Data processing method, data processing device, electronic equipment and medium

Abstract

The invention relates to the technical field of computers, in particular to a data processing method, a data processing device and an electronic equipment medium. Inputting the user data into a target user prediction model, and determining a target user; obtaining a plurality of target features of the target user and feature values corresponding to the plurality of target features; determining a target reason of the target user according to the feature values corresponding to the plurality of target features and the attribution relationship; and determining a user processing mode according to the target reason of the target user, and applying the user processing mode to the target user. Through the technical scheme of the embodiment of the invention, the problem of low efficiency caused by the fact that an author is usually assisted in the same mode in the related technology can be solved.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, and in particular to a data processing method, a data processing device, an electronic device, a computer-readable storage medium, and a computer program product. Background Art

[0002] With the rapid development of terminal devices and the Internet, some short video platforms have also developed rapidly. On short video platforms, users can publish works and earn income. In the process of using short video platforms, users may reduce or stop publishing works on short video platforms.

[0003] In the related art, it is possible to predict which authors are likely to reduce or stop updating based on their behavior data. After determining the authors who are about to reduce or stop updating, corresponding mechanisms can be applied to these authors to prevent them from reducing or stopping updating. However, the solutions in the related art usually use the same mechanism for different authors, which is less targeted and less efficient. Summary of the invention

[0004] The present disclosure provides a data processing method, a data processing device, an electronic device, a computer-readable storage medium, and a computer program product, to at least solve the problem that the related art generally uses the same method to help authors, resulting in low efficiency. The technical solution of the present disclosure is as follows:

[0005] According to a first aspect of an embodiment of the present disclosure, a data processing method is provided, including: obtaining user data; inputting the user data into a target user prediction model to determine a target user; wherein the target user has an activity level, and the target user is a user whose activity level does not meet a preset activity level; obtaining multiple target features of the target user and feature values ​​corresponding to the multiple target features; wherein the feature value is used to indicate the importance of the target feature; determining a target reason for the target user based on the feature values ​​corresponding to the multiple target features and an attribution relationship; wherein the attribution relationship is used to indicate the association between the target feature and the target reason, and the target reason is the reason why the activity level corresponding to the target user does not meet the preset activity level; determining a user processing method based on the target reason of the target user, and applying the user processing method to the target user.

[0006] Optionally, the target reason of the target user is determined according to the feature values ​​corresponding to multiple target features and the attribution relationship, including: determining multiple target positive target features and multiple target negative target features from the multiple target features according to the feature values ​​corresponding to the multiple target features; wherein the number of target positive target features and the number of target negative target features are the same; determining the target reason of the target user according to the multiple target positive target features and the multiple target negative target features, and the attribution relationship.

[0007] Optionally, the target reason of the target user is determined according to multiple target positive target features and multiple target negative target features, and the attribution relationship, including: dividing the multiple target features into multiple positive target features and multiple negative target features; sorting the multiple positive target features according to the feature values ​​corresponding to the multiple positive target features to obtain a positive sorting result, and sorting the multiple negative target features according to the feature values ​​corresponding to the multiple negative target features to obtain a negative sorting result; determining the positive target feature with a preset ranking in the positive sorting result as the target positive target feature, and determining the negative target feature with a preset ranking in the negative sorting result as the target negative target feature.

[0008] Optionally, the target reasons of the target users are determined according to multiple target positive target features and multiple target negative target features, and the attribution relationship, including: dividing the multiple target positive target features into multiple positive multi-level feature categories; wherein the positive multi-level feature categories are used for mapping in the attribution relationship, each positive multi-level feature category corresponds to a feature quantity, and each positive multi-level feature category corresponds to a category attribute; dividing the multiple target negative target features into multiple negative multi-level feature categories; wherein the negative multi-level feature categories are used for mapping in the attribution relationship, each negative multi-level feature category corresponds to a feature quantity, and each negative multi-level feature category corresponds to a category attribute; comparing the feature quantities corresponding to the positive multi-level feature categories and the feature quantities corresponding to the negative multi-level feature categories with the same category attributes, and determining the positive multi-level feature category or the negative multi-level feature category with a larger feature quantity as the reference multi-level feature category; determining the target reasons of the target users according to the multiple reference multi-level feature categories and the attribution relationship.

[0009] Optionally, the attribution relationship includes multiple mapping nodes, the mapping node has at least two secondary mapping nodes, and the target reason of the target user is determined according to multiple reference multi-level feature categories and the attribution relationship, including: obtaining the target reference multi-level feature category corresponding to the mapping node; wherein the multiple reference multi-level feature categories include the target reference multi-level feature category; determining a first candidate secondary mapping node in at least two secondary mapping nodes according to the target reference multi-level feature category, and continuing mapping until mapping to the last-level target mapping node in the attribution relationship; wherein the target mapping node corresponds to a candidate target reason; and determining the candidate target reason as the target reason of the target user.

[0010] Optionally, the attribution relationship corresponds to a hit rule, and a first candidate secondary mapping node is determined in at least two secondary mapping nodes according to the target reference multi-level feature category, and mapping is continued until it is mapped to the last-level target mapping node in the attribution relationship, including: determining a first candidate secondary mapping node in at least two secondary mapping nodes according to the target reference multi-level feature category, and continuing mapping; when the mapping times reaches a preset mapping times, determining a second candidate secondary mapping node; when the second candidate secondary mapping node is the last-level mapping node in the attribution relationship, determining the second candidate secondary mapping node as the last-level target mapping node; when the second candidate secondary mapping node is not the last-level mapping node in the attribution relationship, mapping is continued according to the hit rule until the last-level target mapping node in the attribution relationship; wherein the hit rule is a rule for mapping according to the remaining reference multi-level feature categories, and the remaining reference multi-level feature categories are reference multi-level feature categories in multiple reference multi-level feature categories except the target reference multi-level feature category.

[0011] Optionally, obtaining a target reference multi-level feature category corresponding to the mapping node includes: obtaining a node category attribute corresponding to the mapping node; and determining a reference multi-level feature category corresponding to a category attribute having the same node category attribute as the target reference multi-level feature category corresponding to the mapping node.

[0012] Optionally, the user is a user who has published a work within a second preset time.

[0013] Optionally, obtain sample user data; wherein the sample user data corresponds to a user label; input the sample user data into the model to be trained to determine the predicted target user; update the neural network parameters of the model to be trained according to the user label and the predicted target user to obtain a target user prediction model.

[0014] According to a second aspect of an embodiment of the present disclosure, a data processing device is provided, and the device also includes: a user data acquisition unit, configured to execute acquisition of user data; a target user determination unit, configured to execute input of user data into a target user prediction model to determine the target user; wherein the target user corresponds to an activity level, and the target user is a user whose activity level does not meet a preset activity level; a target feature determination unit, configured to execute acquisition of multiple target features of the target user and feature values ​​corresponding to the multiple target features; wherein the feature value is used to indicate the importance of the target feature; a target cause determination unit, configured to execute determination of a target cause of the target user based on feature values ​​corresponding to the multiple target features and an attribution relationship; wherein the attribution relationship is used to indicate an association relationship between the target feature and the target cause, and the target cause is the reason why the activity level corresponding to the target user does not meet the preset activity level; a user processing unit, configured to execute determination of a user processing method based on the target cause of the target user, and apply the user processing method to the target user.

[0015] Optionally, the target reason of the target user is determined according to the feature values ​​corresponding to multiple target features and the attribution relationship, and the device also includes: a positive and negative target feature determination unit, configured to execute the determination of multiple target positive target features and multiple target negative target features among the multiple target features according to the feature values ​​corresponding to the multiple target features; wherein the number of target positive target features and the number of target negative target features are the same; a first target reason determination unit, configured to execute the determination of the target reason of the target user according to the multiple target positive target features and the multiple target negative target features, and the attribution relationship.

[0016] Optionally, the target reason of the target user is determined according to multiple target positive target features and multiple target negative target features, and the attribution relationship, and the device also includes: a target feature division unit, configured to execute division of the multiple target features into multiple positive target features and multiple negative target features; a sorting result acquisition unit, configured to execute sorting of the multiple positive target features according to the feature values ​​corresponding to the multiple positive target features to obtain a positive sorting result, and sorting of the multiple negative target features according to the feature values ​​corresponding to the multiple negative target features to obtain a negative sorting result; a target target feature determination unit, configured to execute determination of the positive target feature with a preset ranking in the positive sorting result as the target positive target feature, and determination of the negative target feature with a preset ranking in the negative sorting result as the target negative target feature.

[0017] Optionally, the target reason of the target user is determined according to multiple target positive target features and multiple target negative target features, and the attribution relationship, and the device also includes: a positive multi-level feature category division unit, configured to execute the division of multiple target positive target features into multiple positive multi-level feature categories; wherein the positive multi-level feature categories are used for mapping in the attribution relationship, each positive multi-level feature category corresponds to a feature quantity, and each positive multi-level feature category corresponds to a category attribute; a negative multi-level feature category division unit, configured to execute the division of multiple target negative target features into multiple negative multi-level feature categories; wherein The negative multi-level feature categories are used for mapping in the attribution relationship, each negative multi-level feature category corresponds to a feature quantity, and each negative multi-level feature category corresponds to a category attribute; the reference multi-level feature category determination unit is configured to compare the feature quantity corresponding to the positive multi-level feature categories with the same category attribute and the feature quantity corresponding to the negative multi-level feature categories, and determine the positive multi-level feature category or the negative multi-level feature category with a larger feature quantity as the reference multi-level feature category; the second target reason determination unit is configured to determine the target reason of the target user according to multiple reference multi-level feature categories and the attribution relationship.

