Data processing method, device, computer-readable storage medium, and computer equipment

Abstract

The embodiments of the present invention disclose a data processing method, apparatus, computer-readable storage medium, and computer equipment. The method involves obtaining attribute data of multiple users in an application; counting usage of a target function within a preset time period; dividing users into a first group and a second group based on usage; determining a preference score for each user to use the target function; determining a third group that matches the first group from the second group based on the preference score; obtaining the operation data of the first and third groups on the application after a preset time period, and determining the incremental value of using the target function to the application accordingly. In this way, the experimental group and the control group are matched based on the preference scores of users using the target function, and then the incremental value of the target function to the application is determined based on the operation data of the two groups on the application after the intervention period. This method can improve data processing efficiency, thereby improving the efficiency and accuracy of calculating the incremental value of the target function to the application.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a data processing method, apparatus, computer-readable storage medium, and computer equipment. Background Art

[0002] With the rapid development of Internet technology and mobile terminal technology, Internet products are becoming increasingly abundant. At present, Internet products have become an indispensable tool in people's daily life.

[0003] In order to meet people's growing demand for different functions of Internet products, Internet products need to be continuously iterated and updated to launch new product functions.

[0004] However, evaluating the incremental value of new features introduced by Internet products has always been a difficult problem in the industry, and existing evaluation methods are inefficient and inaccurate. Summary of the Invention

[0005] The embodiments of the present application provide a data processing method, apparatus, computer-readable storage medium, and computer equipment. The method can improve the efficiency of data processing, and thereby improve the efficiency of calculating the incremental value generated by the target function for the target application.

[0006] A first aspect of the present application provides a data processing method, the method comprising:

[0007] Get attribute data of a preset number of users in the target application;

[0008] Counting usage of the target function in the target application by the preset number of users within a preset time period;

[0009] dividing the preset number of users into a first user group and a second user group according to the usage situation, wherein the first user group is a group consisting of users who have used the target function;

[0010] determining a preference score for each user to use the target function based on the attribute data and the usage situation;

[0011] Calculating the difference between the preference score of each user in the second user group and the preference score of each user in the first user group, and determining the users in the second group whose difference is less than a preset threshold, to obtain a third user group;

[0012] Operation data of the first user group and the third user group on the target application after the preset time period is obtained, and an incremental value of using the target function for the target application is determined based on the operation data.

[0013] Accordingly, a second aspect of the present application provides a data processing device, comprising:

[0014] An acquisition unit, configured to acquire attribute data of a preset number of users in a target application;

[0015] A statistics unit, configured to count usage of the target function in the target application by the preset number of users within a preset time period;

[0016] a dividing unit, configured to divide the preset number of users into a first user group and a second user group according to the usage situation, wherein the first user group is a group consisting of users who have used the target function;

[0017] a first determining unit, configured to determine a preference score for each user to use the target function based on the attribute data and the usage situation;

[0018] a calculation unit, configured to calculate a difference between a preference score of each user in the second user group and a preference score of each user in the first user group, and determine users in the second group whose difference is less than a preset threshold, to obtain a third user group;

[0019] The second determining unit is configured to obtain operation data of the first user group and the third user group on the target application after the preset time period, and determine an incremental value of using the target function for the target application based on the operation data.

[0020] In some embodiments, the first determining unit includes:

[0021] A first determining subunit, configured to determine tag data corresponding to each user according to the usage situation;

[0022] A training subunit, configured to train a logistic regression model using attribute data of each user as input and label data of each user as output, thereby obtaining a trained logistic regression model;

[0023] The second determining subunit is configured to determine a preference score corresponding to each user based on the trained logistic regression model.

[0024] In some embodiments, the training subunit includes:

[0025] A training module, configured to train multiple logistic regression models using attribute data of each user as input and label data of each user as output, thereby obtaining multiple trained logistic regression models;

[0026] The second determining unit is further configured to:

[0027] Determining a plurality of transition preference scores corresponding to each user based on the plurality of trained logistic regression models;

[0028] The average of the multiple transition preference scores corresponding to each user is calculated to obtain the preference score corresponding to each user.

[0029] In some embodiments, the second determining unit includes:

[0030] A first acquiring subunit is configured to acquire first login data of the first user group and the third user group logging into the target application within a first target time period after the preset time period;

[0031] a second acquiring subunit, configured to acquire second login data of the first user group and the third user group logging into the target application within a second target time period after the first target time period;

[0032] a first calculation subunit, configured to calculate a first retention rate corresponding to the first group and a second retention rate corresponding to the third group based on the first login data and the second login data;

[0033] The second calculation subunit is configured to calculate an incremental value of using the target function for the target application based on the first retention rate and the second retention rate.

[0034] In some embodiments, the second determining unit includes:

[0035] a detection subunit, configured to perform homogeneity detection on the first user group and the third user group;

[0036] The third determination subunit is used to obtain the operation data of the first user group and the third user group on the target application after the preset time period when the result of the homogeneity detection meets the preset conditions, and determine the incremental value of using the target function for the target application based on the operation data.

[0037] In some embodiments, the detection subunit comprises:

[0038] A first acquisition module is configured to acquire third login data of the first user group and the third user group logging into the target application within a third target time period before the preset time period;

[0039] a second acquisition module, configured to acquire fourth login data of the first user group and the third user group in a fourth target time period before the third target time period;

[0040] The first detection module is configured to perform homogeneity detection on the first user group and the third user group according to the third login data and the fourth login data.

[0041] In some embodiments, the detection subunit comprises:

[0042] a detection module, configured to perform a balance check on the attribute data of users in the first user group and the third user group;

[0043] The second detection module is configured to perform a homogeneity detection on the first user group and the third user group when the balance check result is qualified.

[0044] In some embodiments, the inspection module includes:

[0045] a first calculation submodule, configured to calculate a first mean of the attribute data of each user in the first user group, a second mean of the attribute data of each user in the second user group, and a standard deviation of the attribute data of each user in the first user group;

[0046] A second calculation submodule, configured to calculate a difference between the first mean and the second mean;

[0047] a third calculation submodule, configured to calculate a ratio of the difference to the standard deviation;

[0048] The determination submodule is used to determine that the result of the balance check is qualified when the ratio is less than a preset value.

[0049] The third aspect of the present application further provides a computer-readable storage medium, which stores a plurality of instructions, and the instructions are suitable for loading by a processor to execute the steps of the data processing method provided in the first aspect of the present application.

[0050] The fourth aspect of the present application provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein when the processor executes the computer program, the steps of the data processing method provided in the first aspect of the present application are implemented.

[0051] In a fifth aspect, the present application provides a computer program product or computer program, comprising computer instructions stored in a storage medium. A processor of a computer device reads the computer instructions from the storage medium and executes the computer instructions, causing the computer device to perform the steps of the data processing method provided in the first aspect.

