CN112104505B

CN112104505B - Application recommendation method, device, server and computer readable storage medium

Info

Publication number: CN112104505B
Application number: CN202010996415.4A
Authority: CN
Inventors: 李婷婷
Original assignee: Guangdong Oppo Mobile Telecommunications Corp Ltd; Shenzhen Huantai Technology Co Ltd
Current assignee: Guangdong Oppo Mobile Telecommunications Corp Ltd; Shenzhen Huantai Technology Co Ltd
Priority date: 2020-09-21
Filing date: 2020-09-21
Publication date: 2023-06-23
Anticipated expiration: 2040-09-21
Also published as: CN112104505A

Abstract

The application recommendation method, device, server and computer readable storage medium are disclosed. The method comprises the following steps: receiving an acquisition request for an application recommendation page sent by a target terminal, wherein the acquisition request carries a terminal identifier of the target terminal; acquiring user characteristic data corresponding to a target terminal, terminal characteristic data of the target terminal and application characteristic data of each application program to be recommended according to the terminal identification, and acquiring a predicted download probability value corresponding to each application program according to the user characteristic data, the terminal characteristic data and each application characteristic data; and determining a target application program to be displayed in the application recommendation page and the position of the target application program in the application recommendation page based on the predicted download probability value corresponding to each application program, generating the application recommendation page based on the target application program and the position, and sending an application recommendation interface to the target terminal. By adopting the method, the flexibility of application program recommendation can be improved.

Description

Application recommendation method, device, server and computer readable storage medium

Technical Field

The present invention relates to the field of computer technologies, and in particular, to an application recommendation method, an application recommendation device, a server, and a computer readable storage medium.

Background

With the popularization of intelligent terminals, various application programs are developed, and users can select required application programs to download and install through the application market of the intelligent terminals.

Currently, a plurality of application programs to be recommended are generally sequenced by manpower according to a fixed sequence, then the application programs are configured at fixed positions of an application recommendation page according to the sequencing sequence by manpower, after a user opens an application market, the intelligent terminal displays the application recommendation page so as to recommend the application programs in the application recommendation page to the user, and the application programs can be game applications, video applications and the like. Thus, if the user is interested in a certain application recommended in the application recommendation page, the application is downloaded.

However, in the above application recommendation method, the downloading rate of the application program generated based on the application recommendation is low, and the recommendation flexibility of the application program needs to be improved.

Disclosure of Invention

The embodiment of the application recommendation method, device, server and computer readable storage medium can improve flexibility of application recommendation.

In a first aspect, an embodiment of the present application provides an application recommendation method, where the method includes:

Receiving an acquisition request for an application recommendation page sent by a target terminal, wherein the acquisition request carries a terminal identifier of the target terminal;

acquiring user characteristic data corresponding to the target terminal, terminal characteristic data of the target terminal and application characteristic data of each application program to be recommended according to the terminal identification, and acquiring a predicted download probability value corresponding to each application program according to the user characteristic data, the terminal characteristic data and each application characteristic data;

and determining a target application program to be displayed in an application recommendation page and the position of the target application program in the application recommendation page based on the predicted download probability value corresponding to each application program, generating the application recommendation page based on the target application program and the position, and sending the application recommendation interface to the target terminal.

In a second aspect, an embodiment of the present application provides an application recommendation apparatus, where the apparatus includes:

the receiving module is used for receiving an acquisition request for an application recommendation page sent by a target terminal, wherein the acquisition request carries a terminal identifier of the target terminal;

The acquisition module is used for acquiring user characteristic data corresponding to the target terminal, terminal characteristic data of the target terminal and application characteristic data of each application program to be recommended according to the terminal identification, and acquiring a predicted download probability value corresponding to each application program according to the user characteristic data, the terminal characteristic data and each application characteristic data;

and the recommendation module is used for determining a target application program to be displayed in an application recommendation page and the position of the target application program in the application recommendation page based on the predicted download probability value corresponding to each application program, generating the application recommendation page based on the target application program and the position, and sending the application recommendation interface to the target terminal.

In a third aspect, embodiments of the present application provide a server comprising a memory storing a computer program and a processor implementing the steps of the method of the first aspect as described above when the processor executes the computer program.

In a fourth aspect, embodiments of the present application provide a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method of the first aspect described above.