[0018] Optionally, the attribution relationship includes multiple mapping nodes, the mapping node has at least two secondary mapping nodes, and the target reason of the target user is determined according to multiple reference multi-level feature categories and the attribution relationship. The device also includes: a target reference multi-level feature category acquisition unit, configured to execute acquisition of the target reference multi-level feature category corresponding to the mapping node; wherein the multiple reference multi-level feature categories include the target reference multi-level feature category; a first candidate secondary mapping node determination unit, configured to execute determination of a first candidate secondary mapping node in at least two secondary mapping nodes according to the target reference multi-level feature category, and continue mapping until mapping to the last level target mapping node in the attribution relationship; wherein the target mapping node corresponds to a candidate target reason; and a third target reason determination unit, configured to execute determination of the candidate target reason as the target reason of the target user.

[0019] Optionally, the attribution relationship corresponds to a hit rule, and a first candidate secondary mapping node is determined in at least two secondary mapping nodes according to the target reference multi-level feature category, and mapping is continued until it is mapped to the last level target mapping node in the attribution relationship. The device also includes: a mapping continuation unit, configured to execute the determination of the first candidate secondary mapping node in at least two secondary mapping nodes according to the target reference multi-level feature category, and continue mapping; a second candidate secondary mapping node determination unit, configured to determine the second candidate secondary mapping node when the mapping number reaches a preset mapping number; a first node judgment unit, The method comprises the steps of: determining the second candidate secondary mapping node as the last-level target mapping node when the second candidate secondary mapping node is the last-level mapping node in the attribution relationship; and the second node judgment unit is configured to continue mapping according to a hit rule until the last-level target mapping node in the attribution relationship when the second candidate secondary mapping node is not the last-level mapping node in the attribution relationship; wherein the hit rule is a rule for mapping according to the remaining reference multi-level feature categories, and the remaining reference multi-level feature categories are reference multi-level feature categories other than the target reference multi-level feature categories in multiple reference multi-level feature categories.

[0020] Optionally, to obtain a target reference multi-level feature category corresponding to a mapping node, the device also includes: a node category attribute acquisition unit, configured to execute acquisition of a node category attribute corresponding to the mapping node; and a category attribute judgment unit, configured to execute determination of a reference multi-level feature category corresponding to a category attribute having the same node category attribute as the target reference multi-level feature category corresponding to the mapping node.

[0021] Optionally, the user is a user who has published a work within a second preset time.

[0022] Optionally, the device further comprises: a sample acquisition unit configured to acquire sample user data; wherein the sample user data corresponds to a user tag; a target user prediction unit configured to input the sample user data into the model to be trained to determine the predicted target user; and a model training unit configured to perform

[0023] The neural network parameters of the model to be trained are updated according to the user label and the predicted target user to obtain the target user prediction model.

[0024] According to a third aspect of an embodiment of the present disclosure, there is provided an electronic device, including:

[0025] processor;

[0026] a memory for storing processor-executable instructions;

[0027] The processor is configured to execute executable instructions to implement the above-mentioned data processing method.

[0028] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided. When instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to perform the above-mentioned data processing method.

[0029] According to a fifth aspect of an embodiment of the present disclosure, a computer program product, a computer program / instruction, is provided, characterized in that the above-mentioned data processing method is implemented when the computer program / instruction is executed by a processor.

[0030] The technical solution provided by the embodiments of the present disclosure brings at least the following beneficial effects:

[0031] In a data processing method provided by an embodiment of the present disclosure, user data can be obtained, the user data can be input into a target user prediction model, the target user can be determined, multiple target features of the target user and feature values ​​corresponding to the multiple target features can be obtained, the target reason of the target user can be determined according to the feature values ​​corresponding to the multiple target features and the attribution relationship, the user processing method can be determined according to the target reason of the target user, and the user processing method can be applied to the target user. In an embodiment of the present disclosure, the target reason of each user can be determined according to multiple target features and the attribution relationship, and the processing method can be determined according to the target reason, thereby improving the pertinence of the processing and thus improving the efficiency of the processing.

[0032] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description are used to explain the principles of the present disclosure, and do not constitute improper limitations on the present disclosure.

[0034] Figure 1 A schematic diagram schematically illustrates an exemplary system architecture for implementing a data processing method in an exemplary embodiment of the present disclosure;

[0035] Figure 2 is a flow chart of a data processing method according to an exemplary embodiment;

[0036] Figure 3 It is a flowchart showing a method of determining a target reason of a target user according to a plurality of target positive target features and a plurality of target negative target features and an attribution relationship according to an exemplary embodiment;

[0037] Figure 4 is a schematic diagram showing a method of determining a negative target feature with a preset ranking in a negative sorting result as a target negative target feature according to an exemplary embodiment;

[0038] Figure 5 is a flow chart showing a method of determining a target reason of a target user according to multiple reference multi-level feature categories and attribution relationships according to an exemplary embodiment;

[0039] Figure 6 is a flow chart showing a method of determining a candidate target reason as a target reason of a target user according to an exemplary embodiment;

[0040] Figure 7 is a flowchart showing, according to an exemplary embodiment, a method of continuing mapping until the last level target mapping node in the attribution relationship when the third candidate secondary mapping node is not the last level mapping node in the attribution relationship;

[0041] Figure 8 is a flowchart showing a method of determining a reference multi-level feature category corresponding to a category attribute having the same category attribute as a node as a target reference multi-level feature category corresponding to a mapping node according to an exemplary embodiment;

[0042] Fig. 9 It is a flowchart showing a method of updating the neural network parameters of a training model according to a user label and a predicted target user to obtain a target user prediction model according to an exemplary embodiment;

[0043] Fig.10 is a flowchart showing a method of predicting whether a user will publish a work in the next n days by using a target user prediction model according to an exemplary embodiment;

[0044] Fig.11is a schematic diagram showing an attribution relationship according to an exemplary embodiment;

[0045] Fig.12 is a flowchart showing a method of continuing mapping according to a hit rule until the last level target mapping node in the attribution relationship according to an exemplary embodiment;

[0046] Fig.13 is a schematic diagram showing a plurality of positive multi-level feature categories, a plurality of negative multi-level feature categories and determined target causes according to an exemplary embodiment;

[0047] Fig.14 is a flow chart of another data processing method according to an exemplary embodiment;

[0048] Fig.15 is a schematic diagram showing the composition of a data processing device according to an exemplary embodiment;

[0049] Fig.16 It is a schematic diagram showing the structure of an electronic device suitable for implementing the exemplary embodiments of the present disclosure according to an exemplary embodiment. DETAILED DESCRIPTION

[0050] In order to enable ordinary persons in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the accompanying drawings.

[0051] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present disclosure. Instead, they are merely examples of devices and methods consistent with some aspects of the present disclosure as detailed in the appended claims.

[0052] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or these functional entities or parts of these functional entities may be implemented in one or more software hardened modules, or these functional entities may be implemented in different networks and / or processor devices and / or microcontroller devices.

[0053] Figure 1 A schematic diagram showing an exemplary system architecture to which the data processing method according to an embodiment of the present disclosure can be applied.

[0054] like Figure 1 As shown, the system architecture 1000 may include one or more of terminal devices 1001, 1002, 1003, a network 1004, and a server 1005. The network 1004 is used to provide a medium for communication links between the terminal devices 1001, 1002, 1003 and the server 1005. The network 1004 may include various connection types, such as wired, wireless communication links, or optical fiber cables, etc.

[0055] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. According to the implementation requirements, there may be any number of terminal devices, networks and servers. For example, the server 1005 may be a server cluster composed of multiple servers.

[0056] Users can use terminal devices 1001, 1002, 1003 to interact with server 1005 through network 1004 to receive or send messages, etc. Terminal devices 1001, 1002, 1003 can be various electronic devices with display screens, including but not limited to smart phones, tablet computers, portable computers, desktop computers, etc. In addition, server 1005 can be a server that provides various services.

[0057] In one embodiment, the executor of the data processing method of the present disclosure may be server 1005, which may obtain user data sent by terminal devices 1001, 1002, and 1003, input the user data into a target user prediction model, determine the target user, obtain multiple target features of the target user and feature values ​​corresponding to the multiple target features, determine the target reason of the target user based on the feature values ​​corresponding to the multiple target features and the attribution relationship, determine a user processing method based on the target reason of the target user, and apply the user processing method to the target user.

[0058] In addition, the data processing method of the present disclosure can also be executed through terminal devices 1001, 1002, 1003, etc. to obtain user data, input the user data into a target user prediction model, determine the target user, obtain multiple target features of the target user and feature values ​​corresponding to the multiple target features, determine the target reason of the target user according to the feature values ​​corresponding to the multiple target features and the attribution relationship, determine a user processing method according to the target reason of the target user, and apply the user processing method to the target user.

[0059] In addition, the implementation process of the data processing method of the present disclosure can also be implemented by the terminal devices 1001, 1002, 1003 and the server 1005. For example, the terminal devices 1001, 1002, 1003 can obtain user data, input the user data into the target user prediction model, determine the target user, obtain multiple target features of the target user and feature values ​​corresponding to the multiple target features, determine the target reason of the target user according to the feature values ​​corresponding to the multiple target features and the attribution relationship, and return the target reason of the target user to the server 1005, so that the server 1005 determines the user processing method according to the target reason of the target user, and applies the user processing method to the target user.

