CN112612826B

CN112612826B - Data processing method and device

Info

Publication number: CN112612826B
Application number: CN202011522170.8A
Authority: CN
Inventors: 李健伟; 高梓尧; 刘子岳; 黄岑
Original assignee: Beijing Dajia Internet Information Technology Co Ltd
Current assignee: Beijing Dajia Internet Information Technology Co Ltd
Priority date: 2020-12-21
Filing date: 2020-12-21
Publication date: 2024-02-06
Anticipated expiration: 2040-12-21
Also published as: CN112612826A

Abstract

The application discloses a data processing method and device. The method comprises the following steps: acquiring a first historical behavior characteristic of a user account before the current moment; inputting a first historical behavior characteristic into a pre-trained survival probability model, acquiring a first survival prediction probability of a user account in each prediction time period in a prediction period output by the survival probability model, and acquiring a first survival curve of the user account in the prediction period based on the first survival prediction probability; acquiring the active time of the user account for accessing the service for the first time after the current time, and correcting the first survival curve by utilizing the active time to acquire a second survival curve of the user account in the prediction period; and determining the activity gain of the user account based on the difference between the first survival curve and the second survival curve, and determining the influence degree and the influence direction of the first historical behavior characteristic and/or the business characteristic of the business on the activity gain. Through this application, reached the effect that effectively promotes user liveness.

Description

Data processing method and device

Technical Field

The present application relates to the field of data processing, and in particular, to a data processing method and apparatus.

Background

In recent years, the field of internet video has been rapidly developed, wherein various social software, short video APP (Application) and the like become the first choice for billions of users to entertain every day, and numerous internet enterprises are greatly developing related services such as short video and content push, wherein how to improve the activity of users and the viscosity of users have become the core concern of each service.

At present, the main means for improving the user activity is to recommend high-quality push content through a personalized recommendation strategy, and common methods for recommending high-quality push content comprise sorting based on consumption amount, sorting based on conversion rate and the like, which are popular and easy to understand, but have extremely large recommendation algorithm short plates, so that video selection depends on short-term feedback (click rate, duration, praise and the like), which is not matched with the target for improving the user activity, and meanwhile, all videos selected by the method are head videos, which can cause the Martai effect of platform flow distribution to be extremely serious.

The inventors found that in order to solve the drawbacks of recommending business content based on consumption and conversion rate ranking, the related art solution mainly describes video quality from the viewpoint of inherent features of video, for example:

1. Title information of the video;

2. content attribute information, content feedback information, and content source information;

3. the image quality information of the video comprises a frame rate and a code rate.

However, these methods describe more features such as the definition of the video itself, and no effective solution has been proposed for which content can enhance the user's activity.

Disclosure of Invention

The main objective of the present application is to provide a data processing method and apparatus, so as to solve the problem that when pushing content to a user, the user cannot know which content can promote the activity of the user.

To achieve the above object, according to one aspect of the present application, there is provided a data processing method including: acquiring a first historical behavior characteristic of a user account before the current moment; inputting the first historical behavior characteristic to a pre-trained survival probability model, acquiring a first survival prediction probability of the user account in each prediction time period in a prediction period output by the survival probability model, and acquiring a first survival curve of the user account in the prediction period based on the first survival prediction probability; the survival probability model is obtained by taking a second historical behavior characteristic of a sample user account as input and historical activity information of the sample user account as output training, wherein the second historical behavior characteristic is a historical behavior characteristic of the sample user account before a first historical moment, the historical activity information comprises a time interval between the second historical moment and the first historical moment, and the second historical moment is a first active historical moment of the sample user account in a preset time period after the first historical moment; acquiring the active time of the user account for accessing the service for the first time after the current time, and correcting the first survival curve by utilizing the active time to acquire a second survival curve of the user account in the prediction period; and determining the activity gain of the user account based on the difference between the first survival curve and the second survival curve, and determining the influence degree and the influence direction of the first historical behavior characteristic and/or the business characteristic of the business on the activity gain.

Optionally, the determining the influence degree and the influence direction of the first historical behavior feature and/or the service feature of the service on the liveness gain includes: training a tree model by taking the first historical behavior characteristic and the service characteristic of the service as inputs and taking the liveness gain as output; and explaining the trained tree model, and determining the influence degree and the influence direction of the first historical behavior characteristic and/or the business characteristic of the business on the liveness gain.

Optionally, taking the first historical behavior feature and the service feature of the service as input and the liveness gain as output, the training tree model includes: training the tree model by taking the dimension characteristics of the first historical behaviors and the business dimension characteristics of the business as inputs and the liveness gain as output, explaining the trained tree model, and determining the influence degree and the influence direction of the first historical behavior characteristics and/or the business characteristics of the business on the liveness gain comprises: and explaining the trained tree model, and determining the influence degree and the influence direction of the dimension characteristics of the first historical behaviors and/or the service dimension characteristics of the service on the liveness gain.

Optionally, interpreting the trained tree model, determining the influence degree and the influence direction of the first historical behavior feature and/or the service feature of the service on the liveness gain, including: and inputting the trained tree model, the first historical behavior feature and the business feature of the business into a model interpreter, and acquiring the influence degree and the influence direction of the first historical behavior feature and/or the business feature of the business output by the model interpreter on the liveness gain.

Optionally, said modifying said first survival curve with said active time to obtain a second survival curve of said user account during said forecast period includes: acquiring a first lifetime prediction probability of each prediction time period of the user account in a prediction period; acquiring a second survival prediction probability of the user account in a prediction time period to which the active moment belongs; correcting each first survival prediction probability by using the second survival prediction probability; and obtaining the second survival curve based on the corrected first survival prediction probability.

Optionally, after determining the liveness gain of the user account based on the difference between the first survival curve and the second survival curve and determining the influence degree and the influence direction of the first historical behavior feature and/or the service feature of the service on the liveness gain, the method further includes: adding pushing of the business comprising the business feature to the user account corresponding to the first historical behavior feature under the condition that the influence direction of the business feature of the first historical behavior feature and/or the business on the activity gain is positive; and under the condition that the influence direction of the first historical behavior feature and/or the business feature of the business on the activity gain is negative, reducing pushing the business comprising the business feature to the user account corresponding to the first historical behavior feature.

Optionally, after determining an liveness gain of the user account based on the difference between the first survival curve and the second survival curve, and determining an influence degree and an influence direction of the first historical behavior feature and/or the service feature of the service on the liveness gain, the method further includes: if the influence direction of the service characteristics of the service on the liveness gain is a positive direction, pushing the service comprising the service characteristics; if the influence direction of the first historical behavior characteristic on the liveness gain is a positive direction, pushing a service to a client logging in the user account; and pushing the service comprising the service characteristics to a client logging in the user account if the influence direction of the first historical behavior characteristics and the service characteristics of the service on the activity gain is positive.

Optionally, after determining an liveness gain of the user account based on the difference between the first survival curve and the second survival curve, and determining an influence degree and an influence direction of the first historical behavior feature and/or the service feature of the service on the liveness gain, the method further includes: if the influence degree of the service characteristics of the service on the activity gain is greater than a first preset degree and the influence direction is a positive direction, pushing the service comprising the service characteristics; if the influence degree of the first historical behavior characteristic on the activity gain is larger than a second preset degree and the influence direction is a positive direction, pushing a service to a client side logging in the user account; and if the influence directions of the first historical behavior feature and the business feature of the business on the liveness gain are positive directions, the influence degree of the business feature of the business on the liveness gain is larger than a first preset degree, and the influence degree of the first historical behavior feature on the liveness gain is larger than a second preset degree, pushing the business comprising the business feature to a client side logging in the user account.

Optionally, after determining an liveness gain of the user account based on the difference between the first survival curve and the second survival curve, and determining an influence degree and an influence direction of the first historical behavior feature and/or the service feature of the service on the liveness gain, the method further includes: if the influence direction of the service characteristics of the service on the activity gain is a negative direction, reducing pushing the service comprising the service characteristics; if the influence direction of the first historical behavior characteristic on the liveness gain is a negative direction, pushing business to a client side logging in the user account is reduced; and if the influence direction of the first historical behavior characteristic and the business characteristic of the business on the activity gain is a negative direction, pushing the business comprising the business characteristic to a client logging in the user account is reduced.

Optionally, after determining an liveness gain of the user account based on the difference between the first survival curve and the second survival curve, and determining an influence degree and an influence direction of the first historical behavior feature and/or the service feature of the service on the liveness gain, the method further includes: if the influence degree of the service characteristics of the service on the activity gain is greater than a third preset degree and the influence direction is a negative direction, reducing pushing the service comprising the service characteristics; if the influence degree of the first historical behavior characteristic on the liveness gain is larger than a fourth preset degree and the influence direction is a negative direction, pushing services to a client side logging in the user account is reduced; and if the influence direction of the first historical behavior feature and the business feature of the business on the liveness gain is a negative direction, the influence degree of the business feature of the business on the liveness gain is larger than a third preset degree, and the influence degree of the first historical behavior feature on the liveness gain is larger than a fourth preset degree, pushing the business comprising the business feature to a client side logging in the user account is reduced.