According to the application recommendation method, the device, the server and the computer readable storage medium, after receiving the acquisition request for the application recommendation page sent by the target terminal, acquiring the user characteristic data corresponding to the target terminal, the terminal characteristic data of the target terminal and the application characteristic data of each application program to be recommended according to the terminal identification of the target terminal carried by the acquisition request, and acquiring the predicted download probability value corresponding to each application program according to the user characteristic data, the terminal characteristic data and each application characteristic data. In addition, the application program in different positions in the application recommendation page also has influence on whether the user downloads the application program, so that the position of the target application program in the application recommendation page is determined based on the predicted download probability value of each application program, the application recommendation page is generated based on the target application program and the position, and the application recommendation interface is sent to the target terminal. The application program recommendation flexibility is improved.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is an application environment diagram of an application recommendation method in one embodiment;

FIG. 2 is a flow chart of an application recommendation method in one embodiment;

FIG. 3 is a schematic diagram of an exemplary application recommendation page provided in one embodiment;

FIG. 4 is a flow diagram of a server adjusting model parameters of a download probability prediction model in one embodiment;

FIG. 5 is a flow chart of step 203 in one embodiment;

FIG. 6 is a flow diagram of an application recommendation method in one embodiment;

FIG. 7 is a flow chart of an application recommendation method in one embodiment;

FIG. 8 is a block diagram of an application recommendation device in one embodiment;

FIG. 9 is a schematic diagram of an internal structure of a server in one embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application.

The application recommendation of the conventional technology generally comprises the steps of manually sequencing a plurality of application programs to be recommended according to a fixed sequence, wherein the sequence is possibly determined by different client grades or popularization fees of application manufacturers, then manually configuring the application programs at fixed positions of an application recommendation page according to the sequencing sequence, and displaying the application recommendation page by an intelligent terminal after a user opens an application market so as to recommend the application programs in the application recommendation page to the user.

Taking a game application as an example, in a set exposure range, for each user, the game application is manually configured at a fixed position of an application recommendation page, and other application programs, such as a video application, a social application, and the like, are recommended at other positions of the application recommendation page. Therefore, after each user opens the application recommendation page, the positions of the application programs and the application programs can be the same, the recommendation mode is not targeted, and the exposure of the game application is probably large, but users who are actually interested in the game application and download the game application are few, so that the downloading rate of the game application is low, and the recommendation flexibility of the application programs needs to be improved.

In view of this, an embodiment of the present application provides an application recommendation method, in the method, after receiving an acquisition request for an application recommendation page sent by a target terminal, according to a terminal identifier of the target terminal carried by the acquisition request, user feature data corresponding to the target terminal, terminal feature data of the target terminal, and application feature data of each application program to be recommended are acquired, and then, according to the user feature data, the terminal feature data, and each application feature data, a predicted download probability value corresponding to each application program is acquired, where the predicted download probability value of each application program may represent a possibility that each application program is downloaded by a user in the target terminal, so, based on the predicted download probability value corresponding to each application program, a target application program to be displayed in the application recommendation page is determined, thereby being capable of recommending, in a targeted manner, the target application program to the user in combination with the possibility that each application program is downloaded by the user in the target terminal, and being favorable for improving the download rate of the application program. In addition, the application program in different positions in the application recommendation page also has influence on whether the user downloads the application program, so that the position of the target application program in the application recommendation page is determined based on the predicted download probability value of each application program, the application recommendation page is generated based on the target application program and the position, and the application recommendation interface is sent to the target terminal. The method improves the recommendation flexibility of the application program.

Next, an implementation environment related to the application recommendation method provided in the embodiment of the present application will be briefly described.

As shown in fig. 1, the implementation environment may include a server 101 and a plurality of terminals 102 (fig. 1 shows only one target terminal 102 by way of example, and the target terminal 102 may be any one of the plurality of terminals 102), and the target terminal 102 may communicate with the server 101 through a wired or wireless network.

The server 101 may be a server or a server cluster formed by a plurality of servers, and the server 101 may be a tower server, a rack server, a blade server, a high-density server, a single-path server, a dual-path server, or a multi-path server, which is not specifically limited in the type of the server 101 in the embodiment of the present application. The target terminal 102 may be a personal computer, a notebook computer, a media player, a smart television, a smart phone, a tablet computer, a portable wearable device, etc., and the type of the target terminal 102 is not limited in this embodiment.

