Information flow recommendation strategy processing method, device, equipment and computer storage medium

Abstract

The embodiments of the present application provide an information flow recommendation strategy processing method, apparatus, device, and computer-readable storage medium, wherein the method includes: obtaining a preset operational target attribute, and determining at least one influencing feature that affects the operational target attribute; determining a target influencing feature in response to a selection operation for at least one influencing feature; determining a target recommendation population and a target correspondence corresponding to the target influencing feature based on historical information flow and operational target attributes, wherein the target correspondence includes a correspondence between a characteristic value of the target influencing feature and a target value of the operational target attribute; determining and presenting a recommendation strategy based on the target recommendation population and the target correspondence. Through the present application, the target recommendation population and recommendation strategy are automatically determined based on historical data and operational target attributes, which can improve recommendation efficiency and accuracy.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of big data processing technology, and are related to, but not limited to, an information flow recommendation strategy processing method, device, equipment, and computer storage medium. Background Art

[0002] With the development of mobile communications and internet technologies, online information streaming services have gradually replaced traditional media as the primary means for users to obtain information. Every online information platform generates a vast amount of information daily. If all this information were pushed to users, they wouldn't have the time to read it all, or even to filter through the vast amount of information to find the information they're truly interested in. This requires online information platforms to identify target audiences within this vast user base, identify information of interest to them within this vast information flow, and then push it to them, in order to achieve their operational goals. Summary of the Invention

[0003] The embodiments of the present application provide an information flow recommendation strategy processing method, apparatus, device and computer storage medium, which automatically determine the target recommendation population and recommendation strategy through historical data and operation target attributes, thereby improving recommendation efficiency and accuracy.

[0004] The technical solution of the embodiment of the present application is implemented as follows:

[0005] This embodiment of the present application provides a method for processing an information flow recommendation strategy, including:

[0006] Obtaining a preset operating target attribute, and determining at least one influencing feature that affects the operating target attribute;

[0007] In response to a selection operation on the at least one influencing feature, determining a target influencing feature;

[0008] Determine, based on the historical information flow and the operation target attribute, a target recommended population and a target correspondence corresponding to the target influence feature, wherein the target correspondence includes a correspondence between a characteristic value of the target influence feature and a target value of the operation target attribute;

[0009] Based on the target recommendation population and the target correspondence, a recommendation strategy is determined and presented, wherein the recommendation strategy includes a recommended value of the target impact feature and a target value of the operation target attribute corresponding to the recommended value.

[0010] The present invention provides an information flow recommendation strategy processing device, including:

[0011] A first acquisition module is configured to acquire a preset operation target attribute and determine at least one influencing feature that affects the operation target attribute;

[0012] A first determining module is configured to determine a target influencing feature in response to a selection operation on the at least one influencing feature;

[0013] A second determination module is configured to determine, based on the historical information flow and the operation target attribute, a target recommended population corresponding to the target influence feature and a target correspondence relationship, wherein the target correspondence relationship includes a correspondence relationship between a characteristic value of the target influence feature and a target value of the operation target attribute;

[0014] The third determination module is used to determine and present a recommendation strategy based on the target recommendation population and the target correspondence relationship, wherein the recommendation strategy includes a recommended value of the target impact feature and a target value of the operation target attribute corresponding to the recommended value.

[0015] The present invention provides an information flow recommendation strategy processing device, including:

[0016] The memory is used to store executable instructions; the processor is used to implement the above method when executing the executable instructions stored in the memory.

[0017] An embodiment of the present application provides a computer-readable storage medium storing executable instructions for causing a processor to execute the above method.

[0018] The embodiments of the present application have the following beneficial effects:

[0019] In the information flow recommendation strategy processing method provided in the embodiment of the present application, after obtaining the preset operation target attribute, at least one influencing feature that affects the operation target attribute is determined and presented, and in response to the selection operation of the at least one influencing feature, the target influence feature is determined, and then based on the historical information flow and the operation target attribute, the target recommendation population and target correspondence corresponding to the target influence feature are determined, wherein the target correspondence includes the correspondence between the characteristic value of the target influence feature and the target value of the operation target attribute, and finally based on the target recommendation population and the target correspondence, the recommendation strategy is determined and presented, wherein the recommendation strategy includes the recommended value of the target influence feature and the target value of the operation target attribute corresponding to the recommended value. In this way, after determining the operation target attribute and the target influence feature, the target recommendation population and recommendation strategy can be automatically determined by analyzing the historical information flow, which not only has high recommendation efficiency but also can improve recommendation accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A schematic diagram of a network architecture of the information flow recommendation strategy processing system 10 provided in an embodiment of the present application;

[0021] Figure 2 A schematic diagram of the structure of the operation platform 300 provided in an embodiment of the present application;

[0022] Figure 3 A schematic diagram of an implementation flow of the information flow recommendation strategy processing method provided in an embodiment of the present application;

[0023] Figure 4 A schematic diagram of another implementation flow of the information flow recommendation strategy processing method provided in an embodiment of the present application;

[0024] Figure 5 A schematic diagram of another implementation flow of the information flow recommendation strategy processing method provided in an embodiment of the present application;

[0025] Figure 6 This is a schematic diagram of the interface of the intelligent operation console provided in the embodiment of the present application;

[0026] Figure 7 A schematic diagram of another implementation process of the information flow policy processing method provided in an embodiment of the present application;

[0027] Figure 8 A schematic diagram of another implementation flow of the information flow recommendation strategy processing method provided in an embodiment of the present application. DETAILED DESCRIPTION

[0028] In order to make the purpose, technical solutions and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limiting this application. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0029] In the following description, reference is made to "some embodiments," which describe a subset of all possible embodiments. However, it will be understood that "some embodiments" may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict. Unless otherwise defined, all technical and scientific terms used in the embodiments of this application have the same meaning as commonly understood by those skilled in the art to which the embodiments of this application pertain. The terms used in the embodiments of this application are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0030] The relevant data collection and processing in the embodiments of this application should be strictly in accordance with the requirements of relevant laws and regulations when applied in examples, and the informed consent or separate consent of the personal information subject should be obtained. Subsequent data use and processing should be carried out within the scope of authorization of laws and regulations and the personal information subject.

[0031] The following is an explanation of the scientific terms involved in the embodiments of this application.

[0032] 1) Unique Visitors (UV) refers to the number of unique users who access the same webpage or product through the internet. Unique UVs are based on browser cookies. As long as the cookie is not clear, two people logging in from the same browser with different accounts between midnight and midnight will only be counted as one UV. UVs are a key metric for measuring traffic for most products. By monitoring UV changes, you can infer issues encountered across multiple channels, such as promotion and conversion.

[0033] 2) Page Views (PV): The total number of users who access the same webpage or product through the internet. PV is directly proportional to UV, and can be used to infer whether a campaign is popular or engaging. PV is always greater than UV in the same time period.

