed of several data nodes. Based on this, when obtaining the multi-dimensional data features and subject data features in each multi-dimensional example binary, the torque description content in the first multi-dimensional example binary is input into the derivative layer in the original pre-configured network, and the subject data, torsion data and key data are derived through the derivative layer to obtain the first multi-dimensional data features corresponding to the torque description content.

Secondly, the first data description content in the first multi-dimensional example binary is determined as the input of the original information association network, and the derivative layer in the original information association network divides the first data description content into events to obtain several attributes of the first data description content, and then obtains the event queue corresponding to the first data description content based on the several attributes of the first data description content, and generates the first subject data feature corresponding to the first data description content based on the event queue corresponding to the first data description content. Based on this, the first multi-dimensional data feature corresponding to the torque description content in the first multi-dimensional example binary and the first subject data feature corresponding to the first data description content can be obtained.

Similarly, based on the aforementioned similar method, the first multidimensional data feature corresponding to the torque description content in the second multidimensional example binary group and the second theme data feature corresponding to the second data description content can also be obtained. The second multidimensional data feature corresponding to the torque description content after hidden processing and the first theme data feature corresponding to the first data description content in the third multidimensional example binary group are obtained. And the first multidimensional data feature corresponding to the torque description content in the fourth multidimensional example binary group and the third theme data feature corresponding to the first data description content after hidden processing are obtained.

Step 2303: Based on the multi-dimensional data features and the subject data features in each multi-dimensional example binary, the original pre-configured network is configured.

Specifically, during the configuration process of the original preconfigured network: the first multi-dimensional data feature and the first subject data feature are input to the original preconfigured network, and the original preconfigured network outputs the association result of the first multi-dimensional example binary group, and the association result of the first multi-dimensional example binary group is used to characterize the degree of association between the torque description content and the first data description content. Based on this, the association result of the first multi-dimensional example binary group is compared with the actual degree of association between the first torque data and the first data description content, so as to perform reverse gradient propagation to correct the network coefficients of the original preconfigured network.

Optionally, in order to reduce the cost of data annotation of torque data configuration examples and subject information configuration examples associated with each torque data configuration example, the original pre-configuration network adopts a configured event dimension network, which is obtained after configuration based on event content information and description content, and the event dimension network is used to calculate the possibility that the event content information and description content belong to the same event torque data. Based on this, the configured event dimension network is loaded as the original pre-configuration network during the pre-configuration process, and the original pre-configuration network is further pre-configured based on a multi-dimensional example set. Therefore, in the process of pre-configuration of the configured event dimension network, event features can be learned more quickly, so that a small number of torque data configuration examples and subject information configuration examples can be used in the information association network to learn multi-dimensional details such as events and data. On the basis of ensuring the configuration effect of the information association network, the configuration efficiency of the information association network can also be improved.

Step 2304: Obtain each torque data configuration example and a subject information configuration example associated with each torque data configuration example.

In order to make the information association network compatible with the input of three different types of torque data, namely event torque data, attribute torque data and first torque data, and to ensure the effect of the information association network, in the process of configuring the network to obtain the information association network, the torque data configuration examples include event torque data configuration examples, attribute torque data configuration examples and first torque data configuration examples.

Based on this, torque data configuration examples including multiple torque data types and subject information configuration examples associated with each torque data configuration example are obtained, and the actual degree of association between the torque data configuration example and the subject information configuration example is greater than the target value of the association degree.

Step 2305: Based on each torque data configuration example, obtain the torque data feature corresponding to each torque data configuration example, and based on each topic information configuration example, obtain the topic information feature corresponding to each topic information configuration example.

Specifically, the method described in the above embodiment is used to obtain the torque data features corresponding to each torque data configuration example based on each torque data configuration example, and to obtain the theme information features corresponding to each theme information configuration example based on each theme information configuration example.

Step 2306, based on each torque data feature and each subject information feature, obtain the regression analysis correlation degree between each torque data feature and each subject information feature, and debug the original information correlation network based on each actual correlation degree and each regression analysis correlation degree to obtain the information correlation network.

Specifically, the torque data features and the subject information features are determined as the input of the original information association network, and then the original information association network outputs the regression analysis correlation degree between the torque data features and the subject information features. The original information association network is then debugged based on the actual correlation degrees and the regression analysis correlation degrees to obtain the information association network.

Step 2307, obtaining the torque data to be associated and the subject information to be associated, wherein the torque data to be associated includes one or more dimensional data.

Among them, the torque data includes one or more dimensional data, that is, the torque data can be but is not limited to event torque data, first torque data and attribute torque data. If the torque data is event torque data, the torque data includes subject data. Similarly, if the torque data is first torque data, the torque data includes subject data, torsion data and key data.