[0060] With the rapid development of terminal devices and the Internet, some short video platforms have also developed rapidly. On short video platforms, users can publish works and earn income. In the process of using short video platforms, users may reduce or stop publishing works on short video platforms.

[0061] In the related art, it is possible to predict which authors are likely to reduce or stop updating based on their behavior data. After determining the authors who are about to reduce or stop updating, corresponding mechanisms can be applied to these authors to prevent them from reducing or stopping updating. However, the solutions in the related art usually use the same mechanism for different authors, which is less targeted and less efficient.

[0062] Figure 2 is a flow chart of a data processing method according to an exemplary embodiment. Figure 2 As shown, the data processing method may include the following steps.

[0063] In step S210, user data is obtained;

[0064] In step S220, the user data is input into the target user prediction model to determine the target user; wherein the target user corresponds to an activity level, and the target user is a user whose activity level does not meet the preset activity level;

[0065] In step S230, a plurality of target features of the target user and feature values ​​corresponding to the plurality of target features are obtained; wherein the feature value is used to indicate the importance of the target feature;

[0066] In step S240, the target reason of the target user is determined according to the feature values ​​corresponding to the multiple target features and the attribution relationship; wherein the attribution relationship is used to indicate the association relationship between the target feature and the target reason, and the target reason is the reason why the activity level corresponding to the target user does not meet the preset activity level;

[0067] In step S250, a user processing method is determined according to the target reason of the target user, and the user processing method is applied to the target user.

[0068] In a data processing method provided by an embodiment of the present disclosure, user data can be obtained, the user data can be input into a target user prediction model, the target user can be determined, multiple target features of the target user and feature values ​​corresponding to the multiple target features can be obtained, the target reason of the target user can be determined according to the feature values ​​corresponding to the multiple target features and the attribution relationship, the user processing method can be determined according to the target reason of the target user, and the user processing method can be applied to the target user. In an embodiment of the present disclosure, the target reason of each user can be determined according to multiple target features and the attribution relationship, and the processing method can be determined according to the target reason, thereby improving the pertinence of the processing and thus improving the efficiency of the processing.

[0069] Next, we will combine Figure 2 The steps S210 to S250 of the data processing method in this exemplary embodiment are described in more detail in the following embodiments.

[0070] Step S210, obtaining user data;

[0071] It should be noted that the present disclosure does not specifically limit the specific type of user data.

[0072] It should be noted that the user data involved in this disclosure are all data that the users are aware of and authorized or that have been fully authorized by all parties.

[0073] Step S220, inputting user data into a target user prediction model to determine the target user;

[0074] In an exemplary embodiment of the present disclosure, after obtaining user data through the above steps, the user data can be input into a target user prediction model, and the target user can be determined by the target user prediction model. Among them, the target user corresponds to an activity level, and the target user is a user whose activity level does not meet the preset activity level. Specifically, the task of the target user prediction model is to identify whether a certain user does not meet the preset activity level within a future period of time (a first preset time). When the prediction result is that the user does not meet the preset activity level within a future period of time, it is determined to be a target user, and the user needs to be processed to prevent the user from stopping updating.

[0075] For example, not meeting the preset activity level within a future period of time (first preset time) means not publishing works within the future period of time.

[0076] It should be noted that the present disclosure does not specifically limit the specific structure of the target user prediction model.

[0077] Step S230, obtaining multiple target features of the target user and feature values ​​corresponding to the multiple target features;

[0078] In an exemplary embodiment of the present disclosure, after the user data is input into the target user prediction model through the above steps, the target user can be determined, and multiple target features of the target user and feature values ​​corresponding to the multiple target features can be obtained. Specifically, the multiple target features of the target user can be multiple target features output by the target user prediction model. When predicting whether the user is the target user through the target user prediction model, if the target user is determined, the multiple target features of the target user can be output at the same time, and the feature values ​​corresponding to the multiple target features can be output; or, the multiple target features of the target user can be multiple target features of preset configurations. When predicting whether the user is the target user through the target user prediction model, if the target user is determined, the feature values ​​corresponding to the target features can be calculated for the multiple target features of the target user at the same time.

[0079] It should be noted that the present disclosure does not specifically limit the specific types of the multiple target features of the target user and the specific method of obtaining the multiple target features of the target user.

[0080] Among them, the characteristic value is used to indicate the importance of the target feature. Specifically, the importance of the target feature is proportional to the characteristic value of the target feature. The higher the importance of the target feature, the higher the characteristic value of the target feature; the lower the importance of the target feature, the lower the characteristic value of the target feature. The characteristic values ​​corresponding to multiple target features can be used to indicate the contribution of multiple target features to the determination of the user as a target user, that is, the higher the characteristic value corresponding to the target feature, the greater the influence of the target feature on the user being determined as a target user.

[0081] It should be noted that the present disclosure does not specifically limit the specific method of obtaining the feature value corresponding to the target feature.

[0082] Step S240, determining the target reason of the target user according to the feature values ​​corresponding to the multiple target features and the attribution relationship;

[0083] In an exemplary embodiment of the present disclosure, after obtaining multiple target features and feature values ​​corresponding to the multiple target features through the above steps, the target reason of the target user can be determined according to the feature values ​​corresponding to the multiple target features and the attribution relationship. The attribution relationship is used to indicate the association between the target feature and the target reason, and the target reason is the reason why the activity level corresponding to the target user does not meet the preset activity level. Specifically, the attribution relationship can be obtained by summarizing and analyzing historical target reasons and historical target features. In the attribution relationship, the target reason leading to the target user can be determined by multiple target features of the target user.

[0084] It should be noted that the present disclosure does not specifically limit the specific method of determining the attribution relationship.

[0085] In an exemplary embodiment of the present disclosure, multiple target features can be sorted according to feature values ​​corresponding to the target features, multiple target features can be determined from the multiple target features according to the sorted results, and the target reasons of the target users can be determined based on the multiple target features and the attribution relationships.

[0086] For example, the target reasons of the target users are: less production, less production feedback, less application activity, force majeure, etc. Further, the target reasons of the target users are one or more of the above target reasons.

[0087] It should be noted that the present disclosure does not specifically limit the specific type of the target reason of the target user.

[0088] Step S250, determining a user processing method according to the target reason of the target user, and applying the user processing method to the target user.

[0089] In an exemplary embodiment of the present disclosure, after determining the target reason of the target user through the above steps, a user processing method can be determined according to the target reason of the target user, and the user processing method can be applied to the target user. Specifically, different target reasons can correspond to different user processing methods, and the target user can be processed according to the user processing method corresponding to each target reason.

[0090] For example, the target reason of the target user is: more production but less production feedback. For this target reason, it means that the traffic of the target user's works is poor, so the user processing method corresponding to this target reason is to increase the traffic for the target user; for another example, the target reason of the target user is: less production but more production feedback. For this target reason, it means that the target user's works perform well, but the target user's production willingness is low, so the user processing method corresponding to this target reason is to increase the task reward for the target user and encourage the target user to publish works; the target reason of the target user is: less production, less activity, and less production feedback. For this target reason, it means that the target user does not have a high trust in the platform and is not familiar with how to use the platform, so the user processing method corresponding to this target reason is to conduct author education for the target user.

[0091] It should be noted that the present disclosure does not specifically limit the specific content of the method of processing users corresponding to the target reason.

[0092] In an exemplary embodiment of the present disclosure, multiple target positive target features and multiple target negative target features can be determined from multiple target features according to feature values ​​corresponding to multiple target features, and the target reason of the target user can be determined according to the multiple target positive target features and the multiple target negative target features and the attribution relationship. Figure 3 As shown, determining the target reason of the target user according to multiple target positive target features and multiple target negative target features and the attribution relationship may include the following S310 to S320:

[0093] Step S310, determining a plurality of target positive target features and a plurality of target negative target features from the plurality of target features according to the feature values ​​corresponding to the plurality of target features; wherein the number of the target positive target features and the number of the target negative target features are the same;

[0094] In an exemplary embodiment of the present disclosure, after obtaining multiple target features and feature values ​​corresponding to the multiple target features through the above steps, multiple target positive target features and multiple target negative target features can be determined from the multiple target features according to the feature values ​​corresponding to the multiple target features. Specifically, the target positive target feature refers to a promotion feature that determines the user as a target user, and the target negative target feature refers to an antagonistic feature that determines the user as a target user.

[0095] In an exemplary embodiment of the present disclosure, multiple target positive target features and multiple target negative target features can be determined from multiple target features based on feature values ​​corresponding to the multiple target features. Specifically, the multiple target features correspond to feature attributes, respectively, wherein the feature attributes include positive feature attributes and negative feature attributes, and the multiple target features can be distinguished into multiple positive target features and multiple negative target features based on the feature attributes, and the target positive target features and the target negative target features can be determined based on the feature values ​​corresponding to each positive target feature and the feature values ​​corresponding to each negative target feature. For example, a certain number of positive target features can be selected from multiple positive target features as target positive target features, and a certain number of negative target features can be selected from multiple negative target features as target negative target features.

[0096] In an exemplary embodiment of the present disclosure, multiple positive target features and multiple negative target features, as well as feature values ​​corresponding to the multiple positive target features and feature values ​​corresponding to the multiple negative target features can be obtained through shap value (model interpretability).

[0097] It should be noted that the present disclosure does not specifically limit the specific method of determining multiple target positive target features and multiple target negative target features among multiple target features according to the feature values ​​corresponding to the multiple target features.