In order to achieve the above object, according to another aspect of the present application, there is also provided a data processing apparatus including: a first acquisition unit configured to perform acquisition of a first historical behavior feature of the user account before the current time; the input unit is configured to input the first historical behavior characteristic to a pre-trained survival probability model, acquire first survival prediction probabilities of the user account in each prediction time period in a prediction period output by the survival probability model, and acquire a first survival curve of the user account in the prediction period based on the first survival prediction probabilities; the survival probability model is obtained by taking a second historical behavior characteristic of a sample user account as input and historical activity information of the sample user account as output training, wherein the second historical behavior characteristic is a historical behavior characteristic of the sample user account before a first historical moment, the historical activity information comprises a time interval between the second historical moment and the first historical moment, and the second historical moment is a first active historical moment of the sample user account in a preset time period after the first historical moment; a second obtaining unit configured to obtain an active time of the first access service of the user account after the current time, and correct the first survival curve by using the active time, so as to obtain a second survival curve of the user account in the prediction period; and the determining unit is configured to determine the activity gain of the user account based on the difference between the first survival curve and the second survival curve, and determine the influence degree and the influence direction of the first historical behavior characteristic and/or the business characteristic of the business on the activity gain.

Optionally, the determining unit includes: the training module is configured to execute training a tree model by taking the first historical behavior characteristic and the service characteristic of the service as input and taking the liveness gain as output; and the determining module is configured to execute an interpretation training tree model and determine the influence degree and the influence direction of the first historical behavior characteristic and/or the business characteristic of the business on the liveness gain.

Optionally, the training module is configured to perform training the tree model with the dimension feature of the first historical behavior and the service dimension feature of the service as input and the liveness gain as output, and the determining module is configured to perform interpreting the trained tree model to determine the influence degree and the influence direction of the dimension feature of the first historical behavior and/or the service dimension feature of the service on the liveness gain.

Optionally, the determining module is configured to perform: and inputting the trained tree model, the first historical behavior feature and the business feature of the business into a model interpreter, and acquiring the influence degree and the influence direction of the first historical behavior feature and/or the business feature of the business output by the model interpreter on the liveness gain.

Optionally, the second acquisition unit includes: a first acquisition module configured to perform acquisition of a first lifetime prediction probability for each prediction period of the user account over a prediction period; a second obtaining module configured to obtain a second survival prediction probability of the user account in a prediction period to which the active time belongs; a correction module configured to perform correction of each of the first survival prediction probabilities using the second survival prediction probabilities; and a third acquisition module configured to perform obtaining the second survival curve based on the corrected first survival prediction probability.

Optionally, the apparatus further comprises: a first pushing unit configured to determine an activity gain of the user account based on a difference between the first survival curve and the second survival curve, and determine an influence degree and an influence direction of the first historical behavior feature and/or the service feature of the service on the activity gain, and then, if the influence direction of the first historical behavior feature and/or the service feature of the service on the activity gain is a positive direction, add a pushing of the service including the service feature to the user account corresponding to the first historical behavior feature; the first pushing unit is further configured to perform: and under the condition that the influence direction of the first historical behavior feature and/or the business feature of the business on the activity gain is negative, reducing pushing the business comprising the business feature to the user account corresponding to the first historical behavior feature.

Optionally, the apparatus further comprises: a second pushing unit configured to perform pushing a service including the service feature if an influence direction of the service feature of the service on the liveness gain is a positive direction after determining an liveness gain of the user account based on a difference between the first survival curve and the second survival curve and determining an influence degree and an influence direction of the first historical behavior feature and/or the service feature of the service on the liveness gain; the second pushing unit is further configured to perform: pushing a service to a client logging in the user account when the influence direction of the first historical behavior characteristic on the liveness gain is a positive direction; the second pushing unit is further configured to perform: and pushing the business comprising the business characteristics to a client logging in the user account when the influence direction of the first historical behavior characteristics and the business characteristics of the business on the activity gain is positive.

Optionally, the apparatus further comprises: a third pushing unit configured to perform pushing a service including the service feature when the degree of influence of the service feature of the service on the activity gain is greater than a first preset degree and the direction of influence is a positive direction after determining the activity gain of the user account based on the difference between the first survival curve and the second survival curve and determining the degree of influence and the direction of influence of the service feature of the first historical behavior feature and/or the service feature of the service on the activity gain; the third pushing unit is further configured to perform pushing a service to a client logged in the user account when the degree of influence of the first historical behavior feature on the liveness gain is greater than a second preset degree and the direction of influence is a positive direction; the third pushing unit is further configured to perform pushing, to a client logged in to the user account, a service including the service feature when the first historical behavior feature and the service feature of the service affect the liveness gain in a positive direction, and the service feature of the service affects the liveness gain to a degree greater than a first preset degree, and the first historical behavior feature affects the liveness gain to a degree greater than a second preset degree.

Optionally, the apparatus further comprises: a fourth pushing unit configured to perform, after determining an activity gain of the user account based on a difference between the first survival curve and the second survival curve and determining an influence degree and an influence direction of the first historical behavior feature and/or a service feature of the service on the activity gain, when the influence direction of the service feature of the service on the activity gain is a negative direction, to reduce pushing the service including the service feature; the fourth pushing unit is further configured to perform pushing services to a client logged in the user account when the direction of influence of the first historical behavior feature on the liveness gain is a negative direction; the fourth pushing unit is further configured to perform pushing of the service including the service feature to the client logging in the user account when the first historical behavior feature and the service feature of the service affect the liveness gain in a negative direction.

Optionally, the apparatus further comprises: a fifth pushing unit configured to perform, after determining an activity gain of the user account based on a difference between the first survival curve and the second survival curve and determining an influence degree and an influence direction of the first historical behavior feature and/or a service feature of the service on the activity gain, when the influence degree of the service feature of the service on the activity gain is greater than a third preset degree and the influence direction is a negative direction, to reduce pushing the service including the service feature; the fifth pushing unit is further configured to perform pushing services to clients logging in the user account when the degree of influence of the first historical behavior feature on the liveness gain is greater than a fourth preset degree and the influence direction is a negative direction; the fifth pushing unit is further configured to perform pushing the service including the service feature to the client logged in the user account when the first historical behavior feature and the service feature of the service affect the liveness gain in a negative direction, and the service feature of the service affects the liveness gain to a degree greater than a third preset degree, and the first historical behavior feature affects the liveness gain to a degree greater than a fourth preset degree.

To achieve the above object, according to another aspect of the present application, there is also provided a computer-readable storage medium including a stored program, wherein the data processing method described in the present application is performed when the program is executed by a processor.

To achieve the above object, according to another aspect of the present application, there is also provided an electronic device including: a processor; a memory for storing the processor-executable instructions; wherein the processor is configured to execute the instructions to implement the data processing method described herein.

To achieve the above object, according to another aspect of the present application, there is also provided a computer program product comprising a computer program which, when executed by a processor, performs the data processing method described herein.

The method comprises the steps of obtaining a first historical behavior characteristic of a user account before the current moment; inputting a first historical behavior characteristic into a pre-trained survival probability model, acquiring a first survival prediction probability of a user account in each prediction time period in a prediction period output by the survival probability model, and acquiring a first survival curve of the user account in the prediction period based on the first survival prediction probability; acquiring the active time of the user account for accessing the service for the first time after the current time, and correcting the first survival curve by utilizing the active time to acquire a second survival curve of the user account in the prediction period; based on the difference between the first survival curve and the second survival curve, the activity gain of the user account is determined, the influence degree and the influence direction of the first historical behavior feature and/or the business feature of the business on the activity gain are determined, the activity gain is calculated through the difference between the predicted survival data and the actual survival data, and further, the influence weight and the influence direction of the historical behavior feature and the business feature of the user on the activity are determined, so that business content can be pushed according to the preference of different users, the problem that the user activity can be improved when the content is pushed to the users in the related technology is solved, and the effect that the influence of the different business features on the user activity is determined to effectively improve the user activity is achieved.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application, illustrate and explain the application and are not to be construed as limiting the application. In the drawings:

FIG. 1 is a flow chart of a data processing method according to an embodiment of the present application;

FIG. 2 is a schematic diagram of a survival probability curve of the present embodiment;

FIG. 3 is a flow chart of a method of identifying an impact on user activity of the present embodiment;

FIG. 4 is a graph showing the survival probability of the present embodiment;

FIG. 5a is a schematic diagram of the calculated importance of each dimension of the present embodiment;

FIG. 5b is another schematic illustration of the calculated dimension impact directions of the present embodiment;

FIG. 6 is a schematic diagram of a data processing apparatus according to an embodiment of the present application;

fig. 7 is a schematic diagram of the apparatus of the present embodiment.