Step 2308, based on the subject information to be associated, obtain the characteristics of the subject information to be associated through the information association network.

Specifically, the terminal determines the subject information to be associated as the input of the information association network, and the derivative layer in the information association network extracts event features from the element information to be associated in the subject information to be associated to obtain the features of the subject information to be associated. Secondly, since the subject information to be associated is specifically event description information, the derivative layer in the information association network specifically divides the subject information to be associated into events to obtain a number of event analyses (attributes), and then obtains the event queue corresponding to the subject information to be associated based on the several attributes, and generates the features of the subject information to be associated corresponding to the subject information to be associated based on each event queue, and the features of the subject information to be associated are specifically event features.

Step 2309, obtaining the event features to be associated in the torque data features to be associated through the information association network, and obtaining the torque data features to be associated from the preset long short-term memory unit.

The torque data to be associated includes data of multiple dimensions. Specifically, the terminal determines the torque data to be associated as the input of the information association network, and the derivative layer in the information association network extracts event features from the torque data to be associated to obtain event features to be associated.

Secondly, the information association network is obtained by debugging the original information association network obtained by pre-configuration based on each torque data configuration example and the subject information configuration example associated with each torque data configuration example. The torque data configuration example includes one or more dimensional data, and the subject information configuration example is obtained by intelligent analysis based on the torque data configuration example. And the information association network is used to calculate the degree of association between the torque data and the subject information.

Based on this, before associating the torque data with the subject information, each first torque data is obtained in advance, the key features of the first torque data are obtained using the key feature extraction network, and the potential features of the first torque data are obtained using the potential feature extraction network, and the key features and potential features of each first torque data are written into a preset long short-term memory unit for caching. Similarly, before associating the torque data with the subject information, each attribute torque data can also be obtained in advance, the key features of the attribute torque data are obtained using the key feature extraction network, and the key features of each attribute torque data are written into a preset long short-term memory unit for caching.

Furthermore, when the user needs to associate information, an information association request is initiated, and the information association request indicates that the torque data is associated with the subject information to be associated. After the terminal obtains the torque data to be associated and the subject information to be associated, it is necessary to further determine the specific data type of the torque data to be associated. If it is the first torque data or the attribute torque data, the key features to be associated, or the key features to be associated and the potential features to be associated are obtained from the preset long-term and short-term memory units, and the event features to be associated in the torque data features to be associated are obtained through the information association network, thereby completing the acquisition of the torque data features.

Step 2310, based on the characteristics of the torque data to be associated and the characteristics of the subject information to be associated, obtain the association result through the information association network.

Specifically, the terminal obtains the association result through the information association network, obtains the association result based on the regression analysis of the characteristics of the torque data to be associated and the characteristics of the subject information to be associated, and then determines the degree of association between the torque data to be associated and the subject information to be associated based on the association result. Based on this, the degree of association between the torque data to be associated and the subject information to be associated can determine whether the torque data to be associated and the subject information to be associated are associated, so that in actual applications, all torque data in the torque database are associated with the subject information to be associated one by one, so that different association results can be obtained, and then response operations are performed based on specific scenario requirements.

On the basis of the above, a coupling torque data analysis system is shown, which includes a processor and a memory that communicate with each other, and the processor is used to read a computer program from the memory and execute it to implement the above method.

Based on the above, a computer-readable storage medium is also provided, on which a computer program stored implements the above method when running.

In summary, based on the above scheme, the torque data to be associated and the subject information to be associated are obtained, the torque data to be associated includes one or more dimensional data, and then the subject information features to be associated corresponding to the subject information to be associated are obtained, and the torque data features to be associated corresponding to the torque data to be associated are obtained, and based on the torque data features to be associated and the subject information features to be associated, an association result is obtained through an information association network, and the association result is used to characterize the degree of association between the torque data to be associated and the subject information to be associated, and the information association network is based on each torque data configuration example and the subject information configuration example associated with each torque data configuration example, and is obtained after debugging the pre-configured original information association network, the torque data configuration example includes one or more dimensional data, and the subject information configuration example is obtained based on intelligent analysis of the torque data configuration example. Thus, a multidimensional example set is obtained, the multidimensional example set includes several multidimensional example tuples, the multidimensional example tuples include torque description content and data description content, the torque description content includes multi-dimensional data, and the multidimensional data features corresponding to the torque description content in each multidimensional example tuple, and the subject data features corresponding to the data description content in each multidimensional example tuple are obtained, the multidimensional data features include multi-dimensional features, and then based on the multidimensional data features and the subject data features in each multidimensional example tuple, the original preconfigured network is configured to obtain the target preconfigured network, and the target preconfigured network is determined as the original information association network. Through multi-dimensional examples, the torque description content including multi-dimensional data and the indication of associating the data description content are designed for network pre-configuration, and then the torque data configuration example and the subject information configuration example through intelligent analysis are used to debug the original information association network obtained by pre-configuration. The information association network thus obtained can be compatible with single-dimensional torque data and multi-dimensional torque data, so that in the process of associating torque data with time information, multi-dimensional data and event details in the torque data can be obtained more accurately and efficiently, thereby improving the accuracy of the association results, that is, ensuring the accuracy of the association between the subject information and the torque data.