FIG. 2 is a flow chart of an application recommendation method in one embodiment. The application recommendation method in this embodiment is described taking the server running in fig. 1 as an example. As shown in fig. 2, the application recommendation method includes

steps

201, 202, and 203:

Step 201, the server receives an acquisition request for an application recommendation page sent by a target terminal.

In this embodiment of the present application, the target terminal may be any user terminal, and the application recommendation page may be a page that the target terminal presents to the user through an application recommendation program, and the application recommendation program may be, for example, a software store, an application market, and the like installed in the target terminal.

Taking the application recommendation program as a software store as an example, as an implementation manner, if the target terminal detects an operation of opening the software store by a user or the target terminal detects an operation such as a page refreshing operation, a page turning operation and the like input by the user based on the software store, the target terminal sends an acquisition request for an application recommendation page, and the server receives the acquisition request for the application recommendation page sent by the target terminal. And the target terminal requests the application recommendation page to be displayed in the target terminal to the server through the acquisition request.

In this embodiment of the present application, the obtaining request carries a terminal identifier of the target terminal, where the terminal identifier of the target terminal may be an IMEI (International Mobile Equipment Identity, international mobile equipment identifier) of the target terminal or an OAID (Open Anonymous Identifier, anonymous equipment identifier) of the target terminal, and the like, and is not limited herein.

Step 202, the server obtains user characteristic data corresponding to the target terminal, terminal characteristic data of the target terminal and application characteristic data of each application program to be recommended according to the terminal identification, and obtains a predicted download probability value corresponding to each application program according to the user characteristic data, the terminal characteristic data and each application characteristic data.

In the embodiment of the application, the user characteristic data corresponding to the target terminal can represent information such as user attributes of the user, historical behavior characteristics of the user based on the target terminal and the like. For example, the user characteristic data may include information of a user's age, a user's sex, etc., and may further include at least one of application download data, application click data, application browse data, and application search data detected based on the target terminal. As an embodiment, the application download data may be a number of times the user downloads the same type of application as the application in a preset history period, the application click data may be a number of times the user clicks the same type of application as the application in the preset history period, the application browse data may be a duration of the user browsing the same type of application as the application in the preset history period, and the application search data may be a number of times the user searches the same type of application as the application in the preset history period.

The terminal characteristic data of the target terminal may characterize a terminal attribute of the target terminal, e.g., the terminal characteristic data may include at least one of model data and network type data of the target terminal.

The application feature data of each application program to be recommended may represent application attributes of the corresponding application program, for example, the application feature data of each application program at least includes application class data of the application program, and the application feature data of each application program may further include application classification data, online date and other data of the corresponding application program.

In the embodiment of the application, the server recalls the data resources from each data resource library, and after recall, the user characteristic data of each terminal, the terminal characteristic data and the application characteristic data of each application program in the terminal are stored in the server in a corresponding manner with the terminal identification of each terminal. The server can find out the user characteristic data corresponding to the target terminal, the terminal characteristic data of the target terminal and the application characteristic data of each application program to be recommended according to the terminal identification of the target terminal. And then, the server acquires a predicted download probability value corresponding to each application program according to the user characteristic data, the terminal characteristic data and the application characteristic data of each application program.

In one possible implementation manner, for each application program, the server may predict the user feature data, the terminal feature data, and the application feature data of the application program by using a trained neural network-based download probability prediction model, so as to obtain a predicted download probability value corresponding to the application program. In another possible implementation manner, for each application program, the server may further calculate a target hash value of the user feature data, the terminal feature data, and the application feature data of the application program, and then, the server searches, in a preset mapping table including mapping relationships between hash values and download probabilities, a target download probability corresponding to the target hash value, and obtains a predicted download probability value corresponding to the application program.

For an application, the predicted download probability value corresponding to the application may characterize the likelihood that the application is downloaded by the user in the target terminal. The user characteristic data corresponding to the target terminal, the terminal characteristic data of the target terminal and the application characteristic data of the application program are all in certain contact with the possibility that the application program is downloaded by a user in the target terminal.

For example, taking a game application in which an application program is first, if the user age in the user feature data characterizes the user as young, the probability that the game application is downloaded by the user in the target terminal is higher; if the application program downloading data, the application program clicking data, the application program browsing data and the application program searching data in the user characteristic data represent the application program of the user history preference game class, the possibility that the game application is downloaded by the user in the target terminal is higher; if the terminal characteristic data of the target terminal represent that the model of the target terminal is a high-end machine, the game application is downloaded by a user in the target terminal and paid for the user has higher possibility; if the terminal characteristic data of the target terminal characterizes the network of the target terminal to be better, the possibility that the game application is downloaded by a user in the target terminal is higher; if the application characteristic data of the game application characterizes that the application level of the game application is high, for example, S-level, the game application is highly likely to be downloaded by the user in the target terminal, and so on.