Detailed Description

It should be noted that, in the case of no conflict, the embodiments and features in the embodiments may be combined with each other. The present application will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.

In order to make the present application solution better understood by those skilled in the art, the following description will be made in detail and with reference to the accompanying drawings in the embodiments of the present application, it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments herein without making any inventive effort, shall fall within the scope of the present application.

It should be noted that the terms "first," "second," and the like in the description and claims of the present application and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate in order to describe the embodiments of the present application described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

In the technical solution of the embodiment of the present application, the execution subject may be a background server, and in some mobile terminals with relatively strong computing power, the execution subject may also be a mobile terminal. As shown in fig. 1, the method includes the following steps S102-S108.

Step S102: a first historical behavioral characteristic of the user account prior to the current time is obtained.

In one embodiment, the first historical behavioral characteristics may be behavioral characteristics resulting from a user client having logged into the user account accessing the service during a historical period prior to the current time. In one example, the access data of the user client access service includes a user account and a behavior feature.

The first historical behavior feature may be data of browsing, clicking, praying and other behaviors, or may be a feature extracted based on the data of browsing, clicking, praying and other behaviors.

Step S104: and inputting a first historical behavior characteristic into the pre-trained survival probability model, acquiring a first survival prediction probability of the user account in each prediction time period in the prediction period, which is output by the survival probability model, and acquiring a first survival curve of the user account in the prediction period based on the first survival prediction probability. The survival probability model is obtained by taking a second historical behavior characteristic of the sample user account as input and historical activity information of the sample user account as output training, wherein the second historical behavior characteristic is the historical behavior characteristic of the sample user account before the first historical moment, the historical activity information comprises a time interval between the second historical moment and the first historical moment, and the second historical moment is the historical moment when the sample user account is first active in a preset time period after the first historical moment.

The survival probability model is trained by taking actual data as a sample, specifically, the second historical behavior characteristic of the sample user account is taken as an input, and the historical active information of the sample user account is taken as an output. The second historical behavioral characteristics may be data of behaviors such as browsing, clicking, praying, and the like, and may also include attribute characteristics of the user such as age, gender, and the like.

The prediction period may be N days after the current time, where N is a positive integer. In one embodiment, the prediction period may be divided into a plurality of consecutive time periods, which may be referred to as prediction time periods. For example, the prediction period is 7 days, and data within 7 days after the current time is predicted by the survival probability model, and each prediction period in the prediction period is 1 st, 2 nd, and 3 rd days … … th day 7. In an alternative embodiment, the 7 day period of prediction may be set and altered. And then drawing a survival curve for 7 continuous days according to the predicted daily survival probability, wherein the survival curve is the first survival curve.

In one embodiment, the first historical behavior feature is a behavior feature within a preset time period before the current time, and the second historical behavior feature is a behavior feature within a preset time period before the first historical time, and the preset time period can be customized according to the situation, for example, 2 days, 7 days, 14 days, and the like. The prediction period is N days after the current time, N is a positive integer, and N can be customized according to the situation, for example, 3 days, 7 days, 14 days, etc.

For example: assume that the current time is 00 of 10 months and 12 days in 2020: 00, the preset time period is 7 days, the prediction period is 7 days, and then the first historical time is time 2020, 10 months and 5 days 00 which is 7 days before the current time: 00, a first historical behavior characterized by 00, 10 months and 12 days in 2020: behavior characteristics within 7 days before 00, the second historical behavior characteristics being 2020, 10 months 5 days 00: behavior characteristics within 7 days before 00. Again assume that the sample user account is first active for 7 days after the first historical time at a second historical time of 2020, 10 months 8 days 00:00, then the historical activity information includes 3 days (i.e., the time interval between the second historical time and the first historical time).

And training the survival probability model by taking the second historical behavior characteristic within 7 days before the first historical moment as input and taking 3 days as output, wherein when the survival probability model converges, the survival probability model is trained. The first historical behavior feature in 7 days before the current moment is taken as input, and the trained survival probability model is utilized to predict the survival probability (i.e. the first survival prediction probability) of the user account in each day (prediction period) in 7 days (prediction period) after the current moment, for example, the predicted survival probability on the first day is P1, the survival probability on the second day is P2, the survival probability on the third day is P3,... And drawing a first survival curve predicted within 7 days in the future after the current moment by taking the survival probabilities P1, P2, P3, and P7 as ordinate and the predicted time period as abscissa.

Step S106: and acquiring the active time of the user account for accessing the service for the first time after the current time, and correcting the first survival curve by utilizing the active time to acquire a second survival curve of the user account in the prediction period.

In one embodiment, the service may be a video, and the time when the user client accesses the service for the first time after the current time is an active time, where access data of the user client for accessing the service includes a user account and a behavior feature. In one example, when a user client logged into a user account accesses a service at a current time or a historical time before the current time, an active time at which the user client logged into the user account first accessed the service after the current time may be obtained.

For example, the current time is 12 noon on 10 months of 2020: 00, suppose that the user views video 1 in the service at this time through the user client logging into the user account and 11 a.m. on 13 months 2020: 00 accesses this service again through the user client logging into the user account, then the active time is 11 in 11 a.m. 13, 10, 2020: 00.

in one embodiment, the actual survival data (i.e., the active time) of the user account is used to modify the survival prediction probability, and the actual survival curve (i.e., the second survival curve) of the user account is drawn according to the modified survival probability.

Step S108: and determining the activity gain of the user account based on the difference between the first survival curve and the second survival curve, and determining the influence degree and the influence direction of the first historical behavior characteristic and/or the business characteristic of the business on the activity gain.

The first survival curve is corrected through the actual activity data of the user to obtain a second survival curve, the difference of the two survival curves represents the activity gain of the user, and the influence degree and the influence direction of the first historical behavior characteristic and the business characteristic of the business on the activity gain can be reflected. Thus, the service content which is more likely to increase the activity of the user can be pushed to the user during the subsequent service pushing so as to meet different preferences of the user.

This embodiment obtains a first historical behavioral characteristic of the user account prior to the current time; inputting a first historical behavior characteristic into a pre-trained survival probability model, acquiring a first survival prediction probability of a user account in each prediction time period in a prediction period output by the survival probability model, and acquiring a first survival curve of the user account in the prediction period based on the first survival prediction probability; acquiring the active time of the user account for accessing the service for the first time after the current time, and correcting the first survival curve by utilizing the active time to acquire a second survival curve of the user account in the prediction period; according to the method, the system and the device, the activity gain of a user account is determined based on the difference of a first survival curve and a second survival curve, the influence degree and the influence direction of the first historical behavior feature and/or the business feature of the business on the activity gain are determined, the activity gain is calculated through the difference of predicted survival data (the first survival curve drawn based on the first survival prediction probability) and actual survival data (the second survival curve), and then the influence weight and the influence direction of the historical behavior feature of the user and the business feature on the activity are determined, so that business content can be pushed according to the preference of different users, the problem that the user activity can not be improved when the content is pushed to the user in the related technology is solved, and the effect that the influence of the different business features on the user activity is determined so as to effectively improve the user activity is achieved.

In some embodiments, determining the extent and direction of impact of the first historical behavior feature and/or the traffic feature of the traffic on the liveness gain includes: training a tree model by taking first historical behavior characteristics and business characteristics of business as input and activity gain as output; and explaining the trained tree model, and determining the influence degree and the influence direction of the first historical behavior characteristics and/or the business characteristics of the business on the activity gain.

In determining the impact of the user account's liveness and gain, the extent and direction of impact of various factors on liveness may be calculated by a tree model. In one embodiment, a tree model (for example, a decision tree model, a regression tree model, xgboost, lightgbm, etc.) may be used as an initial model, and historical data of an application scenario of the present application is used as a sample to train to obtain a tree model meeting the requirements of the application scenario, and by explaining the trained tree model, the influence degree and the influence direction of the first historical behavior feature and/or the business feature of the business on the activity gain may be determined, where the influence direction is divided into a positive direction and a negative direction, and the positive direction is used to represent improving the activity, and the negative direction is used to represent reducing the activity.

In some embodiments, the determining the degree and direction of influence of the first historical behavior feature and/or the traffic feature of the traffic on the liveness gain is: and explaining the trained tree model, and determining the influence degree and the influence direction of the dimension characteristics of the first historical behaviors and/or the business dimension characteristics of the business on the liveness gain.