[0052] The data processing method provided in the embodiment of the present application obtains attribute data of a preset number of users in a target application; counts the usage of a target function in the target application by the preset number of users within a preset time period; divides the preset number of users into a first user group and a second user group based on the usage, wherein the first user group is a group composed of users who use the target function; determines a preference score for each user to use the target function based on the attribute data and the usage; calculates the difference between the preference score of each user in the second user group and the preference score of each user in the first user group, and determines the users in the second group whose difference is less than a preset threshold to obtain a third user group; obtains operation data of the first user group and the third user group on the target application after a preset time period, and determines the incremental value of using the target function to the target application based on the operation data. In this way, a third user group homogeneous with the first user group is determined from the second user group based on the preference score of the users using the target function, and then determines the incremental value of the target function to the target application based on the operation data of the first user group and the second user group on the target application after the intervention period. This method can effectively determine homogeneous control and experimental groups, improve data processing efficiency, and thus improve the efficiency and accuracy of calculating the incremental value of the target function to the target application. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0054] Figure 1 This is a schematic diagram of a data processing scenario in this application;

[0055] Figure 2 It is a flowchart of the data processing method provided by this application;

[0056] Figure 3 is another flow chart of the data processing method provided by this application;

[0057] Figure 4 It is the time map provided by this application;

[0058] Figure 5 It is a structural diagram of the data processing device provided by this application;

[0059] Figure 6 It is a structural diagram of the computer device provided in this application. DETAILED DESCRIPTION

[0060] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative efforts shall fall within the scope of protection of the present invention.

[0061] Embodiments of the present invention provide a data processing method, apparatus, computer-readable storage medium, and computer equipment. The data processing method can be used in a data processing apparatus. The data processing apparatus can be integrated into a computer device, which can be a terminal or a server. The terminal can be a mobile phone, tablet computer, laptop computer, smart TV, wearable smart device, personal computer (PC), or other device. The server can be an independent physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery network (CDN), and big data and artificial intelligence platforms. The server can be a node in a blockchain.

[0062] See also Figure 1 , which is a schematic diagram of a scenario for data processing provided by this application; as shown in the figure, computer device A obtains attribute data of a preset number of users in the target application from server B corresponding to the target application, and counts the usage of the target function by the target number of users within a preset time period; divides users into a first user group and a second user group according to usage, and determines the preference score of each user for using the target function based on the attribute data and usage; matches users in the second user group with the first user group according to the preference score to obtain a third user group that matches the first user group; obtains operation data of the first user group and the third user group after a preset time period, and determines the incremental value of the target function to the target application based on the operation data. Among them, the incremental value to the target application can be the gain caused to the target application, for example, it can be understood as an increase in user stickiness. Specifically, it can be an increase in the average number of daily logins of users or an increase in online time.

[0063] It should be noted that Figure 1The data processing scenario diagram shown is merely an example. The data processing scenario described in the embodiments of this application is intended to more clearly illustrate the technical solution of this application and does not constitute a limitation on the technical solution provided by this application. Those skilled in the art will appreciate that with the evolution of data processing and the emergence of new business scenarios, the technical solution provided by this application will also be applicable to similar technical problems.

[0064] The following are detailed descriptions based on the above implementation scenarios.

[0065] The embodiments of the present application will be described from the perspective of a data processing device, which can be integrated into a computer device. The computer device can be a terminal or a server. The terminal can be a mobile phone, a tablet computer, a laptop computer, a smart TV, a wearable smart device, a personal computer (PC) or other devices. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers. It can also be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, network acceleration services (Content Delivery Network, CDN), and big data and artificial intelligence platforms. Figure 2 FIG. 1 is a flow chart of a data processing method provided by the present application, which includes:

[0066] Step 101: Obtain attribute data of a preset number of users in a target application.

[0067] The target application can be any application installed on the terminal, or a mini-program that can be used without downloading or installing, simply by attaching it to some instant messaging application. Specifically, the target application includes but is not limited to instant messaging applications, internet shopping applications, video playback applications (including long and short videos), live streaming applications, social applications, and audio playback applications, as well as the mini-programs corresponding to each application.

[0068] The preset number of users in the target application can be users who have registered with the target application's corresponding server and can log in and access the target application's corresponding server through the target application client or target application mini-program installed on their terminal. In other words, users in the target application are users who have registered accounts for the target application and log in and use the target application with a certain frequency. This preset number can be set by the user based on evaluation requirements and is not limited here.

[0069] User attribute data may include basic personal information and data on dimensions such as user social activity. Basic personal information may include information such as the user's age, gender, education level, birthday, and hobbies. Social activity data may include data such as the number of days the user has logged into the target application, the number of logins, and the number of social interactions with other users. Both basic personal information and social activity data are obtained with the user's authorization. The data is encrypted during both acquisition and use to protect the user's privacy.

[0070] Among them, it is understandable that the attribute data of different users are different, which leads to different users having different usage tendencies for different applications of the target application. In this way, when a new function is updated in the target application, it is impossible to simply select an experimental group and a control group to evaluate the incremental value of the new function to the target application. Because in addition to the variable of whether the user uses the new function, the different attribute data of the users is also an uncontrollable variable, so it is impossible to evaluate the incremental value of the new function to the target application by the control variable method. In order to solve the above-mentioned technical problem that the control variable method cannot be used to evaluate the incremental value of the new function to the target application due to the different attribute data of the users, the present application provides a data processing method, which can effectively calculate the incremental value of the new function to the target application. Below, the data processing method provided by the present application is further introduced in detail.

[0071] When it is necessary to evaluate the incremental value of a new function in a target application to the target application, attribute data of a preset number of users in the target application may be obtained first, and further evaluation may be performed based on the obtained attribute data of the users.

[0072] Step 102 : Count usage of a target function in a target application by a preset number of users within a preset time period.

[0073] The preset time period is the period during which the target function has been added to the target application. For example, if a new function, such as a "like" function, is added to the personal status display page of an instant messaging application, users can use the "like" function on the personal status display page of the instant messaging application during this preset time period. Of course, users do not need to use the "like" function.

[0074] The specific length of the preset time period can be set as needed, for example, one day, three days, one week, or two weeks. During this preset time period, statistics are continuously collected on the usage of the newly added target function by a preset number of users. Usage can be divided into use and non-use. Specifically, for example, statistics show that out of 1,000 users of the target application, 100 used the like function within a week, and 900 did not.

[0075] Step 103 : Divide a preset number of users into a first user group and a second user group according to usage.

[0076] After obtaining the usage of the target function in the target application by a preset number of users within a preset time period, the users can be classified according to their usage of the target function. Users who have used the target function can be classified as a first category of users, forming a first user group; users who have not used the target function can be classified as a second category of users, forming a second user group.