[0034] 3) Repeat Visitors (RV) refers to the number of users who repeatedly access the same webpage or product through the internet. The more repeat visitors, the higher the user stickiness.

[0035] 4) Daily Active Users (DAU) refers to the number of active users of a product or webpage in a single day. DAU can be abbreviated as DAU, which reflects the user activity of the product in a short period of time.

[0036] 5) Monthly Active Users (MAU): The number of active users of a product or webpage in a single month. MAU can be abbreviated as DAU, reflecting the user activity of the product over a long period of time.

[0037] 6) User retention rate refers to the proportion of users who continue to launch the application after a period of time among the number of new users in a certain statistical period.

[0038] 7) Sales (GMV, Gross Merchandise Volume) refers to the transaction amount of users on a certain website or product.

[0039] 8) Average Revenue Per User (ARPU) refers to total revenue divided by the number of users. Traditional industries use average order value (total revenue divided by the number of orders). However, the internet industry not only has a larger user base but also places greater emphasis on user operations. Therefore, each user's value to the product is higher, so ARPU is used. ARPU only reflects revenue, not profit.

[0040] 9) Time on Page (TP) refers to the length of time users spend on each page. TP duration can reflect the user's appeal to a particular webpage or activity. TP analysis can also be used to study user behavior preferences and preferences.

[0041] 10) Significant grip features: features that have a significant impact on operational objectives and are modifiable and actionable.

[0042] In order to better understand the information flow recommendation strategy processing method provided in the embodiments of the present application, the information flow recommendation strategy processing method in the related art is first described:

[0043] In related technologies, when determining information flow recommendation strategies, the target audience is categorized through historical experience or long-term data mining. Statistical analysis of the historical performance data of this audience over a certain period of time is then used to determine the recommended strategy to achieve the current operational objectives. This approach not only wastes manpower but also produces poor operational results.

[0044] The above problems are mainly caused by the following two reasons:

[0045] 1. Unable to quickly locate the people who can be operated in the information flow;

[0046] Second, after the population is identified, it is impossible to quickly and accurately quantify it, nor is it possible to quickly determine what operational strategy to implement for that population, resulting in serious user churn or poor user experience.

[0047] Based on this, an embodiment of the present application provides an operation strategy determination method, which uses a recurring event analysis model to automatically select a target population feature set that is useful for the core goals of information flow operations (DAU, duration, etc.) from the user's historical behavior data, and quantify the feature operation values that are operational for specific target populations to guide operations to improve core goals, and provide a system to support operations in independently selecting goals, viewing operation suggestions, and assisting decision-making.

[0048] The following describes an exemplary application of the information flow recommendation strategy processing device provided in an embodiment of the present application. The information flow recommendation strategy processing device provided in an embodiment of the present application can be implemented as any terminal with a screen display function, such as a laptop computer, a tablet computer, a desktop computer, a mobile device (for example, a mobile phone, a portable music player, a personal digital assistant, a dedicated messaging device, a portable gaming device), an intelligent robot, etc., and can also be implemented as a server. The following describes an exemplary application of the information flow recommendation strategy processing device when it is implemented as a server.

[0049] See also Figure 1 , Figure 1 This is a schematic diagram of the network architecture of the information flow recommendation strategy processing system 10 provided in the embodiment of the present application. Figure 1 As shown, the information flow recommendation strategy processing system 10 includes a terminal 100 (for example, a smart phone, tablet computer, desktop computer, laptop computer, etc.), a network 200, an operation platform 300 (for example, a user terminal such as a computer, a server) and an application server 400. Among them, an application is running on the terminal 100, which may be an instant messaging application, a video viewing application, an e-book reading application, etc. When implementing the information flow recommendation strategy processing method of the embodiment of the present application, the operator can set the operation target attributes on the operation platform, such as the number of active users, the transaction amount, the next-day retention rate, etc. After setting the operation target attributes, the operation platform 300 can determine the influencing characteristics that will affect the operation target attributes, and obtain the historical information flow of users running and using the application from the application server 400, and then determine the target recommendation population and recommendation strategy based on the historical information flow and the operation target attributes, and present them to the operator through the front end of the operation platform 100. After obtaining the recommendation strategy, the operator conducts specific strategy experiments and online strategy adjustments, and puts the recommendation strategy online after experiments and adjustments, that is, sends it to the application server 400. The application server 400 sends recommendation information to the terminal 100 corresponding to the target recommendation population based on the recommendation strategy.

[0050] See also Figure 2 , Figure 2 This is a schematic diagram of the structure of the operation backend 300 provided in an embodiment of the present application. Figure 2 The server 300 shown includes: at least one processor 310, a memory 350, at least one network interface 320, and a user interface 330. The various components in the server 300 are coupled together via a bus system 340. It is understood that the bus system 340 is used to achieve connection and communication between these components. In addition to the data bus, the bus system 340 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, the bus system 340 is not described in detail. Figure 2 Various buses are labeled as bus system 340 .

[0051] The processor 310 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., where the general-purpose processor can be a microprocessor or any conventional processor, etc.

[0052] The user interface 330 includes one or more output devices 331 that enable presentation of media content, including one or more speakers and / or one or more visual display screens. The user interface 330 also includes one or more input devices 332, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

[0053] The memory 350 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state memory, hard disk drives, optical disk drives, and the like. The memory 350 may optionally include one or more storage devices physically located away from the processor 310. The memory 350 includes a volatile memory or a non-volatile memory, and may also include both volatile and non-volatile memories. The non-volatile memory may be a read-only memory (ROM), and the volatile memory may be a random access memory (RAM). The memory 350 described in the embodiments of the present application is intended to include any suitable type of memory. In some embodiments, the memory 350 is capable of storing data to support various operations, examples of which include programs, modules, and data structures, or subsets or supersets thereof, as exemplified below.

[0054] Operating system 351, including system programs for processing various basic system services and performing hardware-related tasks, such as the framework layer, core library layer, and driver layer, which are used to implement various basic services and process hardware-based tasks;

[0055] A network communication module 352 for reaching other computing devices via one or more (wired or wireless) network interfaces 320 , exemplary network interfaces 320 including Bluetooth, WiFi, and USB;

[0056] The input processing module 353 is configured to detect one or more user inputs or interactions from one of the one or more input devices 332 and to translate the detected inputs or interactions.

[0057] In some embodiments, the apparatus provided in the embodiments of the present application may be implemented in software. Figure 2An information flow recommendation strategy processing device 354 stored in the memory 350 is shown. The information flow recommendation strategy processing device 354 may be an information flow recommendation strategy processing device in the server 300. The information flow recommendation strategy processing device 354 may be software in the form of a program or plug-in, and includes the following software modules: a first acquisition module 3541, a first determination module 3542, a second determination module 3543, and a third determination module 3544. These modules are logical and can be arbitrarily combined or further separated according to the functions implemented. The functions of each module will be described below.