In view of the above-mentioned existing connection, the server trains a download probability prediction model or establishes a mapping relationship between each hash value and each download probability, so that, for each application program in the target terminal, the server can determine the possibility that the application program is downloaded by the user in the target terminal based on the user characteristic data, the terminal characteristic data and the application characteristic data of the application program.

And 203, the server determines a target application program to be displayed in the application recommendation page and the position of the target application program in the application recommendation page based on the predicted download probability value corresponding to each application program, generates the application recommendation page based on the target application program and the position, and sends an application recommendation interface to the target terminal.

After the server obtains the predicted download probability values corresponding to the application programs in the target terminal by adopting the embodiment, the server determines the target application program to be displayed in the application recommendation page from the application programs according to the predicted download probability values corresponding to the application programs.

In one possible implementation, the server may set a download probability threshold, compare the predicted download probability value for each application to the download probability threshold, and then determine an application in each application for which the predicted download probability value is greater than the download probability threshold as the target application. The download probability threshold may be set at the time of implementation. The target application program is an application program with a larger predicted download probability value in the application programs, so that the possibility that the target application program is downloaded by a user is larger, and the downloading amount of the target application program is improved.

In another possible implementation manner, the server may also sort the applications according to the order of the predicted download probability values from large to small to obtain a sorting result, and then select a preset number of applications with a front sorting as the target application from the sorting result, where the predicted download probability value of the application with the front sorting is larger, so that the possibility that the target application is downloaded by the user is larger, which is beneficial to improving the downloading amount of the target application. The preset number may be the maximum number of application programs that can be presented in the application recommendation page, or the preset number may be less than the maximum number, which is not particularly limited herein.

After determining the target application programs to be displayed in the application recommendation page, the server determines the positions of the target application programs in the application recommendation page according to the predicted download probability values of the target application programs. It will be appreciated that different locations of the application program in the application recommendation page may also have an impact on whether the user downloads the application program, for example, the application program may be more likely to be downloaded by the user when it is displayed in a location in the application recommendation page that is easily noticeable to the user. In this embodiment of the present application, as an implementation manner, the server may sort priorities of positions in the application recommendation page, where a position with a highest priority is most likely to draw attention of a user, determine a position with a highest priority as a position of a target application program with a maximum predicted download probability value, determine a position with a second priority as a position of a target application program with a second predicted download probability value, and so on, recommend the target application program to the user by combining with a position factor based on the predicted download probability value, which is favorable for further improving the download rate of the application program.

In this way, the server configures the target application program with the highest predicted download probability value at the position with the highest priority, the server configures the target application program with the second predicted download probability value at the position with the second priority, and the like, the server generates an application recommendation page based on the target application program and the position, and sends the application recommendation interface to the target terminal, and the target terminal displays the application recommendation interface, so that each target application program is displayed to the user according to the configuration of the server.

FIG. 3 is a schematic diagram of an exemplary application recommendation page. As shown in fig. 3, the priorities of the positions of the application programs in the application recommendation page decrease sequentially from top to bottom, and the predicted download probability value of the target application program "the cloudy recipe 2" in the target terminal is the largest, so that the server configures the "the cloudy recipe 2" at the first position of the application recommendation page. The predicted download probability values of the target application programs of 'jittering short video', 'jindong', 'jittering extremely fast edition' and 'homeland' are sequentially reduced, so that the 'jittering short video', 'jindong', 'jittering extremely fast edition' and 'homeland' are sequentially configured by the server at the position of the application recommendation page with the second priority, the position of the third priority, the position of the fourth priority and the position of the fifth priority and then displayed to a user.

According to the application recommendation method in the embodiment, after receiving an acquisition request for an application recommendation page sent by a target terminal, acquiring user characteristic data corresponding to the target terminal, terminal characteristic data of the target terminal and application characteristic data of each application program to be recommended according to a terminal identifier of the target terminal carried by the acquisition request, and acquiring a predicted download probability value corresponding to each application program according to the user characteristic data, the terminal characteristic data and the application characteristic data, wherein the predicted download probability value of each application program can represent the possibility that each application program is downloaded by a user in the target terminal. In addition, the application program in different positions in the application recommendation page also has influence on whether the user downloads the application program, so that the position of the target application program in the application recommendation page is determined based on the predicted download probability value of each application program, the application recommendation page is generated based on the target application program and the position, and the application recommendation interface is sent to the target terminal. The application program recommendation flexibility is improved.