It should be noted that, the features in this embodiment may be embodied in the form of dimension features, the service may be a video, and multiple feature dimensions may be set for the video, for example, a 64-dimensional feature may be set for the video. The historical behavior feature and the like are also expressed in the form of dimensions, for example, 64-dimensional features are also set for each user, so that the behavior feature of the user is expressed in 64-dimensional features. The more dimensions, the more refined the conclusions are drawn. In this embodiment, the dimension feature of the user behavior refers to user behavior Embedding, and the video dimension feature refers to video Embedding.

In this embodiment, the user behavior and the video coding may be trained by using the DNN model based on the click behavior of the user on the video, and other modes may be used to extract the user behavior coding and the video coding.

In one embodiment, a trained tree model, a first historical behavior feature and a business feature of a business are input into a model interpreter, and the influence degree and the influence direction of the first historical behavior feature and/or the business feature of the business output by the model interpreter on the activity gain are obtained. The interpretability of causal inference based on the tree model is higher, and model results can be better interpreted and landed.

The model interpreter in this embodiment may be a SHAP module, and the importance (influence degree) and the influence direction of each dimension feature on the liveness gain may be calculated by using treeexplaner and shape_values methods of the SHAP module.

In one embodiment, correcting the first survival curve with the active time to obtain a second survival curve for the user account during the forecast period includes: acquiring a first lifetime prediction probability of each prediction time period of a user account in a prediction period; acquiring the actual activity of a user account in a predicted time period to which the active moment belongs; correcting each first life prediction probability by using the actual activity; and obtaining a second survival curve based on the corrected first survival prediction probability.

In one example, based on the trained survival probability model, a first daily survival prediction probability of the user account for n days in the future may be predicted, P1, P2, P3,..once, pm,..once, pn, respectively, and then the survival probability is modified according to the actual activation time of the user account (e.g., the activation on the mth day after the current time). In this example, the user account is active on the mth day, and based on the trained generation probability model, the first survival prediction probability of the activity on the mth day is predicted to be Pm, and the ratio of each first survival prediction probability to the first survival probability (i.e., the second survival prediction probability) at the actual active time of the user account is obtained as follows: P1/Pm, P2/Pm, P3/Pm, & gt, pm/pm=1, & gt, pn/Pm, the obtained ratios are used as survival probability values after the first survival prediction probability is corrected. Since the maximum value of the survival probability does not exceed 1, the corrected probability values are subjected to threshold pruning, and the corrected survival probability values are 1,..once, pm/pm=1,..once, pn/Pm, respectively. The corrected survival probability value not only uses the actual liveness data of the user, ensures that the difference between the actual liveness data and the first survival prediction probability can describe the change of the liveness of the user account, but also keeps the trend of the survival probability of the user account, and reduces the influence of the liveness noise of the user on the change of the liveness.

In some of the above embodiments, after step S108, further includes: under the condition that the influence direction of the first historical behavior feature and/or the business feature of the business on the activity gain is positive, pushing business comprising the business feature is added to the user account corresponding to the first historical behavior feature; and under the condition that the influence direction of the first historical behavior characteristic and/or the business characteristic of the business on the activity gain is negative, pushing the business comprising the business characteristic to the user account corresponding to the first historical behavior characteristic is reduced.

After determining the influence of the first historical behavior feature and/or the business feature of the business on the activity gain, if the influence direction is a positive direction, pushing business content with activity improving effect can be increased to improve the activity of the user; if the direction of influence is negative, pushing of traffic content that has a decreasing effect on liveness is avoided or reduced. In one embodiment, the directions of influence of different business features on the liveness may be opposite, for example, the first dimension feature has a positive direction influence on liveness, the third dimension feature has a negative direction influence on liveness, and if the influence degree of the first dimension feature on liveness is high, the influence degree of the third dimension feature on liveness is also high, namely: for the video, the larger the first dimension characteristic is, the higher the degree of improvement on the activity is; the greater the third dimension characteristic, the higher the degree of reduction in liveness. In this case, the service with high third dimension feature should be pushed as little as possible, or the service with high first dimension feature should be pushed to be increased, so that the user activity can be improved to the greatest extent.

The method and the device can output the combined influence of the video Embedding and the user behavior Embedding on the activity gain. For example, a video group whose video enhancement has a dimension feature greater than one value (e.g., 0.2) has the most positive effect on the increase in liveness of a user group whose user behavior enhancement has a dimension feature greater than one value (e.g., 0.1), i.e., the greater the video enhancement has for a user group whose user behavior enhancement has been greater than 0.1. According to the method, the influence of the combination of different video dimension characteristics and user dimension characteristics on the user liveness can be calculated, the combination of the video dimension characteristics with the largest forward gain and the user dimension characteristics is selected, and the service contents can be obtained to improve the user liveness of the users, so that the video of the video dimension characteristics is pushed to the users with the user dimension characteristics, the effect of pushing personalized high-quality video to the users is achieved, and the liveness of the users is further improved to the greatest extent.

According to the method and the device for improving the user activity, which business contents can improve the user activity, which users can be improved, and which business contents can improve the user activity of which users can be obtained.

Because the user logs in on the webpage or the client through the account, when pushing the service, the service is pushed to the client or the terminal logging in the user account, the pushed service can be displayed on the webpage, and the specific form is not limited in the application.

In one embodiment, the method can obtain which service contents can improve the activity of the user, and if the influence direction of the service characteristics of the service on the activity gain is a positive direction, pushing the service including the service characteristics. In this embodiment, if the direction of influence of the service feature of the service on the activity gain is positive, it indicates that the service plays a role in forward pushing the activity of the user account, and at this time, the service may be pushed to all user accounts, for example, for a certain APP, if it is determined that the direction of influence of the video 1 on the activity gain of the user account is positive, the service may be pushed to all user accounts registered in the APP.

In one embodiment, the method can acquire the user liveness of which users are promoted, and if the influence direction of the first historical behavior characteristic on the liveness gain is a positive direction, pushing the service to the client side logging in the user account. In this embodiment, if the direction of influence of the first historical behavior feature on the liveness gain is positive, it indicates that the liveness of the user account is not affected by the service, and at this time, the service may be pushed to the client logging in the user account according to the service push policy.

In another embodiment, the method can obtain which service content can improve the user liveness of which users, and if the first historical behavior feature and the service feature of the service have positive influence on the liveness gain, the service including the service feature is pushed to the client logging in the user account. In this embodiment, the direction of influence of the first historical behavior feature and the service feature of the service on the activity gain is positive, which indicates that the service plays a role in forward pushing the activity of the user account corresponding to the first historical behavior feature, and at this time, the service can be pushed to the client terminal logging in the user account. In the embodiment of the application, the user account corresponding to the behavior feature indicates that the behavior feature is the behavior feature of the user. For example, if the behavior feature 1 is a behavior feature of the user account 1, the user account corresponding to the behavior feature 1 is the user account 1.

In some embodiments, the method can obtain which service contents can improve the activity of the user, and if the influence degree of the service characteristics of the service on the activity gain is greater than a first preset degree and the influence direction is a positive direction, the service including the service characteristics is pushed. The first preset degree is set according to actual customization, and the embodiment of the application is not particularly limited.

In some embodiments, the method and the system can acquire the user liveness of which users are promoted, and if the influence degree of the first historical behavior feature on the liveness gain is greater than the second preset degree and the influence direction is the positive direction, the service is pushed to the client logging in the user account. The second preset degree is set according to actual customization, and the embodiment of the application is not particularly limited.

In some embodiments, the method can obtain which service content can improve the user liveness of which users, and if the first historical behavior feature and the service feature of the service have positive influence on liveness gain, and the service feature of the service has influence on liveness gain greater than a first preset degree, and the first historical behavior feature has influence on liveness gain greater than a second preset degree, push the service including the service feature to the client side logged in the user account.

In some other embodiments, the platform push policy may not be changed if the first historical behavior feature and/or the traffic feature of the traffic has a very small impact on the liveness gain. Under the condition that the influence is positive and the influence degree of the activity gain is greater than the preset degree, executing the pushing strategy, wherein the first preset degree and the second preset degree can be the same value or different values, and the pushing strategy can be specifically and custom set according to the situation.

In other embodiments, pushing the traffic including the traffic feature is reduced if the direction of impact of the traffic feature on the liveness gain is negative. The scheme can know which service contents reduce the activity of the user.

And if the influence direction of the first historical behavior characteristic on the activity gain is a negative direction, reducing pushing services to the client side logging in the user account. The scheme can know which users have reduced user liveness.