[0058] In other embodiments, the apparatus provided in the embodiments of the present application may be implemented in hardware. As an example, the apparatus provided in the embodiments of the present application may be a processor in the form of a hardware decoding processor, which is programmed to execute the information flow recommendation strategy processing method provided in the embodiments of the present application. For example, the processor in the form of a hardware decoding processor may be one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0059] The following will describe the information flow recommendation strategy processing method provided by the embodiment of the present application in conjunction with the exemplary application and implementation of the operation backend 300 provided by the embodiment of the present application. Figure 3 , Figure 3 A schematic diagram of an implementation flow of the information flow recommendation strategy processing method provided in the embodiment of the present application, which is applied to Figure 1 The operation platform 300 shown below will be combined with Figure 3 The steps shown are explained.

[0060] Step S101: Acquire preset operation target attributes, and determine at least one influencing feature that affects the operation target attributes.

[0061] Here, when step S101 is implemented, it can be based on the setting operation of the operation personnel for the operation target attributes to obtain the preset operation target attributes, wherein the operation target attributes can be the level of activity, DAU, UV, the number of new registered users, the number of consumption conversion users, ARPU, user retention rate, etc.

[0062] In the embodiments of the present application, influencing features that will affect various operational target attributes can be preset. For example, when the application is a video viewing app, for activity level and DAU, its influencing features can be preset as exposure to variety entertainment channels, exposure to sports channels, exposure to TV drama channels, etc. Influencing features are features that can be manipulated and modified by operators. In some embodiments, the influencing features can be called grabbing features.

[0063] The characteristic value of an influencing characteristic can affect the target value of an operational target attribute. Influencing characteristics can have either a positive or negative impact on the operational target attribute. However, for each influencing characteristic, there are generally one or more specific groups of people for whom this influencing characteristic has a positive impact on the operational target attribute. In other words, as the characteristic value of the influencing characteristic increases, the target value of the operational target attribute also increases.

[0064] After obtaining the preset operation target attributes, one or more influencing features that will affect the operation target attributes can be determined based on the operation target attributes, and the one or more influencing features can be presented in the form of a drop-down box on the display device of the operation platform.

[0065] Step S102 : determining a target influencing feature in response to a selection operation on the at least one influencing feature.

[0066] Here, when step S102 is implemented, after one or more impact features are presented on the display device of the operation platform, the operation personnel can select one as a target impact feature according to actual needs.

[0067] Step S103: determining the target recommendation population and target correspondence corresponding to the target impact feature based on the historical information flow and the operation target attribute.

[0068] Here, when step S103 is implemented, the historical information flow is first obtained, and then the characteristic data in the historical information flow is extracted, which may include population characteristic data, page characteristic data, task characteristic data, etc. Currently, when extracting the characteristic data in the historical information flow, it is also necessary to extract the target impact characteristic data and the characteristic data corresponding to the operation target attributes, and then based on the extracted characteristic data and the operation target attributes, determine the target recommended population and target correspondence corresponding to the target impact characteristic, wherein the target correspondence includes the correspondence between the characteristic value of the target influence characteristic and the target value of the operation target attribute. For the target recommended population, the target impact characteristic is a characteristic that has a positive impact on the operation target attribute.

[0069] In the embodiment of the present application, the target influence feature may correspond to one or more target recommendation groups. For example, the target recommendation group may be people aged 30 to 40, people of male gender, or people who clicked on articles containing the keyword "life." Each target recommendation group corresponds to a target correspondence relationship, and the target correspondence relationships corresponding to different target recommendation groups are generally different.

[0070] For example, let's say the operational target attribute is daily active users (DAU), the target impact characteristic is the exposure of a variety entertainment channel, and the target recommendation demographic is women aged 15 to 25. In other words, increasing the exposure of the variety entertainment channel for this target demographic can increase DAU. This target correspondence can be either linear or nonlinear. For example, if the target correspondence is linear, DAU = 120 * exposure of the variety entertainment channel + 3000.

[0071] Step S104: determining and presenting a recommendation strategy based on the target recommendation population and the target correspondence relationship.

[0072] Here, the recommendation strategy includes the recommended value of the target impact feature and the target value of the operational target attribute corresponding to the recommended value. For example, the recommendation strategy may be to increase the exposure of the variety entertainment channel for the target recommended group of women aged 15 to 25 to 20,000, and the corresponding DAU may be increased to 150,000. In some embodiments, the recommendation strategy may also include the adjustment range of the target impact feature and the change range of the corresponding operational target attribute. For example, the recommendation strategy may be to increase the feature value of the target impact feature by 15%, and the change range of the operational target attribute by 20%.

[0073] In an embodiment of the present application, the recommendation strategy may be determined by calculation according to a preset rule based on the target correspondence relationship. For example, the preset rule may be that the characteristic value of the target influencing characteristic does not exceed a preset threshold.

[0074] In the information flow recommendation strategy processing method provided in the embodiment of the present application, after obtaining the preset operation target attribute, at least one influencing feature that affects the operation target attribute is determined and presented, and in response to the selection operation of the at least one influencing feature, the target influence feature is determined, and then based on the historical information flow and the operation target attribute, the target recommendation population and target correspondence corresponding to the target influence feature are determined, wherein the target correspondence includes the correspondence between the characteristic value of the target influence feature and the target value of the operation target attribute, and finally based on the target recommendation population and the target correspondence, the recommendation strategy is determined and presented, wherein the recommendation strategy includes the recommended value of the target influence feature and the target value of the operation target attribute corresponding to the recommended value. In this way, after determining the operation target attribute and the target influence feature, the target recommendation population and recommendation strategy can be automatically determined by analyzing the historical information flow, which not only has high recommendation efficiency but also can improve recommendation accuracy.

[0075] In some embodiments, the above step S103 "determining the target recommendation population and target correspondence corresponding to the target influence feature based on the historical information flow and the operation target attribute" can be implemented by the following steps S1031 to S1033:

[0076] Step S1031: Acquire historical information flow and extract feature data from the historical information flow.

[0077] Here, during implementation, step S1031 can retrieve historical information streams from the application server's data warehouse and perform data cleansing on them. This involves reexamining and verifying the data in the historical information streams. This process aims to remove duplicate information, correct existing errors, and ensure data consistency. Data cleansing is an essential step in the entire data analysis process, and the quality of its results is directly related to the model's effectiveness and ultimate conclusions.

[0078] After data cleaning, step S1031 can be implemented by extracting population characteristic data corresponding to each user account in the historical information flow at the granularity of user account, wherein the population characteristic data includes multiple attribute data; and extracting influence characteristic data in the historical information flow.