For a first game application, experimental data shows that, under the condition of fixed exposure, compared with the traditional technology, the application recommendation method of the embodiment of the application can improve the distribution efficiency by 76%, wherein the distribution efficiency refers to the ratio of the download amount of the game application in the exposure of the game application, and therefore, the application recommendation method of the embodiment of the application greatly improves the download rate of the application program.

In one embodiment, based on the embodiment shown in fig. 2, the present embodiment relates to a process of how the server obtains the predicted download probability value corresponding to each application program according to the user feature data, the terminal feature data, and each application feature data. The process may include step A1:

and A1, for each application program, the server splices the user characteristic data, the terminal characteristic data and the application characteristic data of the application program to obtain spliced characteristic data, and inputs the spliced characteristic data into a download probability prediction model to obtain a predicted download probability value of the application program.

In the embodiment of the application, after obtaining user feature data corresponding to a target terminal, terminal feature data of the target terminal, and application feature data of each application program to be recommended, the server splices the user feature data, the terminal feature data, and the application feature data of the application program for each application program to obtain spliced feature data of the application program.

As an embodiment, for each application program, the server may perform hash processing on the user feature data, the terminal feature data, and the features in the application feature data of the application program, and combine the hash value of each feature and the feature number corresponding to each feature to form the spliced feature data of the application program.

For one application, the user characteristic data includes user age, user gender, application download data detected based on the target terminal, application click data, application browsing data, and application search data, the terminal characteristic data includes model data of the target terminal and network type data, and the application characteristic data of the application includes application class data of the application; the data format of the spliced feature data of the application is sign1: slot1, sign2: slot2. After the server hashes each feature, according to the feature number of each feature, the hash value of the feature with the feature number of 1 is taken as sign1, the slot1 is taken as 1, the hash value of the feature with the feature number of 2 is taken as sign2, the slot2 is taken as 2, and the like, so as to obtain the spliced feature data of the application program.

And the server inputs the spliced characteristic data of the application program into a download probability prediction model, and then obtains a predicted download probability value of the application program.

In this embodiment of the present application, as an implementation manner, the server may train the neural network model by using a large number of sample stitching feature data in advance, and in the model training process, the server uses a download probability value corresponding to each sample stitching feature data as supervision, and repeatedly iterates the training, and after the model converges, the download probability prediction model is obtained. And the same as the splicing characteristic data, one sample splicing characteristic data is obtained by splicing the sample application characteristic data, the user characteristic data and the terminal characteristic data of one application program.

In this way, the server can obtain the predicted download probability value of each application program in the target terminal, and make application recommendation based on the predicted download probability value of each application program.

In one possible implementation, referring to fig. 4, fig. 4 relates to a process of how the server adjusts model parameters of the download probability prediction model. As shown in fig. 4, the process may include steps 401 and 402:

in step 401, the server obtains training adjustment samples according to the download data of different terminals to different application programs.

The training adjustment samples comprise sample user characteristic data, sample terminal characteristic data, sample application characteristic data of an application program and sample download probability values.

In the embodiment of the application, the server generates the application recommendation page based on the target application program and the position, and the target terminal displays the application recommendation interface after sending the application recommendation interface to the target terminal. In this way, each different terminal displays a corresponding application recommendation interface, and based on the displayed application recommendation interface, the terminal collects the download data of each application program and sends the download data to the server.

The server analyzes the downloaded data sent by each terminal to obtain sample user characteristic data, sample terminal characteristic data and sample application characteristic data of each application program corresponding to the terminal. The server determines whether terminals having the same sample user feature data, sample terminal feature data, and sample application feature data have downloaded the corresponding application program by analyzing the download data transmitted from each terminal, whereby the server can determine a sample download probability value for such sample user feature data, sample terminal feature data, and sample application feature data for the application program. Similarly, the server can obtain the sample user characteristic data, the sample terminal characteristic data, the sample application characteristic data and the sample download probability value of each terminal for each application program by analyzing the download data.

In step 402, the server adjusts model parameters of the download probability prediction model based on the training adjustment samples.