And if the influence direction of the first historical behavior characteristic and the business characteristic of the business on the activity gain is a negative direction, pushing the business comprising the business characteristic to the client side logging in the user account is reduced. The scheme can know which service contents reduce the user liveness of which users.

For example, for an APP, if it is determined that the direction of influence of video 2 on the activity gain of a user account is negative, pushing the service to all user accounts registered with the APP may be reduced.

In one embodiment, if the degree of influence of the service feature of the service on the activity gain is greater than a third preset degree and the influence direction is a negative direction, pushing the service including the service feature is reduced. The scheme can know which service contents reduce the activity of the user.

And if the influence degree of the first historical behavior characteristic on the activity gain is larger than the fourth preset degree and the influence direction is a negative direction, reducing pushing the business to the client side logging in the user account. The scheme can know which users have reduced user liveness.

And if the influence direction of the first historical behavior feature and the business feature of the business on the activity gain is a negative direction, the influence degree of the business feature of the business on the activity gain is larger than a third preset degree, and the influence degree of the first historical behavior feature on the activity gain is larger than a fourth preset degree, pushing the business comprising the business feature to the client side logging in the user account is reduced. The scheme can know which service contents reduce the user liveness of which users. The third preset degree and the fourth preset degree are set according to the user definition, and the embodiment is not limited.

For the situation that the influence direction is the negative direction, the purpose of the pushing strategy is to reduce the business pushing of which the influence direction is the negative direction so as to avoid influencing the activity of the user.

Based on the scheme provided by the embodiment, the difference of the video which is promoted to be reserved by the user on different dimensions of the editing can be selected according to the user accounts with different characteristics, and therefore the high-quality video which accords with the personalized preference of the user is screened. According to the embodiment, through the input user history behavior characteristics, the activity condition of a plurality of days after watching the video and the Embedding characteristics of watching the video, the user activity change is calculated according to the survival curve area difference, fig. 2 is a schematic diagram of the survival probability curve of the embodiment, the user activity is higher after the push strategy of the embodiment is adopted by curve 2 in the two curves, the influence weight and the influence direction of each dimension Embedding of the video on the user activity of different characteristics are given, and the short video APP analyst is helped to find the service for improving the user retention.

Referring to fig. 3, the technical solution of this embodiment may be used to identify factors affecting the activity of a user account, where the solution is: and measuring the liveness gain of the user account based on the difference of the areas of the survival probability curves, taking the liveness gain as Label, the user behavior Embedding and the watching video Embedding as Fetures input tree models (for example Xgboost, lightgbm), and combining with a SHAP VALUE algorithm, outputting each dimension Embedding of the video and the influence weight and the influence direction of each user dimension Embedding on the liveness of the user account.

Fig. 3 is a flowchart of a method for identifying an influence on the activity of a user account according to the present embodiment, and as shown in fig. 3, the method mainly includes the following four processes (1) (2) (3) (4).

Assume that the current time is 00 of 10 months and 12 days in 2020: 00, the preset time period is 7 days, the prediction period is 7 days, and then the first historical time is time 2020, 10 months and 5 days 00 which is 7 days before the current time: 00, a first historical behavior characterized by 00, 10 months and 12 days in 2020: behavior characteristics within 7 days before 00, the second historical behavior characteristics being 2020, 10 months 5 days 00: behavior characteristics within 7 days before 00. Again assume that the sample user account is first active for 7 days after the first historical time at a second historical time of 2020, 10 months 8 days 00:00, then the historical activity information includes 3 days (i.e., the time interval between the second historical time and the first historical time).

(1) A survival probability model (M1) such as Cox is trained based on the historical behavior characteristics and the historical activity information of the user account, and then an area enclosed by a survival probability curve of the target user and an abscissa axis is predicted by using the M1 (S1).

The specific process is as follows.

1. Training a model: the data of the sample user account is randomly sampled and the data format is selected as shown in table 1.

TABLE 1

The fields are explained as follows.

User_id is a unique identification of the User, such as a User account.

Event indicates Event occurrence, here indicates whether the user account is active, here indicates Event occurrence with 1 and indicates Event non-occurrence with 0. In the example of Table 1, event is 1, indicating that the user account is again active at the second historical time.

Duration represents the Duration of the interval during which the event occurs again, here the Duration interval during which the user is active again after watching the video, i.e. the time interval between the second historical moment and the first historical moment. The limiting interval in the example is only counted to 7 days at maximum, and can be adjusted according to specific scenes. Here, event and Duration are history active information. When Event is 1, duration is the time interval between the second history time and the first history time, 3 days in the example of table 1; when Event is 0, the Duration field is 0.

Features represent behavioral characteristics such as user attributes, user portraits, and user behaviors before the first historical time.

After the data in table 1 are obtained, for each piece of data, the Cox model is trained by taking Feature as input and Duration as output, and when each parameter in the Cox model converges, the Cox is trained. The trained Cox model can be referred to as the M1 model.

2. Model prediction: the probability of existence model M1 trained in the previous step is used and the target user account (data (i.e. first historical behavior feature) is read in, without Duration (model label) data, in the form as shown in table 2.

TABLE 2

The fields in table 2 are explained as follows.

User_id is a unique identification of the User, such as a User account.

Event indicates the occurrence of an Event, here whether the user is active.

Feature represents behavior characteristics such as user attributes, user portraits, and user behaviors before the current time.

3. And (3) outputting a prediction result: after the video is output and played in the previous step, the survival probability of each user account in each day within 7 days in the future can be accordingly drawn into a first survival curve, such as a dotted line (which may be referred to as a priori survival probability curve) in fig. 4, and a survival curve area S1 enclosed by the first survival curve and the abscissa and ordinate axes is calculated.

(2) And correcting the first survival curve based on the actual active data of the user to obtain a corrected second survival curve (corrected survival probability curve in fig. 4), and calculating the survival curve area enclosed by the second survival curve and the axis of abscissa and ordinate (S2).

The specific process is as follows.

1. A first active day interval (date_diff) is calculated from the time the video is played by the user client logged into the user account until the video is played. Assuming that a user client (e.g., video APP) logging into a user account logs into the user client again on day 3 after playing video, the first active day interval for that user account is considered to be 3 days.

2. And correcting the first survival curve (priori survival probability curve) predicted by the model, and calculating the survival curve area S2 enclosed by the second survival curve and the abscissa axis. The correction process is specifically as follows.

Based on the trained survival probability model M1, the first survival prediction probabilities of the user account for each day in the future 7 days are predicted to be P1, P2, P3,..once, pm,..once, P7, respectively. The user account is active on day 3, then the first survival curve is modified using the first survival prediction probability P3 (i.e., the second survival prediction probability) for the user on day 3. The ratio of each first survival prediction probability to each second survival prediction probability is obtained as follows: P1/P3, P2/P3,..p 3/p3=1,..p, P7/P3), the respective ratios obtained were used as survival probability values after correction of the first survival prediction probability. Since the maximum value of the survival probability does not exceed 1, the corrected probability values are subjected to threshold pruning, and the corrected survival probability values are 1,., P3/p3=1,., and P7/P3, respectively.

(3) And measuring the liveness gain of the user after watching the video (namely after the user client side logging in the user account plays the video) by using the area difference of S2 and S1, and calculating the influence of the video dimension Embedding on the gain by using a tree model (Xgboost, lightgbm).

The specific process is as follows.

1. For each user account, calculating the area difference between the living curve area S2 of the second living curve and the curve area S1 of the first living curve of the user account, wherein the area difference is taken as the activity gain of the user account.

2. For each user account, inputting the Embedding characteristics of the played video and the Embedding characteristics of the user behaviors as tree models Featurs, and the liveness gain as the label of the tree models to obtain a trained model (M2). The data format used for training the model is shown in table 3.

TABLE 3 Table 3

The fields in table 3 are explained as follows.

User_id is a unique identification of the User, such as a User account.

Video Embedding is a representation that abstracts video into multi-dimensional feature vectors.

User behavior Embedding is a representation that abstracts user behavior features into multidimensional feature vectors.

The liveness gain is the difference between S2 and S1.

(4) And calculating the influence weights and the influence directions of the user activities of the various Embedding of the video on the Embedding of different user behaviors through the SHAP module.

The specific process is as follows.

1. And inputting the model M2 and the characteristics of the user behavior Embedding and the video Embedding after the label is removed into the SHAP module.

2. The importance and the influence direction of each Embedding are calculated through the TreeExplatiner and shape_values method of the SHAP module.