[0079] The feature data includes multiple feature attributes and feature values of the multiple feature attributes. For example, demographic feature data includes user name, age, gender, location, and other feature data. It can also include user behavior feature data, such as which articles they browsed, which keywords they searched for, which music they collected, etc.

[0080] Step S1032: cross-process the feature data to obtain a plurality of feature combination data.

[0081] Here, the plurality of feature combination data includes at least the historical feature value of the target impact feature.

[0082] In the embodiment of the present application, step S1032 can be implemented by the following steps:

[0083] Step S321: Based on the feature data, obtain at least one historical feature value of the target impact feature.

[0084] Here, when implementing step S321, based on the feature data, historical feature values of the target impact feature for a certain historical time period or several historical time periods can be obtained. The historical feature value can be the average value of the target impact feature within the time period. For example, the historical feature value of the target impact feature from the current time to seven days ago, the historical feature value of the target impact feature from seven days to 14 days ago, the historical feature value of the target impact feature from 14 days to 21 days ago, or the historical feature value of the target impact feature from 21 to 28 days ago can be obtained.

[0085] Step S322: Obtain the population characteristic data corresponding to each historical characteristic value.

[0086] Here, when implementing step S332, based on the historical time period corresponding to each historical feature value, the demographic data corresponding to each historical feature value is obtained. In other words, the demographic data corresponding to a certain historical feature value is obtained from the feature data within that historical time period. The demographic data includes feature values of multiple feature attributes, including feature values of basic user feature attributes and feature values of behavioral feature attributes.

[0087] Step S323 : combining each historical feature value with the feature value of the feature attribute in the corresponding population feature data to obtain a plurality of feature combination data.

[0088] Here, the characteristic attributes in the crowd characteristic data may include age, gender, location, etc. When implementing step S323, the historical characteristic value may be combined with the characteristic value of one characteristic attribute in the corresponding crowd characteristic data, or the historical characteristic value may be combined with the characteristic values of multiple characteristic attributes. For example, the historical characteristic value may be combined with the characteristic value of age, or the historical characteristic value may be combined with the characteristic values of age and gender.

[0089] Step S1033: Determine the target recommendation population and target correspondence relationship corresponding to the target impact feature based on the plurality of feature combination data.

[0090] Here, step S1033 can be implemented by the following steps:

[0091] Step S331: determine each historical target value of the operation target attribute corresponding to each feature combination data.

[0092] Step S332: Input the various feature combination data and the various historical target values into a data analysis model to obtain the various influencing parameters of the various feature combination data on the operation target attributes and the target correspondence.

[0093] Here, the data analysis model can be a survival model, such as the AG-CP (Anderson Gill) model, or a combination of a fragility model. These two models can address the dependencies between events and the heterogeneity of the population. The data analysis model is used to calculate the impact parameters of each feature combination data on the operational target attributes. These impact parameters can characterize the degree and nature of the impact of the target impact feature on the operational target attributes for the population corresponding to the feature combination data. The impact nature can include positive and negative impacts.

[0094] Step S333: determining the target recommendation population corresponding to the target influence feature based on the various influence parameters.

[0095] Here, step S333 can be implemented by the following steps:

[0096] Step S3331: Determine target feature combination data based on the various influencing parameters and the preset influencing threshold.

[0097] Here, when implementing step S3331, each influencing parameter may be compared with an influencing threshold, and the feature combination data corresponding to the influencing parameter greater than the influencing threshold may be determined as the target feature combination data. In this step, one or more target feature combination data may be determined.

[0098] Step S3332: Determine the target recommended population corresponding to the target influence feature based on the population feature data included in the target feature data.

[0099] Here, the demographic characteristics included in the demographic characteristic data may include one or more, for example, age 20-25, female, females aged 20-25, having searched for the keyword "music", etc. Based on the demographic characteristic data, the target recommendation demographic corresponding to the target influence characteristic may be determined.

[0100] In some embodiments, after step S104, Figure 4 Steps S105 to S109 shown below are combined Figure 4 Each step is explained.

[0101] Step S105 , obtaining the current information flow, and obtaining the current feature value of the target influence feature and the current target value of the operation target attribute corresponding to the target recommendation population based on the current information flow.

[0102] Step S106 : determining a first position of the target recommended group in a preset first coordinate system based on the current feature value and the current target value, and presenting the target recommended group in a first preset manner at the first position.

[0103] Here, in the first coordinate system, the target impact feature is used as the horizontal coordinate, and the operation target attribute is used as the vertical coordinate. Therefore, when implementing step S106, the current feature value can be determined as the value of the horizontal coordinate, and the current target value can be used as the value of the vertical coordinate, so as to determine the first position of the target recommended population in the first coordinate system, and present the target recommended population at the first position according to the first preset method, for example, the target recommended population can be presented as a point in the shape of a circle.

[0104] Step S107 : In response to the adjustment operation on the target impact feature, an adjustment value of the target impact feature is obtained.

[0105] Here, the adjustment operation can be an operation in which the operator inputs the characteristic value of the target impact characteristic, in which case the characteristic value input by the operator can be directly used as the adjustment value; the adjustment operation can also be an operation in which the operator clicks on a certain point in the horizontal axis in the first coordinate system, in which case the adjustment value of the target impact characteristic can be obtained based on the position of the adjustment operation.

[0106] In some embodiments, it is also possible to Figure 6 As shown, below the horizontal coordinate of the first coordinate system, an operation entry for adjusting the target impact feature is provided, and the feature value of the target impact feature can be adjusted by dragging the endpoint provided by the operation entry.

[0107] Step S108: determining a predicted value of the operation target attribute based on the adjustment value and the target correspondence.

[0108] Here, since the target correspondence is the correspondence between the target impact feature and the operation target attribute, after obtaining the adjustment value of the target impact feature, the predicted value of the operation target attribute can be determined according to the target correspondence.

[0109] Step S109 : determining a second position of the target recommended group in the first coordinate system based on the adjustment value and the predicted value, and presenting the target recommended group in the second position according to a first preset manner.

[0110] Here, the adjustment value can be determined as the value of the horizontal coordinate, and the predicted value can be used as the value of the vertical coordinate, so as to determine the second position of the target recommended group in the first coordinate system, and present the target recommended group at the second position according to the first preset method. That is, in this embodiment of the present application, the target recommended group can be presented in the first coordinate system in the form of hash points.

[0111] In some embodiments, different target recommendation groups are presented in the first coordinate system in different display modes. For example, the target recommendation group aged 15 to 20 is presented with circular points, and the target recommendation group of female is presented with triangular points.

[0112] Through the above-mentioned steps S105 to S109, after determining the target correspondence, the operator can adjust the characteristic value of the target influence feature to determine the predicted value of the operation target attribute, and present it intuitively in the first coordinate system, so that the operator can directly understand the degree of influence of the target influence feature on the operation target attribute, which is convenient for determining the final recommendation strategy.