For each application program of each terminal, the server splices the sample user characteristic data of the terminal, the sample terminal characteristic data and the sample application characteristic data of the application program to obtain spliced characteristic data of the terminal and the application program; the server inputs the spliced characteristic data corresponding to each terminal and each application program into a download probability prediction model, adopts a sample download probability value corresponding to each application program in each terminal as supervision, iteratively trains the download probability prediction model, and adjusts model parameters of the download probability prediction model.

Therefore, the server is beneficial to improving the prediction accuracy of the download probability prediction model by adjusting the model parameters of the download probability prediction model, so that the accuracy of the predicted download probability value corresponding to each application program in the target terminal is improved, and the accuracy of application recommendation is improved.

In one embodiment, referring to fig. 5 on the basis of the embodiment shown in fig. 2, this embodiment relates to a process of how the server determines, based on the predicted download probability values corresponding to the respective application programs, the target application program to be displayed in the application recommendation page and the position of the target application program in the application recommendation page. As shown in fig. 5, step 203 may include step 2031, step 2032, and step 2033:

In step 2031, the server obtains the number of predicted resource value transitions corresponding to each application.

In this embodiment of the present application, for each application in the target terminal, the server obtains the number of predicted resource value transitions corresponding to the application. As an embodiment, the server may determine the predicted number of resource value transfers for each application over a future period of time, such as 30 days in the future, 90 days in the future, etc., based on the internal data of the application or the payment data of the application at the same application level as the application.

For example, taking the example that the application program is a game application, the server determines payment data of the game application in the internal measurement stage for 30 days as the predicted resource value transfer quantity corresponding to the game application.

In step 2032, for each application, the server corrects the predicted download probability value of the application according to the number of resource value transitions of the application, to obtain a first value of the application.

The server corrects the predicted download probability value of the application program according to the resource value transition number of the application program, for example, the resource value transition number of the application program may be multiplied by the predicted download probability value of the application program to obtain the first value of the application program.

In step 2033, the server determines, based on the first values of the applications, the target application to be presented in the application recommendation page and the location of the target application in the application recommendation page.

In an actual service scene, various factors are often considered when the application is recommended, and in the embodiment of the application, the recommendation of the target application is performed by combining the resource numerical value transfer quantity of the application on the basis of the predicted downloading probability of the application, so that the data base of the application recommendation is enriched, the data dimension is increased, and the flexibility of the application recommendation is improved.

In one possible implementation of step 2033, step 2033 may include the following steps a, b, c, and d:

and a step a, the server acquires the conversion rate corresponding to each application program.

In the embodiment of the application, for each application program, the conversion rate is used to characterize the ratio of the historical registration amount of the application program to the historical download amount of the application program. The server may count the historical download data and the historical registration data of each application program, and the ratio of the historical registration amount of the application program to the historical download amount of the application program is the conversion rate corresponding to the application program.

And b, for each application program, the server corrects the first value of the application program according to the conversion rate of the application program to obtain the second value of the application program.

The server may perform the correction processing on the first value of the application program according to the conversion rate of the application program, and multiply the conversion rate of the application program with the first value of the application program to obtain the second value of the application program. In this way, the server obtains the second value of each application in the target terminal.

And c, the server performs descending order arrangement on the second values of the application programs to obtain an ordering result.

And d, the server determines a target application program to be displayed in the application recommendation page from the application programs according to the sequencing result, and determines the position of the target application program in the application recommendation page.

The server performs descending order arrangement on the second values of the application programs to obtain an ordering result, and then the server selects a preset number of application programs with the top ordering from the ordering result as target application programs. The server can configure the target application program with the largest second value at the position with the highest priority of the application recommendation interface, the server configures the target application program with the second value at the position with the second priority, and the like, the server generates an application recommendation page and sends the application recommendation interface to the target terminal, and the target terminal displays the application recommendation interface.

In the embodiment of the application program recommendation method, the second value of the application program ranked at the front is larger, namely the comprehensive recommendation score of the application program is larger, so that the target application program can be recommended to the user by combining factors (the predicted download probability value, the predicted resource numerical value transfer quantity and the conversion rate) of multiple dimensions, and the rationality and the flexibility of application recommendation are improved.