[0098] Step S320, determining the target reason of the target user according to a plurality of target positive target features and a plurality of target negative target features and the attribution relationship.

[0099] In an exemplary embodiment of the present disclosure, after obtaining multiple target positive target features and multiple target negative target features through the above steps, the target reason of the target user can be determined based on the multiple target positive target features and multiple target negative target features and the attribution relationship. Specifically, the importance of the target positive target features and the target negative target features can be compared, and the more important target features among the target positive target features and the target negative target features can be determined, and in the attribution relationship, the target reason of the target user can be determined through the more important target features.

[0100] It should be noted that the present disclosure does not specifically limit the specific method of determining the target reason of the target user based on multiple target positive target features and multiple target negative target features and the attribution relationship.

[0101] Through the above steps S310 to S320, multiple target positive target features and multiple target negative target features can be determined from multiple target features according to the feature values ​​corresponding to the multiple target features, and the target reason of the target user can be determined according to the multiple target positive target features and the multiple target negative target features and the attribution relationship. Through the embodiments of the present disclosure, the accuracy of determining the target reason of the target user can be improved, thereby improving processing efficiency.

[0102] In an exemplary embodiment of the present disclosure, multiple target features can be divided into multiple positive target features and multiple negative target features, and the multiple positive target features are sorted according to the feature values ​​corresponding to the multiple positive target features to obtain a positive sorting result, and the multiple negative target features are sorted according to the feature values ​​corresponding to the multiple negative target features to obtain a negative sorting result, and the positive target features with a preset ranking in the positive sorting result are determined as target positive target features, and the negative target features with a preset ranking in the negative sorting result are determined as target negative target features. Figure 4 As shown, determining the negative target feature of the preset ranking in the negative sorting result as the target negative target feature may include the following steps S410 to S430:

[0103] Step S410, dividing the multiple target features into a plurality of positive target features and a plurality of negative target features;

[0104] In an exemplary embodiment of the present disclosure, after obtaining multiple target features and feature values ​​corresponding to the multiple target features through the above steps, the multiple target features can be divided into multiple positive target features and multiple negative target features. Specifically, the positive target feature refers to a promotion feature that determines the user as a target user, and the negative target feature refers to an antagonistic feature that determines the user as a target user. Among them, the multiple target features respectively correspond to feature attributes, wherein the feature attributes include positive feature attributes and negative feature attributes, and the multiple target features can be distinguished into multiple positive target features and multiple negative target features according to the feature attributes.

[0105] It should be noted that the present disclosure does not specifically limit the specific method of dividing multiple target features into multiple positive target features and multiple negative target features.

[0106] Step S420, sorting the multiple positive target features according to the eigenvalues ​​corresponding to the multiple positive target features to obtain a positive sorting result, and sorting the multiple negative target features according to the eigenvalues ​​corresponding to the multiple negative target features to obtain a negative sorting result;

[0107] Step S430, determining the positive target feature of the preset ranking in the positive sorting result as the target positive target feature, and determining the negative target feature of the preset ranking in the negative sorting result as the target negative target feature.

[0108] In an exemplary embodiment of the present disclosure, after the multiple target features are divided into multiple positive target features and multiple negative target features through the above steps, the multiple positive target features can be sorted according to the eigenvalues ​​corresponding to the multiple positive target features to obtain a positive sorting result, and the multiple negative target features can be sorted according to the eigenvalues ​​corresponding to the multiple negative target features to obtain a negative sorting result. Specifically, the multiple positive target features have corresponding eigenvalues, and the multiple negative target features have corresponding eigenvalues, and the multiple positive target features can be sorted according to the eigenvalues ​​corresponding to the multiple positive target features to obtain a positive sorting result, and the multiple negative target features can be sorted according to the eigenvalues ​​corresponding to the multiple negative target features to obtain a negative sorting result.

[0109] Specifically, the feature value can be used to indicate the importance of the target feature. In some embodiments, the smaller the feature value, the higher the importance of the target feature; in some embodiments, the larger the feature value, the higher the importance of the target feature.

[0110] In an exemplary embodiment of the present disclosure, for multiple positive target features, the multiple positive target features can be sorted in order of feature value from small to large to obtain a forward sorting result, and the positive target feature with a preset ranking in the forward sorting result can be determined as the target positive target feature; or, for multiple positive target features, the multiple positive target features can be sorted in order of feature value from large to small to obtain a forward sorting result, and the positive target feature with a preset ranking in the forward sorting result can be determined as the target positive target feature.

[0111] In an exemplary embodiment of the present disclosure, for multiple negative target features, the multiple negative target features can be sorted in order of feature value from small to large to obtain a negative sorting result, and the negative target feature with a preset ranking in the negative sorting result is determined as the target negative target feature; or, for multiple negative target features, the multiple negative target features can be sorted in order of feature value from large to small to obtain a negative sorting result, and the negative target feature with a preset ranking in the negative sorting result is determined as the target negative target feature.

[0112] It should be noted that the present disclosure does not specifically limit the specific method of sorting multiple positive target features according to the eigenvalues ​​corresponding to multiple positive target features to obtain a positive sorting result, and sorting multiple negative target features according to the eigenvalues ​​corresponding to multiple negative target features to obtain a negative sorting result.

[0113] Through the above steps S410 to S430, multiple target features can be divided into multiple positive target features and multiple negative target features, and the multiple positive target features are sorted according to the feature values ​​corresponding to the multiple positive target features to obtain a positive sorting result, and the multiple negative target features are sorted according to the feature values ​​corresponding to the multiple negative target features to obtain a negative sorting result, and the positive target features with preset rankings in the positive sorting results are determined as target positive target features, and the negative target features with preset rankings in the negative sorting results are determined as target negative target features. Through the embodiments of the present disclosure, a target target feature can be determined from multiple target features, and the determined target target feature can be used to determine the target reason of the target user, which can improve the efficiency of processing users.

[0114] In an exemplary embodiment of the present disclosure, multiple target positive target features can be divided into multiple positive multi-level feature categories, and multiple target negative target features can be divided into multiple negative multi-level feature categories. The number of features corresponding to the positive multi-level feature categories and the number of features corresponding to the negative multi-level feature categories of the same category attributes are compared, and the positive multi-level feature category or the negative multi-level feature category with a larger number of features is determined as a reference multi-level feature category. The target reason of the target user is determined based on the multiple reference multi-level feature categories and the attribution relationship. Figure 5As shown, determining the target reason of the target user according to multiple reference multi-level feature categories and attribution relationships may include the following steps S510 to S540:

[0115] Step S510, dividing a plurality of target positive target features into a plurality of positive multi-level feature categories;

[0116] Step S520, classifying a plurality of target negative target features into a plurality of negative multi-level feature categories;

[0117] In an exemplary embodiment of the present disclosure, after obtaining multiple target positive target features and multiple target negative target features through the above steps, the multiple target negative target features can be divided into multiple negative multi-level feature categories. Among them, the positive multi-level feature category is used for mapping in the attribution relationship, each positive multi-level feature category corresponds to a feature quantity, each positive multi-level feature category corresponds to a category attribute, and the negative multi-level feature category is used for mapping in the attribution relationship, each negative multi-level feature category corresponds to a feature quantity, and each negative multi-level feature category corresponds to a category attribute. Specifically, dividing multiple target positive target features into multiple positive multi-level feature categories means that different target positive target features are divided into different positive multi-level feature categories, and each positive multi-level feature category has different category attributes; dividing multiple target negative target features into multiple negative multi-level feature categories means that different target negative target features are divided into different negative multi-level feature categories, and each negative multi-level feature category has different category attributes.

[0118] For example, the target positive target features are the number of normal works deleted by the author in the past n days and the number of comments deleted by the author in the past n days; the positive multi-level feature categories include production categories and production feedback categories. The production category corresponds to production attributes (category attributes) and production feedback categories (category attributes). The number of normal works deleted by the author in the past n days can be classified into the production category, and the number of comments deleted by the author in the past n days can be classified into the production feedback category; the target negative target features are the number of attentions in the past n days and the number of works published in the past n days. The production category corresponds to production attributes (category attributes) and production feedback categories (category attributes). The number of attentions in the past n days can be classified into the production feedback category, and the number of works published in the past n days can be classified into the production category.

[0119] It should be noted that the present disclosure does not specifically limit the specific categories of positive multi-level feature categories and negative multi-level feature categories.

[0120] Step S530, comparing the number of features corresponding to the positive multi-level feature category and the number of features corresponding to the negative multi-level feature category of the same category attribute, and determining the positive multi-level feature category or the negative multi-level feature category with a larger number of features as the reference multi-level feature category;

[0121] In an exemplary embodiment of the present disclosure, after the multiple target positive target features are divided into multiple positive multi-level feature categories, and the multiple target negative target features are divided into multiple negative multi-level feature categories through the above steps, the number of features corresponding to the positive multi-level feature category and the number of features corresponding to the negative multi-level feature category of the same category attribute can be compared, and the positive multi-level feature category or the negative multi-level feature category with a larger number of features can be determined as a reference multi-level feature category. Specifically, the positive multi-level feature category corresponds to a category attribute, the negative multi-level feature category corresponds to a category attribute, and the positive multi-level feature category corresponds to the negative multi-level feature category, and the number of features corresponding to the positive multi-level feature category and the number of features corresponding to the negative multi-level feature category of the same category attribute can be compared. If the number of features of the positive multi-level feature category is larger, the positive multi-level feature category is determined as the reference multi-level feature category, and if the number of features of the negative multi-level feature category is larger, the negative multi-level feature category is determined as the reference multi-level feature category.