[0113] In some embodiments, after step S109, the following steps may be further performed:

[0114] Step S110 : In response to a selection operation on a crowd identifier in the first coordinate system, a selected target identifier is obtained.

[0115] Here, the selection operation may be that the operator clicks or touches a point of a target recommended group in the first coordinate system, thereby obtaining the selected target identifier based on the selection operation.

[0116] Step S111 , obtaining the historical values of the target influence characteristics and the historical values of the operation target attributes of each user in the target recommendation population corresponding to the target identifier.

[0117] Here, each point in the first coordinate system represents a selectable target recommendation population. Once the target identifier is obtained, the corresponding target recommendation population can be determined, and then based on the user identifier of each user in the target recommendation population, the historical value of the target influence feature and the historical value of the operation target attribute corresponding to each user can be obtained.

[0118] Step S112: determining each third position of each user in the preset second coordinate system based on the historical value of the target impact feature and the historical value of the operation target attribute of each user.

[0119] Here, the second coordinate system uses the target influence feature as the horizontal coordinate and the operation target attribute as the vertical coordinate. After obtaining the historical values of the target influence feature and the operation target attribute of each user, the historical values of the target influence feature are used as the values of the horizontal coordinate, and the historical values of the operation target attribute are used as the values of the vertical coordinate, so as to determine the third positions of each user in the second coordinate system.

[0120] Step S113: Present each user at each third position according to a second preset manner.

[0121] Here, when displaying each user, each user may be presented in a manner different from that of displaying the target recommendation group, that is, in a second preset manner.

[0122] Through the above steps S110 to S113, the operator can select a specific target recommendation group in the first coordinate system and display the specific performance of the historical data of the target recommendation group in the second coordinate system.

[0123] Based on the above embodiments, the present invention further provides an information flow recommendation strategy processing method, which is applied to Figure 1 The network architecture shown, Figure 5 A schematic diagram of another implementation flow of the information flow recommendation strategy processing method provided in the embodiment of the present application is shown as follows: Figure 5 As shown, the method includes:

[0124] In step S201, the operation platform receives a setting operation for an operation target attribute and obtains the set operation target attribute.

[0125] Here, the operation platform can provide an application for recommendation strategy processing. After starting the application, the recommendation strategy processing main interface can be displayed. The main interface can provide a setting interface for setting operation target attributes. The operation personnel can set the operation target attributes through the setting interface, for example, it can be set to DAU, UV, etc.

[0126] In step S202 , the operation platform determines at least one influencing feature that affects the operation target attribute, and presents the at least one influencing feature.

[0127] In an embodiment of the present application, influencing features that will affect various operating target attributes can be pre-set. After obtaining the preset operating target attributes, one or more influencing features that will affect the operating target attributes can be determined based on the operating target attributes, and the one or more influencing features can be presented in the form of a drop-down box on the display device of the operating platform.

[0128] Step S203: The operation platform determines a target impact feature in response to the selection operation on the at least one impact feature.

[0129] In step S204, the operation platform obtains historical information flow and extracts feature data from the historical information flow.

[0130] Here, the characteristic data includes multiple characteristic attributes and characteristic values of the multiple characteristic attributes. The characteristic data may include crowd characteristic data and influence characteristic data.

[0131] In step S205 , the operation platform performs cross processing on the feature data to obtain a plurality of feature combination data.

[0132] Here, when implementing step S205, the influence feature data and the population feature data may be cross-processed. Since the population feature data includes multiple feature attributes and feature values corresponding to the multiple feature attributes, cross-processing the influence feature data and the population feature data may be performed by combining the feature values corresponding to one or more feature attributes in the influence feature data and the population feature data to obtain feature combination data. In other words, each feature combination data includes at least the historical feature value of the target influence feature.

[0133] In step S206 , the operation platform determines the target recommendation population and target correspondence relationship corresponding to the target influence feature based on the plurality of feature combination data.

[0134] Here, the target correspondence relationship includes the correspondence relationship between the characteristic value of the target impact characteristic and the target value of the operation target attribute.

[0135] When implementing step S206, the plurality of feature combination data may be input into at least one data analysis model for data processing, thereby determining the target recommendation population and the target correspondence relationship corresponding to the target influence feature.

[0136] In step S207 , the operation platform determines and presents a recommendation strategy based on the target recommendation population and the target correspondence relationship.

[0137] Here, the recommendation strategy includes a recommended value of the target impact feature and a target value of the operation target attribute corresponding to the recommended value.

[0138] In step S208 , the operation platform conducts a strategy experiment on the recommendation strategy, and adjusts the strategy based on the experiment results to obtain a final recommendation strategy.

[0139] Here, when implementing step S208, the recommended strategy can be executed to obtain the actual value of the operation target attribute after executing the recommended strategy to determine whether the recommended strategy is accurate, obtain experimental results, and then adjust the strategy based on the experimental results to obtain the final recommended strategy.

[0140] In step S209, the operation platform sends the final recommendation strategy to the application server to put the recommendation strategy online.

[0141] In step S210 , the application server determines the target recommendation group and recommendation information based on the recommendation strategy.

[0142] Step S211: The application server sends recommendation information to the terminals corresponding to the target recommendation group.

[0143] In the information flow recommendation strategy processing method provided in the embodiment of the present application, after obtaining the preset operation target attribute, at least one influencing feature that affects the operation target attribute is determined and presented. In response to the selection operation of the at least one influencing feature, the target influence feature is determined, and then based on the historical information flow and the operation target attribute, the target recommendation population and the target correspondence corresponding to the target influence feature are determined, wherein the target correspondence includes the correspondence between the feature value of the target influence feature and the target value of the operation target attribute. Finally, based on the target recommendation population and the target correspondence, a recommendation strategy is determined and presented, wherein the recommendation strategy includes the recommended value of the target influence feature and the target value of the operation target attribute corresponding to the recommended value. In this way, after determining the operation target attribute and the target influence feature, the target recommendation population and recommendation strategy can be automatically determined by analyzing the historical information flow, which not only has high recommendation efficiency but also can improve recommendation accuracy. After the operation personnel experiment and adjust the recommendation strategy, they can put the final recommendation strategy online so that the application server executes the recommendation strategy and sends recommendation information to the terminal corresponding to the target recommendation population, thereby achieving the operation target.

[0144] Below, an exemplary application of the embodiment of the present application in a practical application scenario will be described.

[0145] The present application provides an information flow recommendation strategy processing method, which is applied to an operation platform, wherein the operation platform includes: a feature collection and extraction module, a feature cross-calculation module, a significant feature selection (recurrence event analysis model) module and an operation strategy generation module, wherein:

[0146] The feature collection and extraction module is used to extract all feature data from the data flow and store the feature values of these features in the database for easy use in subsequent processes.