[0077] Since the target function is a new function of the target application, the number of users using the new function is generally less than or much less than the number of users not using the target function. That is, in this embodiment of the application, the number of users in the first user group can be less than the number of users in the second user group.

[0078] Among them, since the first user group is a group composed of users who use the target function, that is, a group that has undergone changes, the first user group can be determined as the experimental group, and the preset time period can be called the intervention period.

[0079] Step 104 : Determine each user's preference score for using the target function based on the attribute data and usage.

[0080] After identifying the experimental group that has been affected by the target function, in order to evaluate the incremental value of the target function to the target application, it is necessary to further identify a control group that has not been affected by the target function. Furthermore, to ensure the accuracy of the evaluation, it is necessary to ensure that the control and experimental groups are homogeneous with respect to the target function before the intervention occurs. In other words, the experimental and control groups must have similar usage preferences for the target function. To do this, it is necessary to calculate a preference score for each user using the target function.

[0081] Specifically, in the embodiment of the present application, since the usage of the target function by each user during the preset time period (i.e., the aforementioned intervention period) has been obtained, and this usage is the actual usage data collected without other intervening factors, it can represent the usage preferences of different users for the target function. Therefore, a mapping relationship can be established based on each user's attribute data and the user's usage of the target function during the intervention period, and then the preference score for using the target application corresponding to different attribute data can be further determined based on this mapping relationship, thereby obtaining the preference score for each user using the target function.

[0082] In some embodiments, determining a preference score for each user to use a target function based on attribute data and usage includes:

[0083] 1. Determine the label data corresponding to each user based on usage;

[0084] 2. Using each user's attribute data as input and each user's label data as output, train a logistic regression model to obtain a trained logistic regression model;

[0085] 3. Based on the trained logistic regression model, determine the preference score corresponding to each user.

[0086] In this embodiment of the present application, since the usage of the target function by each user during the intervention period has been obtained, where the usage is a clear binary classification data of whether it is used or not, the label of each user can be determined based on the usage of the target function by each user during the intervention period: when the user uses the target function, the label is determined to be 1; when the user does not use the target function, the label is determined to be 0.

[0087] After determining each user's label data, a logistic regression model can be trained using each user's attribute information as the model input and the user's label information as the model output, thereby obtaining a trained logistic regression model. Specifically, when the user's attribute information is multi-dimensional, the logistic regression model can be a multivariate regression model. Specifically, each user's attribute data can be preprocessed to obtain a feature vector corresponding to each attribute data. Furthermore, each user's attribute data can be preprocessed using a word embedding method to obtain a word vector corresponding to each attribute data, that is, a feature vector corresponding to each user's attribute data.

[0088] After the trained logistic regression model is obtained, the feature vector corresponding to the attribute data of each user is input into the trained logistic regression model to obtain the numerical value corresponding to each user output by the trained logistic regression model. This numerical value is the preference score of each user for using the target function.

[0089] In some embodiments, the logistic regression model is trained using attribute data of each user as input and label data of each user as output to obtain a trained logistic regression model, including:

[0090] 2.1. Using each user's attribute data as input and each user's label data as output, train multiple logistic regression models to obtain multiple trained logistic regression models.

[0091] 2.2. Based on the trained logistic regression model, determine the preference score corresponding to each user, including:

[0092] 2.3. Determine multiple transition preference scores corresponding to each user based on multiple trained logistic regression models;

[0093] 2.4. Calculate the average of multiple transition preference scores corresponding to each user to obtain the preference score corresponding to each user.

[0094] In this embodiment of the present application, after determining the label data for each user, multiple logistic regression models can be trained using the user's attribute data and label data as training sample data to obtain multiple trained logistic regression models. Each user's attribute data can then be input into each trained logistic regression model one by one to obtain the numerical values ​​output by each trained logistic regression model. That is, each user corresponds to the numerical values ​​output by multiple trained logistic regression models. The average of the multiple numerical values ​​corresponding to each user is then calculated to obtain each user's preference score for using the target function.

[0095] In an embodiment of the present application, by using the user's attribute data and label data as training samples to train multiple logistic regression models, and calculating the preference score of each user for using the target function based on the average of the numerical values ​​output by the multiple logistic regression models, the problem of inaccurate calculation of the user's preference score for using the target function caused by the training bias of a single logistic regression model can be avoided, thereby improving the accuracy of calculating the incremental value of the target function to the target application.

[0096] Step 105 , calculating the difference between the preference score of each user in the second user group and the preference score of each user in the first user group, and determining the users in the second group whose difference is less than a preset threshold to obtain a third user group.

[0097] After calculating each user's preference score for the target function, we can further identify a control group of users who are similar to the experimental group. Specifically, we can search for users in the second user group who are similar to the first user group to form a third user group, the control group.

[0098] Specifically, for any target user in the first user group, first determine its corresponding target preference score, and then calculate the difference between the preference score of each user in the second user group and the target preference score of the target user. If the difference between the preference score of a user in the second group and the target preference score is less than a preset threshold, then it can be determined that the user is a user similar to the target user. Then, traverse each user in the first user group and find similar users in the second user group. These users similar to the users in the first user group constitute the third user group. Since the users in the third user group are all users selected from the second user group, the users in the third group did not use the target function during the intervention period, that is, the third user group can be called the control group.

[0099] Step 106 : Obtain operation data of the first user group and the third user group on the target application after a preset time period, and determine the incremental value of using the target function for the target application based on the operation data.

[0100] After determining the experimental and control groups, the incremental value of the target function to the target application can be evaluated based on the performance of the experimental and control groups in the target application after the intervention period. Specifically, the operational data of the experimental group (i.e., the aforementioned first user group) and the control group (i.e., the aforementioned third user group) on the target application after the intervention period (i.e., the aforementioned preset time period) can be obtained. The operational data can include login data, usage data, etc. Then, the incremental value of the target function to the target application can be determined based on the operational data of the experimental and control groups after the intervention period. Specifically, if the operational data of the experimental group after the intervention period shows a positive improvement compared to the operational data of the control group after the intervention period, it means that the target function has positive value to the target application. Specifically, the specific numerical value of the positive value can be calculated using some clear indicators. For example, the change in the average daily online time of the experimental and control groups before and after the intervention period, or the difference in the sub-weekly retention rate of the experimental and control groups after the intervention period, can be used to specifically calculate the specific numerical value of the incremental value of using the target function to the target application.

[0101] In some embodiments, obtaining operation data of the first user group and the third user group on the target application after a preset time period, and determining the incremental value of using the target function for the target application based on the operation data, includes:

[0102] 1. Obtaining first login data of the first user group and the third user group logging into the target application within a first target time period after a preset time period;

[0103] 2. Obtaining second login data of the first user group and the third user group logging into the target application within a second target time period after the first target time period;

[0104] 3. Calculate a first retention rate corresponding to the first group and a second retention rate corresponding to the third group based on the first login data and the second login data;

[0105] 4. Calculate the incremental value of using the target feature to the target application based on the first retention rate and the second retention rate.