It should be understood that the system and its modules shown above can be implemented in various ways. For example, in some embodiments, the system and its modules can be implemented by hardware, software, or a combination of software and hardware. Among them, the hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or a dedicated design hardware. Those skilled in the art will understand that the above methods and systems can be implemented using computer executable instructions and/or included in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on the carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The system and its modules of the present application can not only be implemented by hardware circuits such as ultra-large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., but can also be implemented by software such as executed by various types of processors, and can also be implemented by a combination of the above hardware circuits and software (e.g., firmware).

It should be noted that different embodiments may produce different beneficial effects. In different embodiments, the beneficial effects that may be produced may be any one or a combination of the above, or may be any other possible beneficial effects.

Claims (9)

1. A coupling torque data analysis method, characterized in that the method comprises: Obtaining torque data to be associated and subject information to be associated, wherein the torque data to be associated includes one or more dimensional data; Obtaining a feature of the subject information to be associated corresponding to the subject information to be associated, and obtaining a feature of the torque data to be associated corresponding to the torque data to be associated; Combining the characteristics of the torque data to be associated with the characteristics of the subject information to be associated, obtaining an association result through an information association network, wherein the association result is used to characterize the degree of association between the torque data to be associated and the subject information to be associated, wherein the information association network is obtained by debugging a pre-configured original information association network based on each torque data configuration example and a subject information configuration example associated with each torque data configuration example, wherein the torque data configuration example includes one or more dimensional data, and the subject information configuration example is obtained by performing intelligent analysis in combination with the torque data configuration example; The torque data includes one or more dimensional data, that is, the torque data may be, but is not limited to, event torque data, first torque data, and attribute torque data. If the torque data is event torque data, that is, the torque data includes subject data. Similarly, if the torque data is first torque data, that is, the torque data includes subject data, torsion data, and key data. If the torque data is attribute torque data, that is, the torque data includes subject data and torsion data. Wherein, the information association network is obtained in the following manner: Obtaining each torque data configuration example and a subject information configuration example associated with each torque data configuration example; Based on each of the torque data configuration examples, a torque data feature corresponding to each of the torque data configuration examples is obtained, and based on each of the subject information configuration examples, a subject information feature corresponding to each of the subject information configuration examples is obtained; Based on each of the torque data features and each of the subject information features, a regression analysis correlation degree between each of the torque data features and each of the subject information features is obtained, and based on each of the actual correlation degrees and each of the regression analysis correlation degrees, the original information correlation network is debugged to obtain the information correlation network; The torque data configuration example is a first torque data configuration example, the first torque data configuration example is composed of subject data, torsion data and key data, and the torsion data is composed of a number of data nodes; The step of obtaining torque data features corresponding to each of the torque data configuration examples based on each of the torque data configuration examples includes: Performing event division on the subject data in each of the first torque data configuration examples to obtain an event queue corresponding to the subject data in each of the first torque data configuration examples, and generating, based on each of the event queues: event features corresponding to the subject data in each of the first torque data configuration examples, wherein the event queue includes a plurality of event analysis attributes; Analyzing and processing the torsion data in each of the first torque data configuration examples to obtain key features corresponding to the torsion data in each of the first torque data configuration examples; The key data in each of the first torque data configuration examples are parsed and processed to obtain potential features corresponding to the key data in each of the first torque data configuration examples; wherein the torque data features include the event features, the key features and the potential features. 2. The method according to claim 1, characterized in that the step of obtaining the characteristics of the subject information to be associated corresponding to the subject information to be associated comprises: In combination with the subject information to be associated, obtaining features of the subject information to be associated through the information association network; The obtaining of the to-be-associated torque data feature corresponding to the to-be-associated torque data includes: The event features to be associated in the torque data features to be associated are obtained through the information association network, and the torque data features to be associated are obtained from preset long short-term memory units, wherein the torque data features to be associated include at least one of key features to be associated and potential features to be associated. 3. The method according to claim 2, characterized in that the step of obtaining the to-be-associated torque data feature corresponding to the to-be-associated torque data comprises: If the torque data to be associated is event torque data, the characteristics of the torque data to be associated are obtained through the information association network in combination with the torque data to be associated, and the characteristics of the torque data to be associated are the characteristics of the event to be associated. 