[0147] The feature cross-extraction calculation module is used to classify features, for example, to classify features into: user features, content features, task features, etc., but is not limited to these feature types. This module is also used to cross-organize features, that is, to combine different features.

[0148] The significant feature selection (recurrence event analysis model) module contains multiple data science models. Through continuous calculations in the background, it obtains the impact of different feature combinations on operational objectives, thereby determining the feature combinations that have a positive impact on operational objectives and recording specific feature impact data.

[0149] This module mainly uses the Recurrent Event Model to regress different feature combinations and obtain feature combinations with significant effects.

[0150] This module also includes a significant feature judgment algorithm to determine the feature combination that has a significant effect on the operation target. When it is implemented, the significant feature judgment algorithm can be used to sort the feature impact data of each different feature combination, thereby automatically discovering significant features, reducing manual intervention, and lowering the cost of manual evaluation.

[0151] The operation strategy generation module is used to calculate the significant features and numerical results obtained by the significant feature selection module according to preset rules to obtain the operation guidance strategy, and interact with the front-end page through the system interface.

[0152] The implementation process of using the above-mentioned operation platform to process information flow recommendation strategies may include:

[0153] Step S501: The feature collection and extraction module refines and extracts features from the pipeline data to obtain a usable feature set, which is stored in a feature set database.

[0154] In step S502 , the feature cross extraction calculation module uses different models and feature sets to perform modeling to obtain a result of the feature set.

[0155] In step S503 , the significant feature selection module uses a feature set utility evaluation module to sort and compare the results of the feature sets obtained by all models, select the optimal feature group and feature value, and save them in the result database.

[0156] In step S504, the operation strategy generation module displays the data on the page through the background service, and completes the dynamic display of the data through the interactive page.

[0157] Figure 6 This is a schematic diagram of the interface of the intelligent operation console provided in the embodiment of the present application, such as Figure 6As shown, the interface displays the operational goals supported by the operation system and the corresponding operational significant grabbing features. When implementing, the user can select the core goal 601 (activity level / DAU, etc., not limited to the examples given) on the interface. After determining the core goal 601, the system will return the list of significant grabbing features of the selected core goal and display it in the variable drop-down box. The user can select a significant grabbing feature from the list of significant grabbing features and display it in area 602 of the interface.

[0158] In addition, through this interface, you can view the significant gripping features and numerical performance of the operable population. Furthermore, the user selects the operable variables (significant gripping features) recommended by the system, and the system will pull the operable characteristic population of the selected core target and the selected operable variable. Different populations are represented by different types of shapes and displayed in the characteristic performance chart (i.e. Figure 6 The scatter plot on the left will become more diverse as the system is optimized, not just limited to Figure 6 scatter plot form in ). Figure 6 As shown in the figure, when the user selects the operational goal of "activity (next day retention probability)" and the significant grabbing feature of "novel channel exposure", the system pulls the operational characteristic population of the operational goal and the significant grabbing feature, such as Figure 6 As shown in the scatter plot on the left, the operational feature population has two categories, ranging from 12 to 20 (in Figure 6 square dots in the middle) and females (in Figure 6 are indicated by circular dots).

[0159] And the user is Figure 6 Select a specific operation group in the scatter plot shown on the left ( Figure 6 Each point in the graph is a selectable target group for operation. At this time, the historical data of the operation group can be displayed, such as Figure 6 As shown in the scatter plot on the right, when the operating population aged 12-20 is selected, the specific performance of the historical data of this age group can be displayed (not limited to Figure 6 The scatter plot on the right will have a richer effect).

[0160] pass Figure 6 In the interface shown, users can also modify the value of the significant gripper feature for the target population, view the expected target changes after the value of the significant gripper feature is changed, and Figure 6 The scatter plot on the left shows the change results. In actual operation, Figure 6 The interface shown provides an operation interface 603 for future trend prediction. The user can change the value of the significant gripper feature through the operation interface 603, and the system will give a predicted value of the operation target corresponding to the changed value of the significant gripper feature.

[0161] In some embodiments, the system can also give specific and reasonable operational suggestions and display them, such as increasing exposure to a certain group of people by 20%, which can ultimately increase DAU by 15%.

[0162] Users can conduct specific operational strategy experiments and online strategy adjustments based on the recommendations given by the system. After confirming that the operational results of the operational strategy are correct, they can make corresponding recommendations to the target population based on the operational strategy.

[0163] Figure 7 A schematic diagram of another implementation process of the information flow policy processing method provided in the embodiment of the present application is shown as follows: Figure 7 As shown, the implementation process includes two stages, wherein the first stage includes the following two steps:

[0164] Step S701: problem modeling.

[0165] Here, when implementing step S701, you can first obtain the historical information stream from the data warehouse, then clean the historical data stream to obtain the cleaned data, and then perform feature extraction on the cleaned data. When implementing, you can extract user feature sets, content feature sets, and task feature sets, and store the extracted features in the feature database through the feature engineering pipeline, and cross-sort the extracted features to obtain different feature combinations.

[0166] Step S702: model training.

[0167] Here, when implementing step S702, the different feature combinations obtained in step S701 are input into different models to obtain the processing results of each model, and based on the processing results of each model, the significant grabber feature set that has a positive impact on the team operation target is determined, and the significant grabber feature set is stored in the summary information (summry) database corresponding to the model.

[0168] like Figure 7 As shown, the second phase includes:

[0169] Step S703: deploy the model and output the processing results obtained by the model to the operation console.

[0170] Here, when implementing step S703, after receiving the request from the intelligent operation console, the model service obtains the feature data from the summry database and feature database corresponding to the model, determines the impact data of different feature combinations on the operation objectives, and then selects the significant gripper feature set that has a significant impact on the operation objectives based on the impact data. Finally, based on the significant gripper feature set and the impact data, the rule calculation is performed to obtain the operation guidance strategy, and the operation guidance strategy is returned to the intelligent operation console for presentation and interaction to the user.

[0171] Figure 8 A schematic diagram of another implementation flow of the information flow recommendation strategy processing method provided in the embodiment of the present application is shown as follows: Figure 8 As shown, the process includes:

[0172] Step S801: Obtain original user interaction flow.

[0173] Here, step S801 is implemented to obtain the original user interaction information flow.

[0174] Step S802: performing data cleaning and feature extraction on the user interaction information flow.

[0175] Here, data cleaning and feature extraction can be performed through the feature collection and extraction module.

[0176] Step S803: The extracted features are deposited into a feature library for storage.

[0177] Step S804: perform problem modeling, model training, evaluation, and selection based on the features in the database.