[0106] Among them, in an embodiment of the present application, the incremental value of using the target function for the target application can be calculated based on the retention rate of the experimental group and the control group after the intervention period. Specifically, the first login data of the experimental group and the control group logging into the target application within the first target time period after the intervention period can be obtained first, wherein the first target time period can be 3 days, 5 days, one week or two weeks, etc., which can be set according to actual conditions. Then, the second login data of the experimental group and the control group within the second target time period after the first target time period is obtained. Among them, the time span of the second target time period is the same as the time span of the first target time period. Further, the first retention rate of the experimental group and the second retention rate of the control group are calculated based on the login data of the experimental group and the control group in the first target time period and the second target time period. Among them, when the first target time period and the second target time period are both one week, the first retention rate and the second retention rate can also be called sub-week retention rate. Specifically, the first retention rate is the proportion of the number of users in the experimental group who logged into the target application in the first target time period who logged into the target application in the second target time period to the former. For example, if the experimental group has 100 users, 80 of whom logged into the target application during the first target time period, and 70 of whom logged into the target application during the second target time period, the first retention rate is 87.5%. If the control group also has 100 users, 80 of whom logged into the target application during the first target time period, and 60 of whom logged into the target application during the second target time period, the second retention rate is 75%.

[0107] After calculating the first retention rate for the experimental group and the second retention rate for the control group, we can further calculate the incremental value of using the target feature to the target application based on the first and second retention rates. Specifically, as in the previous example, when the first retention rate is 87.5% and the second retention rate is 75%, the incremental value of using the target feature to the target application is 12.5%.

[0108] In some embodiments, obtaining operation data of the first user group and the third user group on the target application after a preset time period, and determining the incremental value of using the target function for the target application based on the operation data, includes:

[0109] A. Perform homogeneity detection on the first user group and the third user group;

[0110] B. When the result of the homogeneity detection meets the preset conditions, obtain the operation data of the first user group and the third user group on the target application after the preset time period, and determine the incremental value of using the target function for the target application based on the operation data.

[0111] Among them, in the embodiment of the present application, after the experimental group and the control group are determined, it is necessary to first perform a homogeneity test on the experimental group and the control group. Only when the homogeneity test results meet the preset conditions, that is, when it is determined that the experimental group and the control group are user groups with similar attributes, can the incremental value of using the target function to the target application be further calculated based on the operation data of the target application after the intervention period.

[0112] In some embodiments, performing homogeneity detection on the first user group and the third user group includes:

[0113] A1. Obtain third login data of the first user group and the third user group logging into the target application within a third target time period before the preset time period;

[0114] A2. Acquire fourth login data of the first user group and the third user group in a fourth target time period before the third target time period;

[0115] A3. Perform homogeneity detection on the first user group and the third user group based on the third login data and the fourth login data.

[0116] Among them, in an embodiment of the present application, whether the experimental group and the control group are homogeneous can be determined by calculating their retention rates before the intervention period. Specifically, the third login data of the experimental group and the control group logging into the target application within the third target time period before the intervention period can be obtained, and the fourth login data of the experimental group and the control group in the fourth target time period before the third target time period can be obtained. Among them, the time lengths of the third target time period and the fourth target time period are the same. Then, based on the third login data and the fourth login data, the retention rates of the users of the experimental group and the control group before the intervention period are calculated. If the difference in the retention rates of the two is less than the preset threshold, it means that the experimental group and the control group are homogeneous and the matching is successful; if the retention rates of the two are greater than the preset threshold, it means that the experimental group and the control group are not homogeneous, which means that there are large differences in the attribute data of the experimental group and the control group, and the matching fails.

[0117] Specifically, for example, there are 100 users in both the experimental group and the control group. In the fourth login data obtained in the fourth target time period, 80 people in the experimental group logged into the target application, and 100 people in the control group logged into the target application. In the third login data obtained in the third target time period, 60 of the 80 people in the experimental group who logged into the target application in the fourth target time period logged into the target application, and 60 of the 100 people in the control group who logged into the target application in the fourth target time period logged into the target application. If the retention rates of the experimental group and the control group are both 75%, it is determined that the experimental group and the control group are homogeneous. At this point, the incremental value of using the target function to the target application can be further determined based on the retention rates of the two after the intervention period. If the experimental group and the control group are not homogeneous, it is necessary to re-search for users matching the experimental group from the second user group to generate a new control group.

[0118] In some embodiments, performing homogeneity detection on the first user group and the third user group includes:

[0119] a. Performing a balance check on the attribute data of users in the first user group and the third user group;

[0120] b. When the balance check result is qualified, a homogeneity test is performed on the first user group and the third user group.

[0121] Among them, in the embodiment of the present application, before performing a homogeneity test on the experimental group and the control group, a balance check can also be performed on the attribute data of the users in the experimental group and the control group. Performing a balance check on the attribute data of the users in the experimental group and the control group is to detect the similarity of the experimental group and the control group in the attribute dimension of the users, which has nothing to do with the target application, while performing a homogeneity check on the experimental group and the control group is to check the similarity of the two groups from the perspective of the two groups' usage habits of the target application. The embodiment of the present application performs a similarity check on the experimental group and the control group from two different perspectives, thereby further ensuring the similarity of the experimental group and the control group before the intervention period, so that the only control variable is whether the target function is used, so that the result of calculating the incremental value of the target function to the target application is more accurate.

[0122] In some embodiments, performing a balance check on the attribute data of users in the first user group and the third user group includes:

[0123] a1. Calculate a first mean of the attribute data of each user in the first user group, a second mean of the attribute data of each user in the second user group, and a standard deviation of the attribute data of each user in the first user group;

[0124] a2. Calculate the difference between the first mean and the second mean;

[0125] a3. Calculate the ratio of the difference to the standard deviation;

[0126] a4. When the ratio is less than the preset value, the balance check result is determined to be qualified.

[0127] As previously mentioned, user attribute data can include data from multiple dimensions. Therefore, a balance check can be performed on the attribute data for each dimension individually. For the attribute data for any target dimension, the mean of the attribute data for that dimension can be calculated for each user in the experimental and control groups, obtaining a first mean for the experimental group and a second mean for the control group. The standard deviation of the attribute data for that dimension in the experimental group can also be calculated. The difference between the first and second means is then calculated, and the ratio of this difference to the standard deviation of the attribute data for the experimental group is calculated. If this ratio is less than a preset value, the attribute data for that dimension has passed the balance check. The attribute data for each dimension is then traversed. When the balance check for each dimension passes, the balance check for the attribute data for both the experimental and control groups is determined to have passed. If the attribute data for a particular dimension fails the balance check, the attribute data can be adjusted based on the specific business scenario to remove the attribute data for that dimension.