[0122] Step S540 , determining the target reason of the target user according to a plurality of reference multi-level feature categories and attribution relationships.

[0123] In an exemplary embodiment of the present disclosure, after the reference multi-level feature categories are determined through the above steps, the target reason of the target user can be determined based on the multiple reference multi-level feature categories and the attribution relationship. Specifically, multiple reference multi-level feature categories can be determined through the above steps, and the multiple reference multi-level feature categories can be mapped in the attribution relationship so that the attribution process points to the position corresponding to the reference multi-level feature category to determine the target reason of the target user.

[0124] For example, in the attribution relationship, when performing the first mapping, mapping can be performed through production attributes, and the mapping results can include more production and less production. If the reference multi-level feature category of the production attribute determined by the above steps is a negative multi-level feature category, it means that this mapping will be mapped to the mapping node of less production.

[0125] It should be noted that the present disclosure does not specifically limit the specific method of determining the target reason of the target user based on multiple reference multi-level feature categories and attribution relationships.

[0126] Through the above steps S510 to S540, multiple target positive target features can be divided into multiple positive multi-level feature categories, and multiple target negative target features can be divided into multiple negative multi-level feature categories. The number of features corresponding to the positive multi-level feature categories and the number of features corresponding to the negative multi-level feature categories with the same category attributes are compared, and the positive multi-level feature category or the negative multi-level feature category with a larger number of features is determined as the reference multi-level feature category. The target reason of the target user is determined based on the multiple reference multi-level feature categories and the attribution relationship.

[0127] In an exemplary embodiment of the present disclosure, a target reference multi-level feature category corresponding to a mapping node may be obtained, a first candidate secondary mapping node may be determined from at least two secondary mapping nodes according to the target reference multi-level feature category, and mapping may be continued until the last level target mapping node in the attribution relationship is mapped, and the candidate target reason may be determined as the target reason of the target user. Figure 6 As shown, determining the candidate target reason as the target reason of the target user may include the following steps S610 to S630:

[0128] Step S610, obtaining a target reference multi-level feature category corresponding to a mapping node;

[0129] In an exemplary embodiment of the present disclosure, after determining the reference multi-level feature category through the above steps, the target reference multi-level feature category corresponding to the mapping node can be obtained. Among them, the multiple reference multi-level feature categories include the target reference multi-level feature category. Specifically, the attribution relationship includes multiple mapping nodes, the mapping node has at least two secondary mapping nodes, and each mapping node corresponds to a target reference multi-level feature category. For example, the target reference multi-level feature category corresponding to the mapping node is the production category.

[0130] It should be noted that the present disclosure does not specifically limit the specific method for obtaining the target reference multi-level feature category corresponding to the mapping node.

[0131] Step S620, determining a first candidate secondary mapping node from at least two secondary mapping nodes according to the target reference multi-level feature category, and continuing mapping until mapping to the last level target mapping node in the attribution relationship;

[0132] Step S630: determine the candidate target reason as the target reason of the target user.

[0133] In an exemplary embodiment of the present disclosure, after the target reference multi-level feature category is determined through the above steps, a first candidate secondary mapping node can be determined in at least two secondary mapping nodes according to the target reference multi-level feature category, and mapping can be continued until it is mapped to the last level target mapping node in the attribution relationship. Among them, the target mapping node corresponds to a candidate target cause. Specifically, it can be determined whether at least two secondary mapping nodes of the mapping node are positive or negative, and the mapping method can be determined according to whether the target reference multi-level feature category is a positive multi-level feature category or a negative multi-level feature category. When the target reference multi-level feature category is a positive multi-level feature category, it is mapped to a positive secondary mapping node, and when the target reference multi-level feature category is a negative multi-level feature category, it is mapped to a negative secondary mapping node.

[0134] For example, the target reference multi-level feature category corresponding to the mapping node is the production category, and the two secondary mapping nodes corresponding to the mapping node are low production and high production respectively. The target reference multi-level feature category is a negative multi-level feature category, and the category attribute of the negative multi-level feature category is the production category. At this time, it means that it is necessary to map from the mapping node to the secondary mapping node with low production (the first candidate secondary mapping node), and continue to map according to the above rules until it is mapped to the last level target mapping node in the attribution relationship. The target mapping node corresponds to a candidate target reason, and the candidate target reason can be used as the target reason of the target user.

[0135] Specifically, the candidate target reasons of the target mapping node are determined through the mapping process. For example, in the first mapping, the target reason corresponding to the first candidate secondary mapping node is low production, the target reason corresponding to the second candidate secondary mapping node is low production feedback, and the target reason of the last target mapping node is low application activity, that is, the candidate target reasons corresponding to the target mapping node are low production, low production feedback, and low application activity.

[0136] Through the above steps S610 to S630, the target reference multi-level feature category corresponding to the mapping node can be obtained, and the first candidate secondary mapping node is determined in at least two secondary mapping nodes according to the target reference multi-level feature category, and the mapping is continued until it is mapped to the last level target mapping node in the attribution relationship, and the candidate target reason is determined as the target reason of the target user. Through the embodiments of the present disclosure, the target reason of the target user can be determined through the mapping process in the attribution relationship, which can improve the accuracy of target reason discrimination and thus improve the efficiency of processing users.

[0137] In an exemplary embodiment of the present disclosure, a first candidate secondary mapping node can be determined from at least two secondary mapping nodes according to the target reference multi-level feature category, and mapping can be continued. When the mapping times reach a preset mapping times, a second candidate secondary mapping node is determined. When the second candidate secondary mapping node is the last level mapping node in the attribution relationship, the second candidate secondary mapping node is determined as the last level target mapping node. When the second candidate secondary mapping node is not the last level mapping node in the attribution relationship, mapping can be continued according to the hit rule until the last level target mapping node in the attribution relationship is reached. Figure 7 As shown, when the third candidate secondary mapping node is not the last level mapping node in the attribution relationship, continuing mapping until the last level target mapping node in the attribution relationship may include the following steps S710 to S740:

[0138] Step S710, determining a first candidate secondary mapping node from at least two secondary mapping nodes according to the target reference multi-level feature category, and continuing mapping;

[0139] Step S720, when the mapping times reaches a preset mapping times, determining a second candidate secondary mapping node;

[0140] In an exemplary embodiment of the present disclosure, after the target reference multi-level feature category is determined through the above steps, a first candidate secondary mapping node can be determined in at least two secondary mapping nodes according to the target reference multi-level feature category, and mapping can be continued. Specifically, it can be determined whether at least two secondary mapping nodes of the mapping node are positive or negative, and the mapping method can be determined according to whether the target reference multi-level feature category is a positive multi-level feature category or a negative multi-level feature category. When the target reference multi-level feature category is a positive multi-level feature category, it is mapped to a positive secondary mapping node, and when the target reference multi-level feature category is a negative multi-level feature category, it is mapped to a negative secondary mapping node. When mapping to the first candidate secondary mapping node, mapping continues from the first candidate secondary mapping node, and the mapping method is the same as the above mapping method, until the number of mappings reaches a preset number of mappings, and the second candidate secondary mapping node is determined.

[0141] It should be noted that the present disclosure does not impose any special limitation on the specific value of the preset mapping times.

[0142] Step S730, when the second candidate secondary mapping node is the last level mapping node in the attribution relationship, determining the second candidate secondary mapping node as the last level target mapping node;

[0143] Step S740: When the second candidate secondary mapping node is not the last level mapping node in the attribution relationship, continue mapping according to the hit rule until the last level target mapping node in the attribution relationship is reached.

[0144] In an exemplary embodiment of the present disclosure, when the second candidate secondary mapping node is determined as described above, when the second candidate secondary mapping node is the last level mapping node in the attribution relationship, the second candidate secondary mapping node can be determined as the last level target mapping node, and when the second candidate secondary mapping node is not the last level mapping node in the attribution relationship, the mapping is continued according to the hit rule until the last level target mapping node in the attribution relationship. The hit rule is a rule for mapping according to the remaining reference multi-level feature categories, and the remaining reference multi-level feature categories are reference multi-level feature categories other than the target reference multi-level feature category in multiple reference multi-level feature categories. Specifically, when the mapping times reaches the preset mapping times, the last-level mapping node has been reached, and the mapping node can be determined as the target mapping node; if the mapping times reaches the preset mapping times, the last-level mapping node has not been reached, indicating that mapping needs to continue, at this time, mapping can be continued according to the hit rule until the last-level target mapping node in the attribution relationship. Specifically, the secondary mapping node corresponding to the second candidate secondary mapping node can be obtained, and the second candidate secondary mapping node corresponds to a category attribute. The secondary mapping nodes corresponding to each second candidate secondary mapping node are respectively used to indicate the positive and negative directions. At this time, the remaining reference multi-level feature categories with the same category attributes as the second candidate secondary mapping nodes can be searched in multiple remaining reference multi-level feature categories, and the remaining reference multi-level feature categories can be used to determine whether the next mapping points to the positive or negative direction, until the last-level target mapping node in the attribution relationship.