[0178] Here, the features in the database can be cross-calculated through the feature cross-extraction calculation module to combine the features. Then, different algorithm models can be used to determine the feature impact data of different feature combinations on the operation objectives, and the feature impact data can be stored in the feature result library table.

[0179] Step S805: export the model to a model file.

[0180] Step S806: Load the model into the model deployment service.

[0181] Step S807: perform feature selection.

[0182] Here, when step S807 is implemented, the significant feature selection module can use the significant feature judgment algorithm to select features that have significant effects on the operation target from the feature result library table.

[0183] Step S808: put the selection results into the final result database table.

[0184] Step S809: The intelligent operation platform obtains a set of significant gripper features.

[0185] In step S810 , the intelligent operation platform selects at least one significant hand feature from the significant hand feature set, and obtains the operation strategy corresponding to the significant hand feature from the model deployment service.

[0186] Here, the operation strategy generation module will process the content of the final result library table based on the significant gripper features selected by the user, return it to the front end, and present it to the user for interaction. In some embodiments, the front end displays the following Figure 6 The interface shown.

[0187] The following describes the implementation process of selecting significant trigger features for the recurrence event analysis model.

[0188] The recurrence event analysis model inputs the user dimension, content dimension, and behavior dimension in the information flow as features into the survival model, and maps the survival probability of the survival model into the active probability, thereby obtaining the impact value of each combination of features on the operation target. Furthermore, the RFM model features can be extracted based on the flow of the browser dashboard and the statistical features of the information flow can be extracted. Both the RFM features and the statistical features can be user behavior features. In the embodiment of the present application, exposure can be used as an operable and modifiable treatment feature, and the others can be used as group features. The effective treatment+group feature combination is cross-linked, and finally the cross-obtained treatment+group feature combination is input into the preset survival model to obtain the impact value of each feature combination on the operation target. In the embodiment of the present application, the preset survival model can be a combination of the AG-CP model and the fragility model. The use of these two models can solve the dependency between events and the heterogeneity of the population.

[0189] In some embodiments, the survival curve of the data set can also be predicted based on the survival model, the treatment feature is set to the mean, the survival curve is recalculated, the difference between the two survival curves is calculated, and the daily difference value of the survival curve is mapped to the daily visit probability to finally obtain the impact of the treatment feature on DAU.

[0190] In an embodiment of the present application, by analyzing historical data, it is possible to automatically discover features that have a significant effect on operational objectives, and determine the target population and corresponding operational strategies, thereby improving core operational objectives, reducing labor costs and the time cost of discovering useful features; and it is possible to quantify the actual benefits brought about by the target population and changes, reduce possible user damage caused by blind experiments, and improve the accuracy of product operations.

[0191] The following continues to describe the exemplary structure of the information flow recommendation strategy processing device 354 provided in the embodiment of the present application implemented as a software module. In some embodiments, such as Figure 2 As shown, the software module stored in the information flow recommendation strategy processing device 354 of the memory 350 may be the information flow recommendation strategy processing device in the operation background 300, including:

[0192] A first acquisition module 3541 is configured to acquire a preset operation target attribute and determine at least one influencing feature that affects the operation target attribute;

[0193] A first determining module 3542 is configured to determine a target influencing feature in response to a selection operation on the at least one influencing feature;

[0194] A second determining module 3543 is configured to determine, based on the historical information flow and the operation target attribute, a target recommended population corresponding to the target influence feature and a target correspondence relationship, wherein the target correspondence relationship includes a correspondence relationship between a characteristic value of the target influence feature and a target value of the operation target attribute;

[0195] The third determination module 3544 is used to determine and present a recommendation strategy based on the target recommendation population and the target correspondence relationship, wherein the recommendation strategy includes a recommended value of the target impact feature and a target value of the operation target attribute corresponding to the recommended value.

[0196] In some embodiments, the second determining module 3543 is further configured to:

[0197] Acquire a historical information stream, and extract feature data from the historical information stream, wherein the feature data includes a plurality of feature attributes and feature values of the plurality of feature attributes;

[0198] Cross-processing the feature data to obtain a plurality of feature combination data, wherein the plurality of feature combination data at least includes a historical feature value of a target impact feature;

[0199] Based on the multiple feature combination data, the target recommendation population and the target correspondence relationship corresponding to the target impact feature are determined.

[0200] In some embodiments, the second determining module 3543 is further configured to:

[0201] Based on the feature data, obtaining at least one historical feature value of the target impact feature;

[0202] Obtaining population characteristic data corresponding to each historical characteristic value;

[0203] Each historical feature value is combined with the corresponding population feature data to obtain multiple feature combination data.

[0204] In some embodiments, the second determining module 3543 is further configured to:

[0205] Determine the historical target values of the corresponding operational target attributes;

[0206] Inputting the characteristic combination data and the historical target values into a data analysis model to obtain the corresponding relationship between the influencing parameters of the characteristic combination data on the operation target attributes and the target;

[0207] The target recommended population corresponding to the target influence feature is determined based on the various influence parameters.

[0208] In some embodiments, the second determining module 3543 is further configured to:

[0209] Determining target feature combination data based on the various influencing parameters and a preset influencing threshold;

[0210] Based on the population characteristic data included in the target characteristic data, a target recommended population corresponding to the target influence characteristic is determined.

[0211] In some embodiments, the apparatus further comprises:

[0212] A second acquisition module is configured to acquire a current information flow, and based on the current information flow, acquire a current feature value of the target influence feature and a current target value of the operation target attribute corresponding to the target recommendation population;

[0213] a fourth determining module, configured to determine a first position of the target recommended group in a preset first coordinate system based on the current feature value and the current target value, and present the target recommended group at the first position in a first preset manner;

[0214] a third acquisition module, configured to acquire an adjustment value of the target impact feature in response to an adjustment operation on the target impact feature;

[0215] a fifth determining module, configured to determine a predicted value of the operation target attribute based on the adjustment value and the target correspondence;

[0216] A sixth determining module is configured to determine a second position of the target recommended group in the first coordinate system based on the adjustment value and the predicted value, and present the target recommended group in the second position according to a first preset manner.

[0217] In some embodiments, the apparatus further comprises:

[0218] A fourth acquisition module, configured to acquire a selected target identifier in response to a selection operation on the crowd identifier in the first coordinate system;

[0219] A fifth acquisition module is used to acquire the historical value of the target influence feature and the historical value of the operation target attribute of each user in the target recommendation group corresponding to the target identifier;

[0220] A seventh determination module, configured to determine each third position of each user in a preset second coordinate system based on the historical value of the target impact feature and the historical value of the operation target attribute of each user;

[0221] The presentation module is configured to present each user at each third position according to a second preset manner.

[0222] It should be noted that the description of the device embodiment of the present application is similar to the description of the method embodiment described above, and has similar beneficial effects as the method embodiment, so it will not be repeated. For technical details not disclosed in the device embodiment, please refer to the description of the method embodiment of the present application for understanding.