[0128] In this application, a causal inference model is employed to construct a user behavioral value assessment system based on the use of a logistic regression model to construct homogeneous experimental and control groups. Whether or not a user uses the target feature point is used as the sole intervention factor. The causal relationship between the target feature point and the target application is verified by determining the behavioral differences among homogeneous user groups after the target feature point is added to the target application. This effectively avoids the problem of being unable to determine a unique intervention factor due to differences in user attributes, ensures the accuracy of calculating the incremental value of using the target feature to the target application, and facilitates subsequent product strategy iteration.

[0129] According to the above description, the data processing method provided by the embodiment of the present application obtains the attribute data of a preset number of users in the target application; counts the usage of the target function in the target application by the preset number of users within a preset time period; divides the preset number of users into a first user group and a second user group according to the usage, and the first user group is a group composed of users who have used the target function; determines the preference score of each user for using the target function based on the attribute data and the usage; calculates the difference between the preference score of each user in the second user group and the preference score of each user in the first user group, and determines the users in the second group whose difference is less than a preset threshold to obtain a third user group; obtains the operation data of the first user group and the third user group on the target application after the preset time period, and determines the incremental value of using the target function to the target application based on the operation data. In this way, a third user group that is homogeneous with the first user group is determined from the second user group through the preference score of the user using the target function, and then determines the incremental value of the target function to the target application based on the operation data of the first user group and the second user group on the target application after the intervention period. This method can effectively determine homogeneous control and experimental groups, improve data processing efficiency, and thus improve the efficiency and accuracy of calculating the incremental value of the target function to the target application.

[0130] The present application also provides a data processing method, which can be used in a computer device, which can be a terminal or a server. Figure 3 FIG. 1 is another flow chart of the data processing method provided by the present application, which specifically includes:

[0131] Step 201: A computer device obtains attribute data of a preset number of users in a target application.

[0132] In this embodiment of the present application, the target application may be a mobile QQ (an Internet-based instant messaging software) application, and the computer device may obtain attribute data of a certain number of users in the mobile QQ, where the certain number is assumed to be 1000. The user attribute data may include data on dimensions such as user age, gender, number of days logged into mobile QQ within a period of time, average number of logins per day, number of friends to whom messages have been sent, and number of groups to which messages have been sent.

[0133] In step 202 , the computer device collects statistics on usage of a target function in a target application by a preset number of users within a preset time period.

[0134] Among them, in an embodiment of the present application, the target function in the target application can be the like function in the mobile QQ small world service. The preset time period can be a period of time after the like function is added to the mobile QQ small world service, specifically one week. During this week, statistics are taken on whether the above 1,000 users have used the like function. It is understandable that the statistical result of whether or not the like function has been used is a binary classification result, that is, a user either used this like function or did not use this like function. In an embodiment of the present application, the number of users who used the like function is less than the number of users who did not use the like function. For example, among these 1,000 users, 100 users used the like function and 900 users did not use the like function.

[0135] Since some users used the "Like" function and some did not during this week, users who used the "Like" function have been affected by the variable intervention compared to those who did not. Therefore, this week can also be called the intervention period.

[0136] In step 203 , the computer device trains a logistic regression model based on the attribute data and the usage to obtain a trained logistic regression model and obtains a propensity score for each user to use the target function based on the trained logistic regression model.

[0137] After obtaining the attribute data of each user and the usage of the like function by each user during the intervention period, the propensity score of each user to use the like function can be determined accordingly.

[0138] Specifically, a preset logistic regression model can be trained using the user's attribute data as the model input and the user's label data as the output. The user's label data can be determined based on whether the user used the "Like" function during the intervention period. If the user used the "Like" function during the intervention period, the user's label data is determined to be 1; otherwise, the user's label data is determined to be 0. After training the preset logistic regression model to obtain a trained logistic regression model, the attribute data of each user is input into the trained logistic regression model one by one to obtain a propensity score for each user to use the "Like" function. The higher the propensity score, the greater the probability that the user will use the "Like" function.

[0139] In step 204 , the computer device divides a preset number of users into a first user group and a second user group according to usage.

[0140] Based on how users used the "Like" function during the intervention period, users who used the "Like" function can be identified as members of the first user group, while users who did not use the "Like" function can be identified as members of the second user group. Since the users in the first user group have already been subjected to the variable intervention, the first user group can be identified as the experimental group. Furthermore, we need to identify a control group within the second user group that has not been subjected to the variable intervention, similar to the experimental group, so that the only control variable is whether or not the "Like" function was used. This allows us to determine whether or not the "Like" function generated incremental value for the application based on the differences in performance between the two groups after the intervention.

[0141] In step 205 , the computer device determines users in the second user group that match the users in the first user group based on the propensity scores to obtain a third user group.

[0142] Among them, for any user in the experimental group, first determine the target propensity score for using the like function, and then find users from the second user group whose propensity score for using the like function is close to the target propensity score. Specifically, the difference between each propensity score and the target propensity score in the second user group can be calculated. If the absolute value of the difference is less than a preset threshold, it is determined that the user is similar to the user in the experimental group. Then, traverse each user in the experimental group, match one by one to determine the users similar to them in the second user group, and then form the users determined from the second user group into a third user group, i.e., the control group. Among them, when the number of users in the experimental group is 100, the number of users in the control group is also 100.

[0143] Step 206: The computer device performs a balance check on the attribute data of users in the first user group and the third user group.

[0144] After determining the experimental group and the control group, in order to measure the matching effect of the above-mentioned matching and ensure that the attributes of the matched users in the experimental group and the control group are similar, it is necessary to perform a balance check on the attribute data of the users in the experimental group and the control group. Specifically, for the data of any dimension in the attribute data, the difference between the mean of the experimental group and the mean of the control group can be calculated, as well as the standard deviation of the experimental group. Finally, the ratio of the difference and the standard deviation is used to determine the balance check result. When the ratio is less than a preset value, it is determined that the balance check is qualified. If the ratio is greater than the preset value, it is determined that the balance check is unqualified.

[0145] Step 207: When the balance check result is qualified, the computer device performs a homogeneity check on the first user group and the third user group.

[0146] After checking the attribute data of users in the experimental group and the control group and determining that the inspection results are unqualified, further homogeneity testing can be performed on the users in the experimental group and the control group to determine whether the indicators used to evaluate the differences between the two after the intervention are consistent before the intervention.

[0147] Specifically, in the embodiment of the present application, the difference between the experimental group and the control group after the intervention period is evaluated by the sub-week retention rate. Therefore, before the intervention occurs, it is necessary to verify that the sub-week retention rates of the experimental group and the control group are the same or the difference is small.