The above is the main process of identifying video and/or user behavior references affecting user activity by using the scheme provided by the embodiment of the present application. Fig. 5a and 5b are schematic diagrams of the importance and the influence direction of each dimension calculated in this embodiment, and the video for improving the activity of the user group can be recommended to different user groups through the result. The larger the value, the more obvious the influence of each dimension feature on the liveness gain, the more obvious the liveness gain of the user account is, the larger the value is, the larger the size of the scattered points represents the size of the corresponding feature value, and the larger the feature value is, as can be seen from fig. 5a, the positive correlation of the emb27 (namely the video enhancement 27) on the liveness gain of the user account is. In fig. 5b, the abscissa indicates the direction of influence of the video embedded 39 on the liveness gain, and the ordinate indicates the degree of influence of the video embedded 39 on the liveness gain of the user account, for example: the video vector value of the emb39 (i.e. the Embedding 39) is obviously inversely related to the activity gain of the user account, and the video with larger emb39 value can be reduced or filtered in the video recommendation and screening.

The embodiment is used for searching what service (such as video) can better promote the liveness of which user accounts, and has the following beneficial effects:

reflecting the influence of services (such as videos) with different characteristics on users with different characteristics, introducing services (such as videos) and the user behavior Embedding as model characteristics, determining the favorite services Embedding of the user group of each user behavior Embedding, and determining the influence of different service characteristics and/or behavior characteristics on the activity of the users so as to achieve the aim of personalized service screening for the users. According to the embodiment of the application, high-quality services (such as videos) can be screened in batches, resource consumption is reduced, for example, service features and behavior features are abstracted into vectors Embedding, high-quality contents can be searched for users with the same behavior features in batches, screening cost of the services (such as videos) is reduced, and meanwhile, service features and user behavior features with higher dimensionality do not need to be calculated, so that a large amount of computing resources are reduced. The embodiment of the application can meet the core requirements of the short video industry on improving the activity of the user, and from the perspective of improving the activity of the user, the theory of video screening is ensured to be matched with the requirement of the user on the video, so that a larger practical application space is provided; in addition, the embodiment of the application measures the liveness gain through the survival curve area, converts discrete data into continuous data, increases the distinguishing degree of the data, has higher sensitivity, and can capture the influence of services (such as video) on users more sensitively.

In conclusion, the embodiment can be widely applied to the field of Internet (such as the field of video), helps each APP to screen the business Embedding for improving the activity of the user in batches, and has wide application prospect.

It should be noted that the steps illustrated in the flowcharts of the figures may be performed in a computer system such as a set of computer executable instructions, and that although a logical order is illustrated in the flowcharts, in some cases the steps illustrated or described may be performed in an order other than that illustrated herein.

The embodiment of the application provides a data processing device which can be used for executing the data processing method of the embodiment of the application.

FIG. 6 is a schematic diagram of a data processing apparatus according to an embodiment of the present application, as shown in FIG. 6, the apparatus including:

a first obtaining unit 10 configured to perform obtaining a first historical behavior feature of the user account before the current moment;

an input unit 20 configured to perform inputting a first historical behavior feature to a pre-trained survival probability model, obtain a first survival prediction probability of each prediction time period of the user account in the prediction period output by the survival probability model, and obtain a first survival curve of the user account in the prediction period based on the first survival prediction probability; the survival probability model is obtained by taking a second historical behavior characteristic of the sample user account as input and historical activity information of the sample user account as output training, wherein the second historical behavior characteristic is the historical behavior characteristic of the sample user account before the first historical moment, the historical activity information comprises a time interval between the second historical moment and the first historical moment, and the second historical moment is the first active historical moment of the sample user account in a preset time period after the first historical moment;

A second obtaining unit 30 configured to obtain an active time of the first access service of the user account after the current time, and correct the first survival curve by using the active time to obtain a second survival curve of the user account in the prediction period;

the determining unit 40 is configured to determine an activity gain of the user account based on the difference between the first survival curve and the second survival curve, and determine a degree and a direction of influence of the first historical behavior feature and/or the traffic feature of the traffic on the activity gain.

This embodiment employs a first acquisition unit 10 configured to perform acquisition of a first historical behavior feature of the user account prior to the current moment; an input unit 20 configured to perform inputting a first historical behavior feature to a pre-trained survival probability model, obtain a first survival prediction probability of each prediction time period of the user account in the prediction period output by the survival probability model, and obtain a first survival curve of the user account in the prediction period based on the first survival prediction probability; the survival probability model is obtained by taking a second historical behavior characteristic of the sample user account as input and historical activity information of the sample user account as output training, wherein the second historical behavior characteristic is the historical behavior characteristic of the sample user account before the first historical moment, the historical activity information comprises a time interval between the second historical moment and the first historical moment, and the second historical moment is the first active historical moment of the sample user account in a preset time period after the first historical moment; a second obtaining unit 30 configured to obtain an active time of the first access service of the user account after the current time, and correct the first survival curve by using the active time to obtain a second survival curve of the user account in the prediction period; the determining unit 40 is configured to determine the activity gain of the user account based on the difference between the first survival curve and the second survival curve, and determine the influence degree and the influence direction of the first historical behavior feature and/or the business feature of the business on the activity gain, so as to solve the problem that the content cannot be known to improve the activity of the user when the content is pushed to the user in the related art, and further achieve the effect of determining the influence of different business features on the activity of the user to effectively improve the activity of the user.

Alternatively, the determining unit 40 includes: the training module is configured to execute training tree models by taking first historical behavior characteristics and business characteristics of business as input and activity gain as output; the determining module is configured to execute the interpretation training tree model and determine the influence degree and the influence direction of the first historical behavior characteristic and/or the business characteristic of the business on the activity gain.

Optionally, the training module is configured to perform training of the tree model with the dimension feature of the first historical behavior and the dimension feature of the business as input and the liveness gain as output, and the determining module is configured to perform interpretation of the trained tree model and determine the influence degree and the influence direction of the dimension feature of the first historical behavior and/or the dimension feature of the business on the liveness gain.

Optionally, the determining module is configured to perform: and inputting the trained tree model, the first historical behavior characteristic and the business characteristic of the business into the model interpreter, and acquiring the influence degree and the influence direction of the first historical behavior characteristic and/or the business characteristic of the business output by the model interpreter on the activity gain.

Optionally, the second acquisition unit includes: a first acquisition module configured to perform acquisition of a first lifetime prediction probability of the user account for each prediction period within the prediction period; a second obtaining module configured to obtain a second survival prediction probability of a prediction period to which the user account belongs at the active time; a correction module configured to perform correction of each first survival prediction probability using the second survival prediction probabilities; and a third acquisition module configured to perform obtaining a second survival curve based on the corrected first survival prediction probability.

Optionally, the apparatus further comprises: the first pushing unit is configured to determine the activity gain of the user account based on the difference between the first survival curve and the second survival curve, and after determining the influence degree and the influence direction of the first historical behavior feature and/or the business feature of the business on the activity gain, push the business including the business feature to the user account corresponding to the first historical behavior feature under the condition that the influence direction of the business feature of the first historical behavior feature and/or the business on the activity gain is positive; the first pushing unit is further configured to perform: and under the condition that the influence direction of the service characteristics of the first historical behavior characteristics and/or the service characteristics of the service on the activity gain is negative, reducing pushing the service comprising the service characteristics to the user account corresponding to the first historical behavior characteristics.

Optionally, the apparatus further comprises: a second pushing unit configured to perform pushing the service including the service feature if the influence direction of the service feature of the service on the activity gain is a positive direction after determining the activity gain of the user account based on the difference between the first survival curve and the second survival curve and determining the influence degree and the influence direction of the first historical behavior feature and/or the service feature of the service on the activity gain; the second pushing unit is further configured to perform: pushing a service to a client which logs in a user account when the influence direction of the first historical behavior characteristic on the activity gain is a positive direction; the second pushing unit is further configured to perform: and pushing the business comprising the business characteristics to the client logging in the user account when the influence direction of the business characteristics of the first historical behavior characteristics and the business on the activity gain is positive.

Optionally, the apparatus further comprises: a third pushing unit configured to perform pushing the service including the service feature when the degree of influence of the service feature of the service on the activity gain is greater than a first preset degree and the direction of influence is a positive direction after determining the activity gain of the user account based on the difference between the first survival curve and the second survival curve and determining the degree of influence and the direction of influence of the first historical behavior feature and/or the service feature of the service on the activity gain; the third pushing unit is further configured to perform pushing of the service to the client logged in the user account when the degree of influence of the first historical behavior feature on the activity gain is greater than a second preset degree and the influence direction is a positive direction; the third pushing unit is further configured to perform pushing the service including the service feature to the client logged in the user account when the first historical behavior feature and the service feature of the service have a positive influence on the liveness gain, and the service feature of the service has a degree of influence on the liveness gain greater than a first preset degree, and the first historical behavior feature has a degree of influence on the liveness gain greater than a second preset degree.