[0223] The embodiment of the present application provides a storage medium storing executable instructions, wherein the executable instructions are stored. When the executable instructions are executed by a processor, the processor will execute the method provided by the embodiment of the present application, for example, Figure 4 The method shown.

[0224] In some embodiments, the storage medium can be a computer-readable storage medium, such as a ferroelectric random access memory (FRAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM); it can also be various devices including one or any combination of the above memories.

[0225] In some embodiments, executable instructions may be in the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0226] By way of example, executable instructions may, but need not necessarily, correspond to a file in a file system, may be stored as part of a file storing other programs or data, such as one or more scripts in a Hypertext Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinating files (e.g., files storing one or more modules, subroutines, or code portions). By way of example, executable instructions may be deployed for execution on one computing device, or on multiple computing devices located at one site, or on multiple computing devices distributed across multiple sites and interconnected by a communication network.

[0227] The above description is merely an embodiment of the present application and is not intended to limit the scope of protection of the present application. Any modifications, equivalent replacements, and improvements made within the spirit and scope of the present application are included in the scope of protection of the present application.

Claims

1. A method for processing information flow recommendation strategies, characterized in that: include: Obtaining a preset operating target attribute, and determining at least one influencing feature that affects the operating target attribute; In response to a selection operation on the at least one influencing feature, determining a target influencing feature; Cross-processing the feature data extracted from the historical information stream to obtain a plurality of feature combination data, wherein the feature data includes a plurality of feature attributes and feature values of the plurality of feature attributes, and the plurality of feature combination data includes at least a historical feature value of the target impact feature; Determining, based on the historical target values of the operating target attributes corresponding to the respective feature combination data and the respective feature combination data, the impact parameters of the respective feature combination data on the operating target attributes and the target correspondence relationship, wherein the target correspondence relationship includes the correspondence relationship between the feature values of the target impact features and the target values of the operating target attributes; Determine a target recommended population corresponding to the target influence feature based on the influence parameter; Based on the target recommendation population and the target correspondence, a recommendation strategy is determined and presented, wherein the recommendation strategy includes a recommended value of the target impact feature and a target value of the operation target attribute corresponding to the recommended value.

2. The method according to claim 1, wherein The feature data extracted from the historical information stream is cross-processed to obtain multiple feature combination data, including: Based on the feature data extracted from the historical information flow, obtaining at least one historical feature value of the target impact feature; Obtaining population characteristic data corresponding to each of the historical characteristic values; Each of the historical feature values and the corresponding population feature data are combined to obtain a plurality of feature combination data.

3. The method according to claim 1, wherein Determining the influence parameter and target correspondence of each feature combination data on the operation target attribute based on the historical target value of the operation target attribute corresponding to each feature combination data and each feature combination data includes: Determining each historical target value of the operation target attribute corresponding to each of the feature combination data; The various feature combination data and the various historical target values are input into a data analysis model to obtain the various influencing parameters and target correspondences of the various feature combination data on the operation target attributes.

4. The method according to claim 1, wherein The determining, based on the influence parameter, a target recommended population corresponding to the target influence feature includes: Determining target feature combination data based on the influencing parameters and a preset influencing threshold; Based on the population characteristic data included in the target characteristic data, a target recommended population corresponding to the target influence characteristic is determined.

5. The method according to any one of claims 1 to 4, characterized in that The method further comprises: Acquire a current information flow, and based on the current information flow, acquire a current feature value of the target influence feature and a current target value of the operation target attribute corresponding to the target recommendation population; Based on the current feature value and the current target value, determining a first position of the target recommended group in a preset first coordinate system, and presenting the target recommended group at the first position in a first preset manner; In response to an adjustment operation on the target impact feature, obtaining an adjustment value of the target impact feature; Determining a predicted value of the operational target attribute based on the adjustment value and the target correspondence; Based on the adjustment value and the predicted value, a second position of the target recommended group in the first coordinate system is determined, and the target recommended group is presented at the second position according to a first preset manner.

6. The method according to claim 5, characterized in that The method further comprises: In response to a selection operation on a crowd identifier in the first coordinate system, obtaining a selected target identifier; Obtaining historical values of target influence characteristics and operation target attributes of each user in the target recommendation population corresponding to the target identifier; Determining each third position of each user in a preset second coordinate system based on the historical value of the target impact feature and the historical value of the operation target attribute of each user; Each user is presented at each third position according to a second preset manner.

7. An information flow recommendation strategy processing device, characterized in that: include: A first acquisition module is configured to acquire a preset operation target attribute and determine at least one influencing feature that affects the operation target attribute; A first determining module is configured to determine a target influencing feature in response to a selection operation on the at least one influencing feature; A second determination module is configured to perform cross-processing on feature data extracted from the historical information stream to obtain a plurality of feature combination data, wherein the feature data includes a plurality of feature attributes and feature values of the plurality of feature attributes, and the plurality of feature combination data includes at least the historical feature value of the target impact feature; determine, based on the historical target value of the operation target attribute corresponding to each of the feature combination data and each of the feature combination data, an impact parameter and a target correspondence relationship of each of the feature combination data on the operation target attribute, wherein the target correspondence relationship includes a correspondence between the feature value of the target impact feature and the target value of the operation target attribute; and determine, based on the impact parameter, a target recommendation population corresponding to the target impact feature; The third determination module is used to determine and present a recommendation strategy based on the target recommendation population and the target correspondence relationship, wherein the recommendation strategy includes a recommended value of the target impact feature and a target value of the operation target attribute corresponding to the recommended value.

8. The device according to claim 7, characterized in that The second determining module is further configured to obtain at least one historical feature value of the target impact feature based on feature data extracted from the historical information flow; Obtaining population characteristic data corresponding to each of the historical characteristic values; Each of the historical feature values and the corresponding population feature data are combined to obtain a plurality of feature combination data.

9. The device according to claim 7, characterized in that The second determination module is further configured to determine each historical target value of the operation target attribute corresponding to each of the feature combination data; The various feature combination data and the various historical target values are input into a data analysis model to obtain the various influencing parameters and target correspondences of the various feature combination data on the operation target attributes.

10. The device according to claim 7, characterized in that The second determining module is further configured to determine target feature combination data based on the influencing parameter and a preset influencing threshold; Based on the population characteristic data included in the target characteristic data, a target recommended population corresponding to the target influence characteristic is determined.

11. An information flow recommendation strategy processing device, characterized in that: include: a memory for storing executable instructions; A processor, configured to implement the method according to any one of claims 1 to 6 when executing the executable instructions stored in the memory.

12. A computer-readable storage medium, characterized in that Executable instructions are stored, which are used to cause a processor to execute and implement the method according to any one of claims 1 to 6.

Images