[0148] In this way, we can obtain the login data of the experimental group and the control group on mobile QQ within two weeks before the intervention period, and determine the next-week retention rate of the experimental group and the control group based on the proportion of users who logged in in the previous week and logged in in the next week. If the difference in the next-week retention rates of the two is less than a preset threshold, then it can be determined that the experimental group and the control group are homogeneous.

[0149] like Figure 4 As shown, it is a time diagram, and the scheme of this application can be further described in detail with reference to this time diagram. As shown in the figure, the time span between time T1 and time T2 is one week, and the time span between time T2 and time T3 is also one week. Similarly, the time span between any two adjacent times is one week. Among them, the intervention period is the period from time T3 to time T4. After determining the experimental group and the control group, a homogeneity test is performed on the experimental group and the control group. The number of users who logged into mobile QQ in each group from time T1 to time T2 and logged in from time T2 to time T3 is detected, and the sub-week retention rate before the intervention is determined based on the ratio of the number of users who logged into mobile QQ from time T2 to time T3 to the number of logins in the previous week. If the sub-week retention rates of the experimental group and the control group are equal or the difference is less than the preset difference, then it is determined that the homogeneity test of the experimental group and the control group has passed.

[0150] Step 208 : When the homogeneity test is qualified, the computer device obtains the next-week retention rate of the first user group and the third user group after a preset time period.

[0151] Among them, after confirming that the experimental group and the control group are homogeneous, the login data of the experimental group and the control group within two weeks after the intervention period can be obtained, and then the retention rate of the experimental group and the control group in the second week after the intervention period can be determined based on the proportion of users who logged in in the previous week and logged in in the next week.

[0152] Please continue reading Figure 4 , obtain the sub-week retention rates of the experimental and control groups, that is, obtain the number of users who logged into QQ between T4 and T5 and logged in between T5 and T6, and calculate the ratio of retained users to obtain the sub-week retention rates of the experimental and control groups after the intervention.

[0153] In step 209 , the computer device calculates the incremental value of using the target function for the target application based on the sub-week retention rates of the first user group and the third user group after a preset time period.

[0154] Among them, after obtaining the next-week retention rates of the experimental group and the control group after the intervention period, the incremental value of using the like function on the next-week retention rate of mobile QQ users can be determined based on the difference in the next-week retention rates between the two.

[0155] Among them, since the experimental group and control group in this application are determined by observation based on the actual use of the target function by users, rather than by random grouping. Therefore, it is impossible to weaken the influence of confounding variables between the experimental group and the control group based on the effect of the law of large numbers, and it is easy to produce systematic deviations. The data processing method provided in this application can use the user's tendency to use the target function to match similar users, thereby eliminating interference factors between groups, and then accurately obtain the incremental value of the target function to the target application.

[0156] According to the above description, the data processing method provided by the embodiment of the present application obtains the attribute data of a preset number of users in the target application; counts the usage of the target function in the target application by the preset number of users within a preset time period; divides the preset number of users into a first user group and a second user group according to the usage, and the first user group is a group composed of users who have used the target function; determines the preference score of each user for using the target function based on the attribute data and the usage; calculates the difference between the preference score of each user in the second user group and the preference score of each user in the first user group, and determines the users in the second group whose difference is less than a preset threshold to obtain a third user group; obtains the operation data of the first user group and the third user group on the target application after the preset time period, and determines the incremental value of using the target function to the target application based on the operation data. In this way, a third user group that is homogeneous with the first user group is determined from the second user group through the preference score of the user using the target function, and then determines the incremental value of the target function to the target application based on the operation data of the first user group and the second user group on the target application after the intervention period. This method can effectively determine homogeneous control and experimental groups, improve data processing efficiency, and thus improve the efficiency and accuracy of calculating the incremental value of the target function to the target application.

[0157] In order to better implement the above method, an embodiment of the present invention further provides a data processing device, which can be integrated into a terminal or a server.

[0158] For example, Figure 5 , which is a schematic diagram of the structure of a data processing device provided in an embodiment of the present application, the data processing device may include an acquisition unit 301, a statistics unit 302, a division unit 303, a first determination unit 304, a calculation unit 305, and a second determination unit 306, as follows:

[0159] An acquisition unit, configured to acquire attribute data of a preset number of users in a target application;

[0160] A statistics unit, used to count usage of a target function in a target application by a preset number of users within a preset time period;

[0161] a division unit, configured to divide a preset number of users into a first user group and a second user group according to usage, wherein the first user group is a group consisting of users who have used the target function;

[0162] A first determining unit is configured to determine a preference score for each user to use the target function based on the attribute data and the usage situation;

[0163] a calculation unit, configured to calculate a difference between a preference score of each user in the second user group and a preference score of each user in the first user group, and determine users in the second group whose difference is less than a preset threshold, to obtain a third user group;

[0164] The second determining unit is configured to obtain operation data of the first user group and the third user group on the target application after a preset time period, and determine an incremental value of using the target function on the target application according to the operation data.

[0165] In some embodiments, the first determining unit includes:

[0166] A first determining subunit, configured to determine tag data corresponding to each user according to usage conditions;

[0167] A training subunit, configured to train a logistic regression model using attribute data of each user as input and label data of each user as output, thereby obtaining a trained logistic regression model;

[0168] The second determining subunit is configured to determine a preference score corresponding to each user based on the trained logistic regression model.

[0169] In some embodiments, the training subunit includes:

[0170] A training module, configured to train multiple logistic regression models using attribute data of each user as input and label data of each user as output, thereby obtaining multiple trained logistic regression models;

[0171] The second determining unit is further configured to:

[0172] Determining multiple transition preference scores corresponding to each user based on multiple trained logistic regression models;

[0173] The average of the multiple transition preference scores corresponding to each user is calculated to obtain the preference score corresponding to each user.

[0174] In some embodiments, the second determining unit includes:

[0175] A first acquisition subunit is configured to acquire first login data of the first user group and the third user group logging into the target application within a first target time period after a preset time period;

[0176] A second acquiring subunit is configured to acquire second login data of the first user group and the third user group logging into the target application within a second target time period after the first target time period;

[0177] a first calculation subunit, configured to calculate a first retention rate corresponding to the first group and a second retention rate corresponding to the third group based on the first login data and the second login data;

[0178] The second calculation subunit is configured to calculate the incremental value of using the target function for the target application based on the first retention rate and the second retention rate.

[0179] In some embodiments, the second determining unit includes:

[0180] a detection subunit, configured to perform homogeneity detection on the first user group and the third user group;

[0181] The third determination subunit is used to obtain the operation data of the first user group and the third user group on the target application after a preset time period when the result of the homogeneity detection meets the preset conditions, and determine the incremental value of using the target function for the target application based on the operation data.