In one embodiment, based on the embodiment shown in fig. 2 and described above, referring to fig. 6, this embodiment relates to a process how the server determines an application to be recommended. As shown in fig. 6, the application recommendation method of the present embodiment further includes step 204 and step 205:

in step 204, the server obtains a plurality of candidate applications, and determines recommended conversion values corresponding to the candidate applications respectively.

In this embodiment of the present application, the server may obtain recommended conversion values corresponding to each candidate application program in the target terminal, where the recommended conversion values include at least one of average application person download data, application exposure, and number of application resource numerical transfers.

As an implementation mode, the server can determine application exposure, download data of each terminal, resource value transfer quantity and the like by inquiring the feedback log sent by each terminal, so that the application average download data can be obtained.

In step 205, the server screens out candidate application programs whose recommendation conversion values do not reach the preset recommendation index values from the plurality of candidate application programs, and takes the screened candidate application programs as application programs to be recommended.

In this embodiment of the present application, the preset recommendation index value may be a preset application person average download index, an application exposure index, and a resource numerical value transfer number index, and the preset recommendation index value may be set by itself when implemented. If the recommendation conversion value of a certain application program reaches the preset recommendation index value, the application program is characterized to be not required to be recommended, so that the server screens candidate application programs, the recommendation conversion value of which does not reach the preset recommendation index value, from a plurality of candidate application programs, and takes the screened candidate application programs as the application programs to be recommended, thereby avoiding the application programs which do not need to be recommended from being recommended to users by mistake and ensuring the service compliance of application recommendation.

Taking the recommendation conversion value including the application average download data as an example, the application program is a game application, if the recommendation conversion value of the game application does not reach the application average download index, the server determines the game application as the application program to be recommended, and continues to recommend the game application in a targeted manner by combining with the predicted download probability value of the game application so as to promote the download amount of the game application.

Taking the recommended conversion value including the application exposure as an example, the application program is a game application, if the recommended conversion value of the game application does not reach the application exposure index, the server determines the game application as the application program to be recommended, and continues to recommend the game application in a targeted manner by combining with the predicted download probability value of the game application so as to promote the exposure of the game application.

Taking the case that the recommended conversion value comprises the number of application resource value transitions, and the application program is a game application, if the recommended conversion value of the game application does not reach the index of the number of resource value transitions, the server determines the game application as the application program to be recommended, and continues to recommend the game application in a targeted manner by combining with the predicted download probability value of the game application so as to promote the number of resource value transitions of the game application.

Therefore, the embodiment combines different business requirements (the application personnel average download index, the application exposure index or the resource numerical value transfer quantity index), can realize flexible recommendation of the application program, and improves the recommendation flexibility of the application program.

Referring to fig. 7, a flowchart of another application recommendation method provided in an embodiment of the present application is shown, where the application recommendation method may be applied to a server in the implementation environment shown in fig. 1. As shown in fig. 7, the application recommendation method may include the steps of:

In step 7001, the server receives an acquisition request for an application recommendation page sent by the target terminal.

The acquisition request carries the terminal identification of the target terminal.

In step 7002, the server obtains a plurality of candidate applications, and determines recommended conversion values corresponding to the candidate applications respectively.

Wherein, the recommended conversion value comprises at least one of application person average download data, application exposure and application resource numerical transfer quantity.

In step 7003, the server screens out candidate application programs whose recommendation conversion values do not reach preset recommendation index values from the plurality of candidate application programs, and takes the screened candidate application programs as application programs to be recommended.

In step 7004, the server obtains user feature data corresponding to the target terminal, terminal feature data of the target terminal, and application feature data of each application program to be recommended according to the terminal identifier.

The user characteristic data comprises at least one of application program downloading data, application program clicking data, application program browsing data and application program searching data which are detected based on the target terminal; the terminal characteristic data comprises at least one of model data and network type data of the target terminal; the application characteristic data of each application program at least includes application level data of the application program.

Step 7005, for each application program, the server splices the user feature data, the terminal feature data and the application feature data of the application program to obtain spliced feature data, and inputs the spliced feature data into the download probability prediction model to obtain a predicted download probability value of the application program.

Step 7006, the server obtains the number of predicted resource value transitions corresponding to each application program; and for each application program, correcting the predicted download probability value of the application program according to the resource value transfer quantity of the application program to obtain a first value of the application program.

Step 7007, the server obtains the conversion rate corresponding to each application program; for each application program, carrying out correction processing on a first value of the application program according to the conversion rate of the application program to obtain a second value of the application program; the second values of the application programs are arranged in a descending order to obtain a sequencing result; and according to the sorting result, determining a target application program to be displayed in the application recommendation page from the application programs, and determining the position of the target application program in the application recommendation page.