[0145] For example, the second candidate secondary mapping node is less production feedback, and the second candidate secondary mapping node includes three secondary mapping nodes: revenue, traffic, and experience. The secondary mapping node (revenue) corresponds to more revenue (positive) and less revenue (negative), the secondary mapping node (traffic) corresponds to more traffic (positive) and less traffic (negative), and the secondary mapping node (experience) corresponds to good experience (positive) and poor experience (negative). At this time, if the remaining reference multi-level feature categories include the number of exposures in the past n days (target negative target feature), it can be mapped from the second candidate secondary mapping node to the secondary mapping node with less traffic. If the secondary mapping node is the last level mapping node, the secondary mapping node is determined to be the target mapping node. If not, continue to map according to the above-mentioned hit rules until it is mapped to the last level target mapping node in the attribution relationship.

[0146] It should be noted that the present disclosure does not specifically limit the specific manner of continuing mapping according to the hit rule.

[0147] Through the above steps S710 to S740, a first candidate secondary mapping node can be determined among at least two secondary mapping nodes according to the target reference multi-level feature category, and mapping can be continued. When the mapping times reaches a preset mapping times, a second candidate secondary mapping node can be determined. When the second candidate secondary mapping node is the last-level mapping node in the attribution relationship, the second candidate secondary mapping node is determined as the last-level target mapping node. When the second candidate secondary mapping node is not the last-level mapping node in the attribution relationship, mapping can be continued according to the hit rule until the last-level target mapping node in the attribution relationship is reached.

[0148] In an exemplary embodiment of the present disclosure, the node category attribute corresponding to the mapping node can be obtained, and the reference multi-level feature category corresponding to the category attribute with the same node category attribute is determined as the target reference multi-level feature category corresponding to the mapping node. Figure 8 As shown, determining the reference multi-level feature category corresponding to the category attribute with the same node category attribute as the target reference multi-level feature category corresponding to the mapping node may include the following steps S810-S820:

[0149] Step S810, obtaining the node category attribute corresponding to the mapping node;

[0150] Step S820: Determine the reference multi-level feature category corresponding to the category attribute with the same node category attribute as the target reference multi-level feature category corresponding to the mapping node.

[0151] In an exemplary embodiment of the present disclosure, the node category attribute corresponding to the mapping node can be obtained, and the reference multi-level feature category corresponding to the category attribute with the same node category attribute is determined as the target reference multi-level feature category corresponding to the mapping node. Specifically, the mapping node corresponds to a node category attribute, and the reference multi-level feature category corresponds to a category attribute. The reference multi-level feature category with the same category attribute as the node category attribute can be searched in multiple reference multi-level feature categories as the target reference multi-level feature category corresponding to the mapping node.

[0152] For example, multiple reference multi-level feature categories and the category attributes corresponding to the reference multi-level feature categories are: reference multi-level feature category A-production attribute, reference multi-level feature category B-production feedback attribute, reference multi-level feature category C-application activity attribute, and the node category attribute corresponding to the mapping node is the production attribute. At this time, the reference multi-level feature category A can be determined as the target reference multi-level feature category.

[0153] It should be noted that the present disclosure does not specifically limit the specific manner of determining the reference multi-level feature category corresponding to the category attribute with the same node category attribute as the target reference multi-level feature category corresponding to the mapping node.

[0154] Through the above steps S810 to S820, the node category attribute corresponding to the mapping node can be obtained, and the reference multi-level feature category corresponding to the category attribute with the same node category attribute is determined as the target reference multi-level feature category corresponding to the mapping node.

[0155] In an exemplary embodiment of the present disclosure, the user is a user who has published a work within the second preset time. Specifically, the user who has published a work within the second preset time refers to a user who has published a work within the second preset time in the past.

[0156] In an exemplary embodiment of the present disclosure, sample user data may be obtained, the sample user data may be input into a model to be trained, a predicted target user may be determined,

[0157] According to the user label and the predicted target user, the neural network parameters of the training model are updated to obtain the target user prediction model. Fig. 9 As shown, updating the neural network parameters of the training model according to the user label and the predicted target user to obtain the target user prediction model may include the following steps S910 to S930:

[0158] Step S910, obtaining sample user data; wherein the sample user data corresponds to a user tag;

[0159] Step S920, inputting sample user data into the model to be trained to determine the predicted target user;

[0160] Step S930: According to the user tag and the predicted target user

[0161] The neural network parameters of the model to be trained are updated to obtain the target user prediction model.

[0162] In an exemplary embodiment of the present disclosure, after obtaining sample user data, the sample user data may be input into a model to be trained to determine that the sample user is a predicted target user. Specifically, the model to be trained may be used to predict whether a user is a target user.

[0163] In an exemplary embodiment of the present disclosure, the model to be trained may include multiple hidden layers, and the hidden layers may include convolutional layers, normalization layers, excitation layers, etc. The sample user data corresponding to the sample users may be sequentially input into the multiple hidden layers of the model to be trained to obtain hidden layer calculation results, and whether the user is a predicted target user is determined by the hidden layer calculation results.

[0164] It should be noted that the present disclosure does not specifically limit the specific method of determining sample users as predicted target users.

[0165] In an exemplary embodiment of the present disclosure, after determining that the user is a predicted target user through the above steps, the neural network parameters of the model to be trained can be updated according to the user label and the prediction result of the predicted target user to obtain the target user prediction model. Among them, whether the user is the predicted target user is a predicted value, and the user label can be used to indicate the true value of whether the user is the target user. At this time, the predicted target user (prediction value) can be compared with the user label (true value) to obtain the predicted difference between the predicted target user (prediction value) and the user label (true value), and the neural network parameters of the model to be trained can be updated according to the prediction difference to obtain the target user prediction model.

[0166] Specifically, the neural network parameters of the model to be trained may include the number of model layers, the number of feature vector channels, the learning rate, etc. When the neural network parameters of the model to be trained are updated according to the prediction difference, the number of model layers, the number of feature vector channels, and the learning rate of the model to be trained can be updated to train the target user prediction model.

[0167] In an exemplary embodiment of the present disclosure, the neural network parameters of the model to be trained may be updated by a back propagation algorithm, and after the training is completed, a target user prediction model may be obtained.

[0168] It should be noted that the present disclosure does not specifically limit the specific method of updating the neural network parameters of the training model according to the user tags and the predicted target users.

[0169] In an exemplary embodiment of the present disclosure, the neural network parameters of the model to be trained can be updated according to the user label and the predicted target user, and when the model to be trained meets the convergence condition, the model to be trained is determined as the target user prediction model. Specifically, the model to be trained meets the convergence condition, which means that the prediction accuracy of the model to be trained is high and can be applied. For example, the convergence condition may include the number of training times, such as when the model to be trained is trained N times, the training ends; for another example, the convergence condition may include the training duration, such as when the model to be trained is trained for T duration, the training ends.

[0170] It should be noted that the present disclosure does not make any special restrictions on the specific content of the convergence conditions. By applying convergence conditions to the model, the training process of the model to be trained can be better controlled to avoid the problem of over-training of the neural network, thereby improving the training efficiency of the model to be trained.

[0171] In an exemplary embodiment of the present disclosure, for example, Fig.10As shown, when the to-be-trained model is being trained, the input sample users may be users who have published works in the past n days, the positive samples are users who have not published works in the past n days, and the negative samples are users who have published works in the past n days; after the to-be-trained model is trained and the target user prediction model is obtained, the target user prediction model can be used to predict whether the user will publish works in the next n days (whether the preset activity level is met).

[0172] Through the above steps S910 to S930, sample user data can be obtained, the sample user data can be input into the model to be trained, the predicted target user can be determined, and the neural network parameters of the model to be trained can be updated according to the user label and the predicted target user to obtain the target user prediction model. Through the embodiments of the present disclosure, the model to be trained can be trained to obtain the target user prediction model, and the target user prediction model can be used to predict whether the user is the target user, so as to process the target user, thereby improving the efficiency of processing users.

[0173] In an exemplary embodiment of the present disclosure, Fig.11 As shown, it is a schematic diagram of an attribution relationship. In the attribution relationship, multiple mapping nodes are included (for example, active application, inactive application, high production, low production, etc.), and the mapping nodes correspond to target reference multi-level feature categories. For example, for the mapping node of active application, the mapping node has two secondary mapping nodes (high production, low production), and the target reference multi-level feature category corresponding to the mapping node is a negative multi-level feature category indicating high production or a positive multi-level feature category indicating low production. The category attributes of the negative multi-level feature category and the positive multi-level feature category are the same, both of which are production attributes. If the determined target reference multi-level feature category is a positive multi-level feature category, it is mapped from the mapping node of active application to the secondary mapping node of low production, and the mapping continues until the last level target mapping node, and the target reason corresponding to the target mapping node is determined as the target reason of the target user.

[0174] Further, through the above steps, map to the secondary mapping node of less production (first candidate secondary mapping node), continue mapping, when the mapping times reaches 3 times (preset mapping times), the second candidate secondary mapping node can be determined at this time, when the second candidate secondary mapping node is the last level mapping node in the attribution relationship, the second candidate secondary mapping node is determined as the last level target mapping node, and the target reason corresponding to the target mapping node is determined as the target reason of the target user. For example, for the mapping process: less production-good production feedback-force majeure, the mapping times have reached the preset mapping times, and it is mapped to the mapping node of force majeure (the second candidate secondary mapping node), and this mapping node is the last level mapping node in the attribution relationship. The next-level target mapping node can determine the target reason corresponding to the target mapping node as the target reason of the target user; when the second candidate secondary mapping node is not the last-level mapping node in the attribution relationship, continue mapping according to the hit rule until the last-level target mapping node in the attribution relationship. For example, for the mapping process: low production-good production feedback-profit, the mapping times have reached the preset mapping times, and it is mapped to the profit mapping node (the second candidate secondary mapping node), and this mapping node is not the last-level mapping node in the attribution relationship. It can continue mapping according to the hit rule until the last-level target mapping node in the attribution relationship, and determine the target reason corresponding to the target mapping node as the target reason of the target user.