[0182] In some embodiments, the detection subunit comprises:

[0183] A first acquisition module is configured to acquire third login data of the first user group and the third user group logging into the target application within a third target time period before the preset time period;

[0184] A second acquisition module is configured to acquire fourth login data of the first user group and the third user group in a fourth target time period before the third target time period;

[0185] The first detection module is configured to perform homogeneity detection on the first user group and the third user group according to the third login data and the fourth login data.

[0186] In some embodiments, the detection subunit comprises:

[0187] A detection module, configured to perform a balance check on the attribute data of users in the first user group and the third user group;

[0188] The second detection module is configured to perform a homogeneity detection on the first user group and the third user group when the balance check result is qualified.

[0189] In some embodiments, the inspection module includes:

[0190] A first calculation submodule, configured to calculate a first mean of the attribute data of each user in the first user group, a second mean of the attribute data of each user in the second user group, and a standard deviation of the attribute data of each user in the first user group;

[0191] A second calculation submodule, configured to calculate a difference between the first mean and the second mean;

[0192] A third calculation submodule is used to calculate the ratio of the difference to the standard deviation;

[0193] The determination submodule is used to determine that the result of the balance check is qualified when the ratio is less than a preset value.

[0194] In specific implementation, the above units can be implemented as independent entities, or can be arbitrarily combined to be implemented as the same or several entities. The specific implementation of the above units can be found in the previous method embodiments and will not be repeated here.

[0195] According to the above description, the data processing device provided by the embodiment of the present application obtains attribute data of a preset number of users in the target application through the acquisition unit 301; the statistics unit 302 counts the usage of the target function of the target application by the preset number of users within a preset time period; the division unit 303 divides the preset number of users into a first user group and a second user group based on the usage, where the first user group is a group composed of users who have used the target function; the first determination unit 304 determines the preference score of each user for using the target function based on the attribute data and the usage; the calculation unit 305 calculates the difference between the preference score of each user in the second user group and the preference score of each user in the first user group, and determines the users in the second group whose difference is less than a preset threshold to obtain a third user group; the second determination unit 306 obtains the operation data of the first user group and the third user group on the target application after the preset time period, and determines the incremental value of using the target function to the target application based on the operation data. In this way, a third user group with the same characteristics as the first user group is determined from the second user group based on the preference scores of the users using the target function, and then the incremental value of the target function to the target application is determined based on the operation data of the first user group and the second user group on the target application after the intervention period. This method can effectively determine homogeneous control and experimental groups, improve data processing efficiency, and thus improve the efficiency and accuracy of calculating the incremental value of the target function to the target application.

[0196] The embodiment of the present application also provides a computer device, which can be a terminal or a server. Figure 6 The figure shows the structure of the computer device provided by this application. Specifically:

[0197] The computer device may include components such as a processing unit 401 with one or more processing cores, a storage unit 402 with one or more storage media, a power module 403, and an input module 404. Those skilled in the art will appreciate that Figure 6The computer device structure shown in the figure does not constitute a limitation on the computer device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange components differently.

[0198] Processing unit 401 is the control center of the computer device. It utilizes various interfaces and lines to connect the various components of the entire computer device. By running or executing software programs and / or modules stored in storage unit 402 and accessing data stored in storage unit 402, it performs various functions of the computer device and processes data, thereby performing overall testing of the computer device. Optionally, processing unit 401 may include one or more processing cores; preferably, processing unit 401 may integrate an application processor and a modem processor, wherein the application processor primarily processes the operating system, user interface, and application programs, while the modem processor primarily handles wireless communications. It is understood that the modem processor may not be integrated into processing unit 401.

[0199] The storage unit 402 can be used to store software programs and modules. The processing unit 401 executes various functional applications and data processing by running the software programs and modules stored in the storage unit 402. The storage unit 402 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, at least one application required for a function (such as a sound playback function, an image playback function, and web page access, etc.); the data storage area may store data created based on the use of the computer device, etc. In addition, the storage unit 402 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage device. Accordingly, the storage unit 402 may also include a memory controller to provide the processing unit 401 with access to the storage unit 402.

[0200] The computer device also includes a power supply module 403 for supplying power to various components. Preferably, the power supply module 403 can be logically connected to the processing unit 401 via a power management system, thereby enabling the power management system to manage charging, discharging, and power consumption. The power supply module 403 can also include one or more DC or AC power supplies, a recharging system, a power failure detection circuit, a power converter or inverter, a power status indicator, and other arbitrary components.

[0201] The computer device may further include an input module 404, which may be configured to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0202] Although not shown, the computer device may further include a display unit, etc., which will not be described in detail here. Specifically, in this embodiment, the processing unit 401 in the computer device will load the executable files corresponding to one or more application processes into the storage unit 402 according to the following instructions, and the processing unit 401 will run the application stored in the storage unit 402 to implement various functions as follows:

[0203] Obtain attribute data of a preset number of users in a target application; count usage of a target function in the target application by a preset number of users within a preset time period; divide the preset number of users into a first user group and a second user group based on usage, the first user group being a group composed of users who have used the target function; determine a preference score for each user to use the target function based on the attribute data and usage; calculate the difference between the preference score of each user in the second user group and the preference score of each user in the first user group, and determine the users in the second group whose difference is less than a preset threshold to obtain a third user group; obtain operation data of the first user group and the third user group on the target application after the preset time period, and determine the incremental value of using the target function to the target application based on the operation data.

[0204] It should be noted that the computer device provided in the embodiment of the present application and the method in the above embodiment belong to the same concept. The specific implementation of the above operations can be referred to the previous embodiment and will not be repeated here.

[0205] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be accomplished by instructions, or by controlling related hardware through instructions. The instructions may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0206] To this end, an embodiment of the present invention provides a computer-readable storage medium storing a plurality of instructions that can be loaded by a processor to execute the steps of any method provided in an embodiment of the present invention. For example, the instructions can execute the following steps:

[0207] Obtain attribute data of a preset number of users in a target application; count usage of a target function in the target application by a preset number of users within a preset time period; divide the preset number of users into a first user group and a second user group based on usage, the first user group being a group composed of users who have used the target function; determine a preference score for each user to use the target function based on the attribute data and usage; calculate the difference between the preference score of each user in the second user group and the preference score of each user in the first user group, and determine the users in the second group whose difference is less than a preset threshold to obtain a third user group; obtain operation data of the first user group and the third user group on the target application after the preset time period, and determine the incremental value of using the target function to the target application based on the operation data.

[0208] The specific implementation of the above operations can be found in the previous embodiments and will not be repeated here.