Optionally, the apparatus further comprises: a fourth pushing unit configured to perform pushing to reduce the service including the service feature when the influence direction of the service feature of the service is a negative direction after determining the liveness gain of the user account based on the difference between the first survival curve and the second survival curve and determining the influence degree and the influence direction of the first historical behavior feature and/or the service feature of the service on the liveness gain; the fourth pushing unit is further configured to perform pushing of the service to the client logged in the user account when the direction of influence of the first historical behavior feature on the liveness gain is negative; the fourth pushing unit is further configured to perform pushing of the service including the service feature to the client logging in the user account is reduced when the direction of influence of the service feature of the service and the first historical behavior feature is negative.

Optionally, the apparatus further comprises: a fifth pushing unit configured to perform pushing the service including the service feature when the influence degree of the service feature of the service on the activity gain is greater than a third preset degree and the influence direction is a negative direction after determining the activity gain of the user account based on the difference between the first survival curve and the second survival curve and determining the influence degree and the influence direction of the first historical behavior feature and/or the service feature of the service on the activity gain; the fifth pushing unit is further configured to perform pushing of the service to the client logged in the user account when the degree of influence of the first historical behavior feature on the liveness gain is greater than a fourth preset degree and the influence direction is a negative direction; the fifth pushing unit is further configured to perform pushing the service including the service feature to the client logged in the user account when the first historical behavior feature and the service feature of the service affect the activity gain in a negative direction, and the service feature of the service affects the activity gain to a degree greater than a third preset degree, and the first historical behavior feature affects the activity gain to a degree greater than a fourth preset degree.

The data processing device comprises a processor and a memory, wherein the first acquisition unit, the input unit, the correction unit and the like are all stored in the memory as program units, and the processor executes the program units stored in the memory to realize corresponding functions.

The processor includes a kernel, and the kernel fetches the corresponding program unit from the memory. The kernel can be provided with one or more than one kernel, and the user activity is effectively improved by adjusting the kernel parameters.

The memory may include volatile memory, random Access Memory (RAM), and/or nonvolatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM), among other forms in computer readable media, the memory including at least one memory chip.

The embodiment of the application provides a storage medium having a program stored thereon, which when executed by a processor, implements the data processing method.

The embodiment of the application provides a processor for running a program, wherein the data processing method is executed when the program runs.

An embodiment of the present application provides an apparatus, where fig. 7 is a schematic diagram of the apparatus of the present embodiment, and as shown in fig. 7, the apparatus includes at least one processor, and at least one memory and a bus connected to the processor; the processor and the memory complete communication with each other through a bus; the processor is used for calling the program instructions in the memory to execute the data processing method. The device herein may be a server, PC, PAD, cell phone, etc.

The present application also provides a computer program product adapted to perform, when executed on a data processing device, a program initialized with the method steps of: acquiring a first historical behavior characteristic of a user account before the current moment; inputting a first historical behavior characteristic into a pre-trained survival probability model, acquiring a first survival prediction probability of a user account in each prediction time period in a prediction period output by the survival probability model, and acquiring a first survival curve of the user account in the prediction period based on the first survival prediction probability; acquiring the active time of the user account for accessing the service for the first time after the current time, and correcting the first survival curve by utilizing the active time to acquire a second survival curve of the user account in the prediction period; and determining the activity gain of the user account based on the difference between the first survival curve and the second survival curve, and determining the influence degree and the influence direction of the first historical behavior characteristic and/or the business characteristic of the business on the activity gain.

It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

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

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

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

In one typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include volatile memory in a computer-readable medium, random Access Memory (RAM) and/or nonvolatile memory, etc., such as Read Only Memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

Computer readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of storage media for a computer include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium, which can be used to store information that can be accessed by a computing device. Computer-readable media, as defined herein, does not include transitory computer-readable media (transmission media), such as modulated data signals and carrier waves.

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

It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and changes may be made to the present application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. which are within the spirit and principles of the present application are intended to be included within the scope of the claims of the present application.

Claims

1. A method of data processing, comprising:

acquiring a first historical behavior characteristic of a user account before the current moment;

inputting the first historical behavior characteristic to a pre-trained survival probability model, acquiring a first survival prediction probability of the user account in each prediction time period in a prediction period output by the survival probability model, and acquiring a first survival curve of the user account in the prediction period based on the first survival prediction probability; the survival probability model is obtained by taking a second historical behavior characteristic of a sample user account as input and historical activity information of the sample user account as output training, wherein the second historical behavior characteristic is a historical behavior characteristic of the sample user account before a first historical moment, the historical activity information comprises a time interval between the second historical moment and the first historical moment, and the second historical moment is a first active historical moment of the sample user account in a preset time period after the first historical moment;

acquiring the active time of the user account for accessing the service for the first time after the current time, and correcting the first survival curve by utilizing the active time to acquire a second survival curve of the user account in the prediction period;

And determining the activity gain of the user account based on the difference between the first survival curve and the second survival curve, and determining the influence degree and the influence direction of the first historical behavior characteristic and/or the business characteristic of the business on the activity gain.

2. The method according to claim 1, wherein said determining the extent and direction of influence of the first historical behavior feature and/or the traffic feature of the traffic on the liveness gain comprises:

training a tree model by taking the first historical behavior characteristic and the service characteristic of the service as inputs and taking the liveness gain as output;

and explaining the trained tree model, and determining the influence degree and the influence direction of the first historical behavior characteristic and/or the business characteristic of the business on the liveness gain.

3. The method of claim 2, wherein the step of determining the position of the substrate comprises,

taking the first historical behavior characteristic and the service characteristic of the service as input and the liveness gain as output, training a tree model comprising: training the tree model by taking the dimension characteristics of the first historical behaviors and the business dimension characteristics of the business as inputs and the liveness gain as output,

Interpreting the trained tree model, and determining the influence degree and the influence direction of the first historical behavior feature and/or the business feature of the business on the liveness gain comprises: and explaining the trained tree model, and determining the influence degree and the influence direction of the dimension characteristics of the first historical behaviors and/or the service dimension characteristics of the service on the liveness gain.

4. Method according to claim 2, characterized in that interpreting the trained tree model, determining the extent and direction of influence of the first historical behavior feature and/or the traffic feature of the traffic on the liveness gain, comprises:

and inputting the trained tree model, the first historical behavior feature and the business feature of the business into a model interpreter, and acquiring the influence degree and the influence direction of the first historical behavior feature and/or the business feature of the business output by the model interpreter on the liveness gain.

5. The method of claim 1, wherein said modifying said first survival curve with said active time to obtain a second survival curve for said user account during said forecast period comprises:

Acquiring a first lifetime prediction probability of each prediction time period of the user account in a prediction period;

acquiring a second survival prediction probability of the user account in a prediction time period to which the active moment belongs;

correcting each first survival prediction probability by using the second survival prediction probability;

and obtaining the second survival curve based on the corrected first survival prediction probability.

6. The method of claim 1, wherein after determining an liveness gain of the user account and determining a degree and direction of impact of the first historical behavioral characteristic and/or the business characteristic of the business on the liveness gain based on a difference between the first and second survival curves, the method further comprises:

adding pushing of the business comprising the business feature to the user account corresponding to the first historical behavior feature under the condition that the influence direction of the business feature of the first historical behavior feature and/or the business on the activity gain is positive;

and under the condition that the influence direction of the first historical behavior feature and/or the business feature of the business on the activity gain is negative, reducing pushing the business comprising the business feature to the user account corresponding to the first historical behavior feature.

7. The method according to claim 1, wherein after determining an liveness gain of the user account based on the difference between the first and second survival curves and determining a degree and direction of influence of the first historical behavior feature and/or the traffic feature of the traffic on the liveness gain, the method further comprises:

if the influence direction of the service characteristics of the service on the liveness gain is a positive direction, pushing the service comprising the service characteristics;

if the influence direction of the first historical behavior characteristic on the liveness gain is a positive direction, pushing a service to a client logging in the user account;

and pushing the service comprising the service characteristics to a client logging in the user account if the influence direction of the first historical behavior characteristics and the service characteristics of the service on the activity gain is positive.

8. The method according to claim 1, wherein after determining an liveness gain of the user account based on the difference between the first and second survival curves and determining a degree and direction of influence of the first historical behavior feature and/or the traffic feature of the traffic on the liveness gain, the method further comprises:

If the influence degree of the service characteristics of the service on the activity gain is greater than a first preset degree and the influence direction is a positive direction, pushing the service comprising the service characteristics;

if the influence degree of the first historical behavior characteristic on the activity gain is larger than a second preset degree and the influence direction is a positive direction, pushing a service to a client side logging in the user account;

and if the influence directions of the first historical behavior feature and the business feature of the business on the liveness gain are positive directions, the influence degree of the business feature of the business on the liveness gain is larger than a first preset degree, and the influence degree of the first historical behavior feature on the liveness gain is larger than a second preset degree, pushing the business comprising the business feature to a client side logging in the user account.