[0175] In an exemplary embodiment of the present disclosure, Fig.12 As shown, mapping can continue according to the hit rule until the last level target mapping node in the attribution relationship.

[0176] S1210, determining as a reference multi-level feature category. Specifically, multiple target positive target features are divided into multiple positive multi-level feature categories, and multiple target negative target features are divided into multiple negative multi-level feature categories, and the number of features corresponding to the positive multi-level feature categories and the number of features corresponding to the negative multi-level feature categories of the same category attributes are compared, and the positive multi-level feature category or the negative multi-level feature category with a larger number of features is determined as the reference multi-level feature category;

[0177] S1220, determining a first candidate secondary mapping node. Specifically, obtaining a target reference multi-level feature category corresponding to the mapping node, and determining a first candidate secondary mapping node from at least two secondary mapping nodes according to the target reference multi-level feature category;

[0178] S1230, determining a second candidate secondary mapping node. Specifically, when the mapping times reaches a preset mapping times, determining a second candidate secondary mapping node;

[0179] S1240, continue mapping according to the hit rule until the target mapping node. Specifically, when the second candidate secondary mapping node is not the last level mapping node in the attribution relationship, continue mapping according to the hit rule until the last level target mapping node in the attribution relationship.

[0180] In an exemplary embodiment of the present disclosure, a plurality of positive multi-level feature categories, a plurality of negative multi-level feature categories, and a determined target cause are as follows: Fig.13 As shown, multiple target positive target features can be divided into multiple positive multi-level feature categories, and the number of target positive target features in each positive multi-level feature category is recorded; multiple target negative target features can be divided into multiple negative multi-level feature categories, and the number of target positive target features in each positive multi-level feature category is recorded; the number of features corresponding to the positive multi-level feature category and the number of features corresponding to the negative multi-level feature category of the same category attributes are compared; the positive multi-level feature category or the negative multi-level feature category with a larger number of features is determined as the reference multi-level feature category; and the target reason of the target user is determined based on the multiple reference multi-level feature categories and the attribution relationship.

[0181] In an exemplary embodiment of the present disclosure, Fig.14 As shown, after the target reason of the target user is determined through the scheme of the present invention, the target type can be determined according to the target reason of the target user; if the target reason is not determined, the traffic is increased for the target user, and if the target user is not active within the third preset time period, the risk of stopping updating of the target user is predicted, and if the user's risk of stopping updating is high, the income of the target user is increased; if the target user is still not active within the fourth preset time period, the target user with undetermined target reason is processed manually, and the target user with determined target reason is processed by intelligent method (determining the user processing method according to the target reason of the target user).

[0182] In a data processing method provided by an embodiment of the present disclosure, user data can be obtained, the user data can be input into a target user prediction model, the target user can be determined, multiple target features of the target user and feature values ​​corresponding to the multiple target features can be obtained, the target reason of the target user can be determined according to the feature values ​​corresponding to the multiple target features and the attribution relationship, the user processing method can be determined according to the target reason of the target user, and the user processing method can be applied to the target user. In an embodiment of the present disclosure, the target reason of each user can be determined according to multiple target features and the attribution relationship, and the processing method can be determined according to the target reason, thereby improving the pertinence of the processing and thus improving the efficiency of the processing.

[0183] Fig.15 is a block diagram of a data processing device according to an exemplary embodiment. Fig.15The data processing device 1500 includes a user data acquisition unit 1510, a target user determination unit 1520, a target feature determination unit 1530, a target reason determination unit 1540, and a user processing unit 1550.

[0184] Among them, the user data acquisition unit is configured to execute acquisition of user data; the target user determination unit is configured to execute input of user data into a target user prediction model to determine the target user; wherein the target user has an activity level, and the target user is a user whose activity level does not meet the preset activity level; the target feature determination unit is configured to execute acquisition of multiple target features of the target user and feature values ​​corresponding to the multiple target features; wherein the feature value is used to indicate the importance of the target feature; the target cause determination unit is configured to execute determination of the target cause of the target user based on the feature values ​​corresponding to the multiple target features and the attribution relationship; wherein the attribution relationship is used to indicate the association between the target feature and the target cause, and the target cause is the reason why the activity level corresponding to the target user does not meet the preset activity level; the user processing unit is configured to execute determination of a user processing method based on the target cause of the target user, and apply the user processing method to the target user.

[0185] Optionally, the target reason of the target user is determined according to the feature values ​​corresponding to multiple target features and the attribution relationship, and the device also includes: a positive and negative target feature determination unit, configured to execute the determination of multiple target positive target features and multiple target negative target features among the multiple target features according to the feature values ​​corresponding to the multiple target features; wherein the number of target positive target features and the number of target negative target features are the same; a first target reason determination unit, configured to execute the determination of the target reason of the target user according to the multiple target positive target features and the multiple target negative target features, and the attribution relationship.

[0186] Optionally, the target reason of the target user is determined according to multiple target positive target features and multiple target negative target features, and the attribution relationship, and the device also includes: a target feature division unit, configured to execute division of the multiple target features into multiple positive target features and multiple negative target features; a sorting result acquisition unit, configured to execute sorting of the multiple positive target features according to the feature values ​​corresponding to the multiple positive target features to obtain a positive sorting result, and sorting of the multiple negative target features according to the feature values ​​corresponding to the multiple negative target features to obtain a negative sorting result; a target target feature determination unit, configured to execute determination of the positive target feature with a preset ranking in the positive sorting result as the target positive target feature, and determination of the negative target feature with a preset ranking in the negative sorting result as the target negative target feature.

[0187] Optionally, the target reason of the target user is determined according to multiple target positive target features and multiple target negative target features, and the attribution relationship, and the device also includes: a positive multi-level feature category division unit, configured to execute the division of multiple target positive target features into multiple positive multi-level feature categories; wherein the positive multi-level feature categories are used for mapping in the attribution relationship, each positive multi-level feature category corresponds to a feature quantity, and each positive multi-level feature category corresponds to a category attribute; a negative multi-level feature category division unit, configured to execute the division of multiple target negative target features into multiple negative multi-level feature categories; wherein The negative multi-level feature categories are used for mapping in the attribution relationship, each negative multi-level feature category corresponds to a feature quantity, and each negative multi-level feature category corresponds to a category attribute; the reference multi-level feature category determination unit is configured to compare the feature quantity corresponding to the positive multi-level feature categories with the same category attribute and the feature quantity corresponding to the negative multi-level feature categories, and determine the positive multi-level feature category or the negative multi-level feature category with a larger feature quantity as the reference multi-level feature category; the second target reason determination unit is configured to determine the target reason of the target user according to multiple reference multi-level feature categories and the attribution relationship.

[0188] Optionally, the attribution relationship includes multiple mapping nodes, the mapping node has at least two secondary mapping nodes, and the target reason of the target user is determined according to multiple reference multi-level feature categories and the attribution relationship. The device also includes: a target reference multi-level feature category acquisition unit, configured to execute acquisition of the target reference multi-level feature category corresponding to the mapping node; wherein the multiple reference multi-level feature categories include the target reference multi-level feature category; a first candidate secondary mapping node determination unit, configured to execute determination of a first candidate secondary mapping node in at least two secondary mapping nodes according to the target reference multi-level feature category, and continue mapping until mapping to the last level target mapping node in the attribution relationship; wherein the target mapping node corresponds to a candidate target reason; and a third target reason determination unit, configured to execute determination of the candidate target reason as the target reason of the target user.

[0189] Optionally, the attribution relationship corresponds to a hit rule, and a first candidate secondary mapping node is determined in at least two secondary mapping nodes according to the target reference multi-level feature category, and mapping is continued until it is mapped to the last level target mapping node in the attribution relationship. The device also includes: a mapping continuation unit, configured to execute the determination of the first candidate secondary mapping node in at least two secondary mapping nodes according to the target reference multi-level feature category, and continue mapping; a second candidate secondary mapping node determination unit, configured to determine the second candidate secondary mapping node when the mapping number reaches a preset mapping number; a first node judgment unit, The method comprises the steps of: determining the second candidate secondary mapping node as the last-level target mapping node when the second candidate secondary mapping node is the last-level mapping node in the attribution relationship; and the second node judgment unit is configured to continue mapping according to a hit rule until the last-level target mapping node in the attribution relationship when the second candidate secondary mapping node is not the last-level mapping node in the attribution relationship; wherein the hit rule is a rule for mapping according to the remaining reference multi-level feature categories, and the remaining reference multi-level feature categories are reference multi-level feature categories other than the target reference multi-level feature categories in multiple reference multi-level feature categories.

[0190] Optionally, to obtain a target reference multi-level feature category corresponding to a mapping node, the device also includes: a node category attribute acquisition unit, configured to execute acquisition of a node category attribute corresponding to the mapping node; and a category attribute judgment unit, configured to execute determination of a reference multi-level feature category corresponding to a category attribute having the same node category attribute as the target reference multi-level feature category corresponding to the mapping node.

[0191] Optionally, the user is a user who has published a work within a second preset time.

[0192] Optionally, the device also includes: a sample acquisition unit, configured to execute acquisition of sample user data; wherein the sample user data corresponds to a user label; a target user prediction unit, configured to execute input of the sample user data into the model to be trained, and determine the sample user as the predicted target user; and a model training unit, configured to execute updating of the neural network parameters of the model to be trained according to the user label and the predicted target user, so as to obtain a target user prediction model.