4. The method according to claim 2, characterized in that the step of obtaining the to-be-associated torque data feature corresponding to the to-be-associated torque data comprises: If the torque data to be associated is the first torque data, combining the torque data to be associated, obtaining the event feature to be associated through the information association network; In combination with the to-be-associated torque data, the to-be-associated key features and the to-be-associated potential features are obtained from the long short-term memory unit; Wherein, the torque data features to be associated include the event features to be associated, the key features to be associated and the potential features to be associated; Among them, the long short-term memory unit includes: key features obtained by extracting key features from each first torque data in the first torque database through a key feature extraction network, and potential features obtained by extracting potential features from each first torque data through a potential feature extraction network. 5. The method according to claim 2, characterized in that the step of obtaining the to-be-associated torque data feature corresponding to the to-be-associated torque data comprises: If the torque data to be associated is attribute torque data, combining the torque data to be associated, and obtaining the event feature to be associated through the information association network; In combination with the to-be-associated torque data, obtaining the to-be-associated key features from the long short-term memory unit; Wherein, the torque data features to be associated include the event features to be associated and the key features to be associated; Wherein, the long short-term memory unit includes: key features obtained by extracting key features of each attribute torque data in the attribute torque database through a key feature extraction network. 6. The method according to claim 1, characterized in that the information association network is based on each torque data configuration example and a subject information configuration example associated with each torque data configuration example, wherein the subject information configuration example is obtained by analyzing a multi-dimensional example set, and obtaining the multi-dimensional example set further comprises: Obtaining torque description content, first data description content, and second data description content, wherein the correlation degree between the torque description content and the first data description content is greater than a correlation degree target value, and the correlation degree between the torque description content and the second data description content is less than the correlation degree target value; Performing a hiding process on the torque description content to obtain the torque description content after the hiding process; Performing a hiding process on the first data description content to obtain the first data description content after the hiding process; The multidimensional example set includes at least: a first multidimensional example binary group including the torque description content and the first data description content, a second multidimensional example binary group including the torque description content and the second data description content, a third multidimensional example binary group including the torque description content after hidden processing and the first data description content, and a fourth multidimensional example binary group including the torque description content and the first data description content after hidden processing; The multi-dimensional data features corresponding to the torque description content in each multi-dimensional example binary group and the subject data features corresponding to the data description content in each multi-dimensional example binary group are obtained, including: Obtaining a first multi-dimensional data feature corresponding to the torque description content in the first multi-dimensional example binary group, and a first subject data feature corresponding to the first data description content; Obtaining a first multi-dimensional data feature corresponding to the torque description content in the second multi-dimensional example binary group, and a second subject data feature corresponding to the second data description content; Obtaining a second multi-dimensional data feature corresponding to the torque description content after the hidden processing in the third multi-dimensional example binary group, and a first theme data feature corresponding to the first data description content; Obtaining the first multi-dimensional data feature corresponding to the torque description content in the fourth multi-dimensional example binary group, and the third theme data feature corresponding to the first data description content after the hidden processing; The original pre-configured network is configured based on the multi-dimensional data features and the subject data features in each multi-dimensional example binary, including: During configuration of the original pre-configured network: Combining the first multi-dimensional data feature and the first subject data feature, obtaining an association result of the first multi-dimensional example binary group, wherein the association result is used to characterize the degree of association between the torque description content and the first data description content; The network coefficients of the original pre-configured network are optimized in combination with the association results of the first multi-dimensional example bigrams. 7. The method according to claim 6, characterized in that the multi-dimensional example binary group specifically includes first content information and first description content, the first content information includes subject data, twist data and key data; the multi-dimensional data features specifically include event features, key features and potential features; Among them, the original pre-configured network is an event dimension network that has been configured, and the event dimension network is obtained after configuration based on event content information and description content, and the event dimension network is used to calculate the possibility that the event content information and description content belong to the same event torque data. 8. The method according to claim 1, characterized in that the original information association network is obtained in a manner comprising: Obtaining a multi-dimensional example set, wherein the multi-dimensional example set includes a plurality of multi-dimensional example 2-tuples, wherein the multi-dimensional example 2-tuples include torque description content and data description content, wherein the torque description content includes data of multiple dimensions; Obtaining multi-dimensional data features corresponding to the torque description content in each of the multi-dimensional example binary groups, and subject data features corresponding to the data description content in each of the multi-dimensional example binary groups, wherein the multi-dimensional data features include multiple dimensional features; Based on the multi-dimensional data features and the subject data features in each of the multi-dimensional example dyads, the original pre-configured network is configured to obtain a target pre-configured network, and the target pre-configured network is determined as the original information association network. 9. A coupling torque data analysis system, characterized in that it comprises a processor and a memory that communicate with each other, wherein the processor is used to read a computer program from the memory and execute it to implement the method described in any one of claims 1 to 8.