[0209] The computer-readable storage medium may include a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0210] Since the instructions stored in the computer-readable storage medium can execute the steps in any method provided in the embodiments of the present invention, the beneficial effects that can be achieved by any method provided in the embodiments of the present invention can be achieved. Please refer to the previous embodiments for details and will not be repeated here.

[0211] According to one aspect of the present application, a computer program product or computer program is provided, wherein the computer program product or computer program includes computer instructions, and the computer instructions are stored in a storage medium. A processor of a computer device reads the computer instructions from the storage medium, and the processor executes the computer instructions, so that the computer device performs the above-mentioned Figure 2 or Figure 3 Methods provided in various optional implementations.

[0212] The above is a detailed introduction to a data processing method, device, computer-readable storage medium and computer equipment provided by an embodiment of the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

Claims

1. A data processing method, characterized in that: The method comprises: Get attribute data of a preset number of users in the target application; Counting usage of the target function in the target application by the preset number of users within a preset time period; dividing the preset number of users into a first user group and a second user group according to the usage situation, wherein the first user group is a group consisting of users who have used the target function; determining a preference score for each user to use the target function based on the attribute data and the usage situation; Calculating the difference between the preference score of each user in the second user group and the preference score of each user in the first user group, and determining the users in the second user group whose difference is less than a preset threshold, to obtain a third user group; Obtaining operation data of the first user group and the third user group on the target application after the preset time period, and determining the incremental value of using the target function for the target application based on the operation data, including: obtaining first login data of the first user group and the third user group logging into the target application within a first target time period after the preset time period; obtaining second login data of the first user group and the third user group logging into the target application within a second target time period after the first target time period; calculating a first retention rate corresponding to the first user group based on the first login data and the second login data of the first user group; calculating a second retention rate corresponding to the third user group based on the first login data and the second login data of the third user group; and calculating the incremental value of using the target function for the target application based on the first retention rate and the second retention rate.

2. The method according to claim 1, characterized in that The determining, based on the attribute data and the usage situation, a preference score for each user to use the target function includes: Determine tag data corresponding to each user according to the usage; Taking the attribute data of each user as input and the label data of each user as output, a logistic regression model is trained to obtain a trained logistic regression model; Based on the trained logistic regression model, a preference score corresponding to each user is determined.

3. The method according to claim 2, characterized in that The method of training a logistic regression model using attribute data of each user as input and label data of each user as output to obtain a trained logistic regression model includes: Taking the attribute data of each user as input and the label data of each user as output, multiple logistic regression models are trained to obtain multiple trained logistic regression models; Determining the preference score corresponding to each user based on the trained logistic regression model includes: Determining a plurality of transition preference scores corresponding to each user based on the plurality of trained logistic regression models; The average of the multiple transition preference scores corresponding to each user is calculated to obtain the preference score corresponding to each user.

4. The method according to claim 1, wherein Before the step of obtaining the operation data of the first user group and the third user group on the target application after the preset time period, and determining the incremental value of using the target function for the target application based on the operation data, the method further includes: performing homogeneity detection on the first user group and the third user group; When the result of the homogeneity detection meets the preset conditions, the operation data of the first user group and the third user group on the target application after the preset time period is obtained, and the incremental value of using the target function for the target application is determined based on the operation data.

5. The method according to claim 4, characterized in that The performing homogeneity detection on the first user group and the third user group includes: Acquire third login data of the first user group and the third user group logging into the target application within a third target time period before the preset time period; Acquire fourth login data of the first user group and the third user group in a fourth target time period before the third target time period; A homogeneity detection is performed on the first user group and the third user group according to the third login data and the fourth login data.

6. The method according to claim 4, characterized in that The performing homogeneity detection on the first user group and the third user group includes: Performing a balance check on the attribute data of users in the first user group and the third user group; When the balance check result is qualified, a homogeneity test is performed on the first user group and the third user group.

7. The method according to claim 6, characterized in that The performing a balance check on the attribute data of users in the first user group and the third user group includes: Calculating a first mean of the attribute data of each user in the first user group, a second mean of the attribute data of each user in the second user group, and a standard deviation of the attribute data of each user in the first user group; calculating a difference between the first mean and the second mean; calculating a ratio of the difference to the standard deviation; When the ratio is less than a preset value, the result of the balance check is determined to be qualified.

8. A data processing device, characterized in that: The device comprises: An acquisition unit, configured to acquire attribute data of a preset number of users in a target application; A statistics unit, configured to count usage of the target function in the target application by the preset number of users within a preset time period; a dividing unit, configured to divide the preset number of users into a first user group and a second user group according to the usage situation, wherein the first user group is a group consisting of users who have used the target function; a first determining unit, configured to determine a preference score for each user to use the target function based on the attribute data and the usage situation; a calculation unit, configured to calculate a difference between a preference score of each user in the second user group and a preference score of each user in the first user group, and determine users in the second user group whose difference is less than a preset threshold, to obtain a third user group; a second determining unit, configured to obtain operation data of the first user group and the third user group on the target application after the preset time period, and determine an incremental value of using the target function for the target application based on the operation data; The second determining unit includes: A first acquiring subunit is configured to acquire first login data of the first user group and the third user group logging into the target application within a first target time period after the preset time period; a second acquiring subunit, configured to acquire second login data of the first user group and the third user group logging into the target application within a second target time period after the first target time period; a first calculation subunit, configured to calculate a first retention rate corresponding to the first user group based on the first login data and the second login data of the first user group; and calculate a second retention rate corresponding to the third user group based on the first login data and the second login data of the third user group; The second calculation subunit is configured to calculate an incremental value of using the target function for the target application based on the first retention rate and the second retention rate.

9. The device according to claim 8, characterized in that The first determining unit includes: A first determining subunit, configured to determine tag data corresponding to each user according to the usage situation; A training subunit, configured to train a logistic regression model using attribute data of each user as input and label data of each user as output, thereby obtaining a trained logistic regression model; The second determining subunit is configured to determine a preference score corresponding to each user based on the trained logistic regression model.

10. The device according to claim 9, characterized in that The training subunit comprises: A training module, configured to train multiple logistic regression models using attribute data of each user as input and label data of each user as output, thereby obtaining multiple trained logistic regression models; The second determining unit is further configured to: Determining a plurality of transition preference scores corresponding to each user based on the plurality of trained logistic regression models; The average of the multiple transition preference scores corresponding to each user is calculated to obtain the preference score corresponding to each user.

11. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor to execute the steps of the data processing method according to any one of claims 1 to 7.

12. A computer device, characterized in that: The method comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the data processing method according to any one of claims 1 to 7 when executing the computer program.

13. A computer program product, characterized in that The computer program product includes computer instructions, which are stored in a storage medium; a processor of a computer device reads the computer instructions from the storage medium, and the processor executes the computer instructions, so that the computer device performs the steps of the data processing method according to any one of claims 1 to 7.

Images