[0193] In a data processing method provided by an embodiment of the present disclosure, user data can be obtained, the user data can be input into a target user prediction model, the target user can be determined, multiple target features of the target user and feature values ​​corresponding to the multiple target features can be obtained, the target reason of the target user can be determined according to the feature values ​​corresponding to the multiple target features and the attribution relationship, the user processing method can be determined according to the target reason of the target user, and the user processing method can be applied to the target user. In an embodiment of the present disclosure, the target reason of each user can be determined according to multiple target features and the attribution relationship, and the processing method can be determined according to the target reason, thereby improving the pertinence of the processing and thus improving the efficiency of the processing.

[0194] Regarding the data processing device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0195] Refer to the following Fig.16 16 is a diagram to describe an electronic device 1600 according to such an embodiment of the present disclosure. Fig.16 The electronic device 1600 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0196] like Fig.16 As shown, the electronic device 1600 is in the form of a general computing device. The components of the electronic device 1600 may include, but are not limited to: the at least one processing unit 1610, the at least one storage unit 1620, a bus 1630 connecting different system components (including the storage unit 1620 and the processing unit 1610), and a display unit 1640.

[0197] The storage unit stores program codes, which can be executed by the processing unit 1610, so that the processing unit 1610 performs the steps according to various exemplary embodiments of the present disclosure described in the above “Exemplary Method” section of this specification. For example, the processing unit 1610 can perform the following steps: Figure 2In step S210 shown in, user data is obtained; in step S220, the user data is input into a target user prediction model to determine the target user; wherein, the target user has an activity level, and the target user is a user whose activity level does not meet the preset activity level; in step S230, multiple target features of the target user and feature values ​​corresponding to the multiple target features are obtained; wherein the feature value is used to indicate the importance of the target feature; in step S240, the target reason of the target user is determined according to the feature values ​​corresponding to the multiple target features and the attribution relationship; wherein the attribution relationship is used to indicate the association between the target feature and the target reason, and the target reason is the reason why the activity level corresponding to the target user does not meet the preset activity level; in step S250, a user processing method is determined according to the target reason of the target user, and the user processing method is applied to the target user.

[0198] The storage unit 1620 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 1621 and / or a cache memory unit 1622 , and may further include a read-only memory unit (ROM) 1623 .

[0199] The storage unit 1620 may also include a program / utility 1624 having a set (at least one) of program modules 1625, such program modules 1625 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0200] Bus 1630 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0201] The electronic device 1600 may also communicate with one or more external devices 1670 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 1600, and / or communicate with any device that enables the electronic device 1600 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 1650. Furthermore, the electronic device 1600 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter 1660. As shown, the network adapter 1660 communicates with other modules of the electronic device 1600 via a bus 1630. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 1600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0202] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the embodiment of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the embodiment of the present disclosure.

[0203] In an exemplary embodiment, a computer-readable storage medium including instructions is also provided, such as a memory including instructions, and the above instructions can be executed by a processor of the device to complete the above method. Optionally, the computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.

[0204] In an exemplary embodiment, a computer program product is also provided, including a computer program / instruction, and when the computer program / instruction is executed by a processor, the data processing method in the above embodiment is implemented.

[0205] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the following claims.

[0206] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A data processing method, characterized in that: include: Get user data; Inputting user data into a target user prediction model to determine a target user; wherein the target user corresponds to an activity level, and the target user is a user whose activity level does not meet a preset activity level; Acquire multiple target features of the target user and feature values ​​corresponding to the multiple target features; wherein the feature value is used to indicate the importance of the target feature; Determining the target reason of the target user according to the feature values ​​corresponding to the multiple target features and the attribution relationship; wherein the attribution relationship is used to indicate the association relationship between the target feature and the target reason, and the target reason is the reason why the activity level corresponding to the target user does not meet the preset activity level; A user processing method is determined according to the target reason of the target user, and the user processing method is applied to the target user.

2. The data processing method according to claim 1, characterized in that: The determining the target reason of the target user according to the feature values ​​corresponding to the multiple target features and the attribution relationship includes: Determining a plurality of target positive target features and a plurality of target negative target features from the plurality of target features according to the feature values ​​corresponding to the plurality of target features; wherein the number of the target positive target features and the number of the target negative target features are the same; The target reason of the target user is determined according to the multiple target positive target features and the multiple target negative target features and the attribution relationship.

3. The data processing method according to claim 2, characterized in that: The determining the target reason of the target user according to the multiple target positive target features and the multiple target negative target features and the attribution relationship includes: Dividing the plurality of target features into a plurality of positive target features and a plurality of negative target features; Sorting the multiple positive target features according to the feature values ​​corresponding to the multiple positive target features to obtain a positive sorting result, and sorting the multiple negative target features according to the feature values ​​corresponding to the multiple negative target features to obtain a negative sorting result; The positive target feature of the preset ranking in the positive sorting result is determined as the target positive target feature, and the negative target feature of the preset ranking in the negative sorting result is determined as the target negative target feature.

4. The data processing method according to claim 2, characterized in that: The determining the target reason of the target user according to the multiple target positive target features and the multiple target negative target features and the attribution relationship includes: Dividing the plurality of target positive target features into a plurality of positive multi-level feature categories; wherein the positive multi-level feature categories are used for mapping in the attribution relationship, each of the positive multi-level feature categories corresponds to a feature quantity, and each of the positive multi-level feature categories corresponds to a category attribute; Divide the multiple target negative target features into multiple negative multi-level feature categories; wherein the negative multi-level feature categories are used for mapping in the attribution relationship, each of the negative multi-level feature categories corresponds to a feature quantity, and each of the negative multi-level feature categories corresponds to a category attribute; Comparing the number of features corresponding to the positive multi-level feature category and the number of features corresponding to the negative multi-level feature category of the same category attribute, and determining the positive multi-level feature category or the negative multi-level feature category with a larger number of features as a reference multi-level feature category; The target reason of the target user is determined according to the multiple reference multi-level feature categories and the attribution relationship.

5. The data processing method according to claim 4, characterized in that: The attribution relationship includes a plurality of mapping nodes, the mapping node has at least two secondary mapping nodes, and the determining the target reason of the target user according to the plurality of reference multi-level feature categories and the attribution relationship includes: Obtaining a target reference multi-level feature category corresponding to the mapping node; wherein the plurality of reference multi-level feature categories include a target reference multi-level feature category; Determine a first candidate secondary mapping node in at least two secondary mapping nodes according to the target reference multi-level feature category, and continue mapping until mapping to the last level target mapping node in the attribution relationship; wherein the target mapping node corresponds to a candidate target cause; The candidate target reason is determined as the target reason of the target user.

6. The data processing method according to claim 5, characterized in that: The attribution relationship corresponds to a hit rule, and the first candidate secondary mapping node is determined in at least two secondary mapping nodes according to the target reference multi-level feature category, and the mapping is continued until the last level target mapping node in the attribution relationship is mapped, including: Determining a first candidate secondary mapping node from at least two secondary mapping nodes according to the target reference multi-level feature category, and continuing mapping; When the mapping number reaches a preset mapping number, determining a second candidate secondary mapping node; When the second candidate secondary mapping node is the last level mapping node in the attribution relationship, determining the second candidate secondary mapping node as the last level target mapping node; When the second candidate secondary mapping node is not the last level mapping node in the attribution relationship, continue mapping according to the hit rule until the last level target mapping node in the attribution relationship; The hit rule is a rule for mapping according to the remaining reference multi-level feature categories, and the remaining reference multi-level feature categories are reference multi-level feature categories among the multiple reference multi-level feature categories except the target reference multi-level feature category.

7. The data processing method according to claim 5, characterized in that: The obtaining of the target reference multi-level feature category corresponding to the mapping node includes: Obtaining the node category attribute corresponding to the mapping node; The reference multi-level feature category corresponding to the category attribute having the same category attribute as the node is determined as the target reference multi-level feature category corresponding to the mapping node.

8. The data processing method according to claim 1, characterized in that: The user is a user who has published a work within a second preset time.

9. The data processing method according to claim 1, characterized in that: The method further comprises: Acquire sample user data; wherein the sample user data corresponds to a user tag; Inputting the sample user data into the model to be trained, and determining the sample user as a predicted target user; According to the user tag and the predicted target user The neural network parameters of the model to be trained are updated to obtain a target user prediction model.

10. A data processing device, characterized in that: include: A user data acquisition unit, configured to acquire user data; A target user determination unit is configured to input user data into a target user prediction model to determine a target user; wherein the target user has an activity level corresponding to it, and the target user is a user whose activity level does not meet a preset activity level; A target feature determination unit is configured to obtain a plurality of target features of the target user and feature values ​​corresponding to the plurality of target features; wherein the feature value is used to indicate the importance of the target feature; A target reason determination unit is configured to determine the target reason of the target user according to the feature values ​​corresponding to the multiple target features and the attribution relationship; wherein the attribution relationship is used to indicate the association relationship between the target feature and the target reason, and the target reason is the reason why the activity level corresponding to the target user does not meet the preset activity level; The user processing unit is configured to determine a user processing method according to the target reason of the target user, and apply the user processing method to the target user.

11. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the executable instructions to implement the data processing method according to any one of claims 1 to 9. 12 . A computer-readable storage medium, when instructions in the computer-readable storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the data processing method according to claim 1 .

Images