9. The method according to claim 1, wherein after determining an liveness gain of the user account based on the difference between the first and second survival curves and determining a degree and direction of influence of the first historical behavior feature and/or the traffic feature of the traffic on the liveness gain, the method further comprises:

If the influence direction of the service characteristics of the service on the activity gain is a negative direction, reducing pushing the service comprising the service characteristics;

if the influence direction of the first historical behavior characteristic on the liveness gain is a negative direction, pushing business to a client side logging in the user account is reduced;

and if the influence direction of the first historical behavior characteristic and the business characteristic of the business on the activity gain is a negative direction, pushing the business comprising the business characteristic to a client logging in the user account is reduced.

10. The method according to claim 1, wherein after determining an liveness gain of the user account based on the difference between the first and second survival curves and determining a degree and direction of influence of the first historical behavior feature and/or the traffic feature of the traffic on the liveness gain, the method further comprises:

if the influence degree of the service characteristics of the service on the activity gain is greater than a third preset degree and the influence direction is a negative direction, reducing pushing the service comprising the service characteristics;

if the influence degree of the first historical behavior characteristic on the liveness gain is larger than a fourth preset degree and the influence direction is a negative direction, pushing services to a client side logging in the user account is reduced;

And if the influence direction of the first historical behavior feature and the business feature of the business on the liveness gain is a negative direction, the influence degree of the business feature of the business on the liveness gain is larger than a third preset degree, and the influence degree of the first historical behavior feature on the liveness gain is larger than a fourth preset degree, pushing the business comprising the business feature to a client side logging in the user account is reduced.

11. A data processing apparatus, comprising:

a first acquisition unit configured to perform acquisition of a first historical behavior feature of the user account before the current time;

the input unit is configured to input the first historical behavior characteristic to a pre-trained survival probability model, acquire first survival prediction probabilities of the user account in each prediction time period in a prediction period output by the survival probability model, and acquire a first survival curve of the user account in the prediction period based on the first survival prediction probabilities; the survival probability model is obtained by taking a second historical behavior characteristic of a sample user account as input and historical activity information of the sample user account as output training, wherein the second historical behavior characteristic is a historical behavior characteristic of the sample user account before a first historical moment, the historical activity information comprises a time interval between the second historical moment and the first historical moment, and the second historical moment is a first active historical moment of the sample user account in a preset time period after the first historical moment;

A second obtaining unit configured to obtain an active time of the first access service of the user account after the current time, and correct the first survival curve by using the active time, so as to obtain a second survival curve of the user account in the prediction period;

and the determining unit is configured to determine the activity gain of the user account based on the difference between the first survival curve and the second survival curve, and determine the influence degree and the influence direction of the first historical behavior characteristic and/or the business characteristic of the business on the activity gain.

12. The apparatus according to claim 11, wherein the determining unit comprises:

the training module is configured to execute training a tree model by taking the first historical behavior characteristic and the service characteristic of the service as input and taking the liveness gain as output;

and the determining module is configured to execute an interpretation training tree model and determine the influence degree and the influence direction of the first historical behavior characteristic and/or the business characteristic of the business on the liveness gain.

13. The apparatus of claim 12, wherein the device comprises a plurality of sensors,

the training module is configured to perform training the tree model with the dimension characteristic of the first historical behavior, the business dimension characteristic of the business as input, the liveness gain as output,

The determining module is configured to execute the trained tree model and determine the influence degree and the influence direction of the dimension characteristics of the first historical behavior and/or the service dimension characteristics of the service on the liveness gain.

14. The apparatus of claim 12, wherein the determination module is configured to perform:

15. The apparatus of claim 11, wherein the second acquisition unit comprises:

a first acquisition module configured to perform acquisition of a first lifetime prediction probability for each prediction period of the user account over a prediction period;

a second obtaining module configured to obtain a second survival prediction probability of the user account in a prediction period to which the active time belongs;

a correction module configured to perform correction of each of the first survival prediction probabilities using the second survival prediction probabilities;

And a third acquisition module configured to perform obtaining the second survival curve based on the corrected first survival prediction probability.

16. The apparatus of claim 11, wherein the apparatus further comprises:

a first pushing unit configured to perform pushing of a service including the service feature to a user account corresponding to the first historical behavior feature, where the first historical behavior feature and/or the service feature of the service has a positive direction of the influence on the activity gain after determining the activity gain of the user account based on the difference between the first survival curve and the second survival curve and determining the influence degree and the influence direction of the first historical behavior feature and/or the service feature of the service on the activity gain;

the first pushing unit is further configured to perform: and under the condition that the influence direction of the first historical behavior feature and/or the business feature of the business on the activity gain is negative, reducing pushing the business comprising the business feature to the user account corresponding to the first historical behavior feature.

17. The apparatus of claim 11, wherein the apparatus further comprises:

A second pushing unit configured to perform pushing a service including the service feature if an influence direction of the service feature of the service on the liveness gain is a positive direction after determining an liveness gain of the user account based on a difference between the first survival curve and the second survival curve and determining an influence degree and an influence direction of the first historical behavior feature and/or the service feature of the service on the liveness gain;

the second pushing unit is further configured to perform: pushing a service to a client logging in the user account when the influence direction of the first historical behavior characteristic on the liveness gain is a positive direction;

the second pushing unit is further configured to perform: and pushing the business comprising the business characteristics to a client logging in the user account when the influence direction of the first historical behavior characteristics and the business characteristics of the business on the activity gain is positive.

18. The apparatus of claim 11, wherein the apparatus further comprises:

a third pushing unit configured to perform pushing a service including the service feature when the degree of influence of the service feature of the service on the activity gain is greater than a first preset degree and the direction of influence is a positive direction after determining the activity gain of the user account based on the difference between the first survival curve and the second survival curve and determining the degree of influence and the direction of influence of the service feature of the first historical behavior feature and/or the service feature of the service on the activity gain;

The third pushing unit is further configured to perform pushing a service to a client logged in the user account when the degree of influence of the first historical behavior feature on the liveness gain is greater than a second preset degree and the direction of influence is a positive direction;

the third pushing unit is further configured to perform pushing, to a client logged in to the user account, a service including the service feature when the first historical behavior feature and the service feature of the service affect the liveness gain in a positive direction, and the service feature of the service affects the liveness gain to a degree greater than a first preset degree, and the first historical behavior feature affects the liveness gain to a degree greater than a second preset degree.

19. The apparatus of claim 11, wherein the apparatus further comprises:

a fourth pushing unit configured to perform, after determining an activity gain of the user account based on a difference between the first survival curve and the second survival curve and determining an influence degree and an influence direction of the first historical behavior feature and/or a service feature of the service on the activity gain, when the influence direction of the service feature of the service on the activity gain is a negative direction, to reduce pushing the service including the service feature;

The fourth pushing unit is further configured to perform pushing services to a client logged in the user account when the direction of influence of the first historical behavior feature on the liveness gain is a negative direction;

the fourth pushing unit is further configured to perform pushing of the service including the service feature to the client logging in the user account when the first historical behavior feature and the service feature of the service affect the liveness gain in a negative direction.

20. The apparatus of claim 11, wherein the apparatus further comprises:

a fifth pushing unit configured to perform, after determining an activity gain of the user account based on a difference between the first survival curve and the second survival curve and determining an influence degree and an influence direction of the first historical behavior feature and/or a service feature of the service on the activity gain, when the influence degree of the service feature of the service on the activity gain is greater than a third preset degree and the influence direction is a negative direction, to reduce pushing the service including the service feature;

the fifth pushing unit is further configured to perform pushing services to clients logging in the user account when the degree of influence of the first historical behavior feature on the liveness gain is greater than a fourth preset degree and the influence direction is a negative direction;

The fifth pushing unit is further configured to perform pushing the service including the service feature to the client logged in the user account when the first historical behavior feature and the service feature of the service affect the liveness gain in a negative direction, and the service feature of the service affects the liveness gain to a degree greater than a third preset degree, and the first historical behavior feature affects the liveness gain to a degree greater than a fourth preset degree.

21. An electronic device, comprising:

a processor;

a memory for storing the processor-executable instructions;

wherein the processor is configured to execute the instructions to implement the data processing method of any one of claims 1 to 10.

22. A computer-readable storage medium, characterized in that the computer-readable storage medium comprises a stored program, wherein the data processing method of any one of claims 1 to 10 is performed when the program is executed by a processor.

23. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, performs the data processing method of any of claims 1 to 10.