Wherein the conversion rate is used to characterize the ratio of the historical registration amount of the application program to the historical download amount of the application program.

In step 7008, the server generates an application recommendation page based on the target application program and the location, and sends the application recommendation interface to the target terminal.

In step 7009, the server obtains training adjustment samples according to the downloaded data of different terminals to different application programs.

The training adjustment sample comprises sample user characteristic data, sample terminal characteristic data, sample application characteristic data of an application program and a sample downloading probability value;

step 7010, the server adjusts model parameters of the download probability prediction model based on the training adjustment sample.

It should be understood that, although the steps in the flowcharts of fig. 2-7 are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in fig. 2-7 may include multiple sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, nor does the order in which the sub-steps or stages are performed necessarily occur in sequence, but may be performed alternately or alternately with at least a portion of the sub-steps or stages of other steps or other steps.

Fig. 8 is a block diagram of an application recommendation apparatus according to an embodiment. As shown in fig. 8, the application recommendation apparatus includes:

the receiving module 10 is configured to receive an acquisition request for an application recommendation page sent by a target terminal, where the acquisition request carries a terminal identifier of the target terminal;

the obtaining module 20 is configured to obtain, according to the terminal identifier, user feature data corresponding to the target terminal, terminal feature data of the target terminal, and application feature data of each application program to be recommended, and obtain, according to the user feature data, the terminal feature data, and each application feature data, a predicted download probability value corresponding to each application program;

the recommendation module 30 is configured to determine a target application program to be displayed in an application recommendation page and a position of the target application program in the application recommendation page based on the predicted download probability value corresponding to each application program, generate the application recommendation page based on the target application program and the position, and send the application recommendation interface to the target terminal.

Optionally, the obtaining module 20 is specifically configured to splice, for each application program, the user feature data, the terminal feature data, and the application feature data of the application program to obtain spliced feature data, and input the spliced feature data into a download probability prediction model to obtain a predicted download probability value of the application program.

Optionally, the apparatus further comprises:

the sample acquisition module is used for acquiring training adjustment samples according to the download data of different terminals to different application programs, wherein the training adjustment samples comprise sample user characteristic data, sample terminal characteristic data, sample application characteristic data of the application programs and sample download probability values;

and the training module is used for adjusting the model parameters of the download probability prediction model based on the training adjustment sample.

Optionally, the user characteristic data includes at least one of application download data, application click data, application browsing data, and application search data detected based on the target terminal; the terminal characteristic data comprises at least one of model data and network type data of the target terminal; the application characteristic data of each application program at least comprises application grade data of the application program.

Optionally, the recommendation module 30 includes:

the obtaining unit is used for obtaining the number of the predicted resource value transitions corresponding to each application program;

the correction unit is used for correcting the predicted download probability value of each application program according to the resource value transfer quantity of the application program to obtain a first value of the application program;

And the recommending unit is used for determining a target application program to be displayed in an application recommending page and the position of the target application program in the application recommending page based on the first value of each application program.

Optionally, the recommending unit is specifically configured to obtain a conversion rate corresponding to each application program, where the conversion rate is used to characterize a ratio of a historical registration amount of the application program to a historical download amount of the application program; for each application program, correcting the first value of the application program according to the conversion rate of the application program to obtain the second value of the application program; the second values of the application programs are arranged in a descending order to obtain a sorting result; and determining the target application program to be displayed in the application recommendation page from the application programs according to the sorting result, and determining the position of the target application program in the application recommendation page.

Optionally, the apparatus further comprises:

the screening module is used for acquiring a plurality of candidate application programs, determining recommended conversion values corresponding to the candidate application programs respectively, screening candidate application programs with recommended conversion values which do not reach preset recommended index values from the candidate application programs, and taking the screened candidate application programs as application programs to be recommended, wherein the recommended conversion values comprise at least one of application person average download data, application exposure and application resource numerical transfer quantity.

For specific limitations of the application recommendation device, reference may be made to the above limitation of the application recommendation method, and no further description is given here. The respective modules in the above application recommendation apparatus may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in a server, or may be stored in software in a memory in the server, so that the processor may call and execute the operations of the above modules.

The above-mentioned division of each module in the application recommendation device is only used for illustration, and in other embodiments, the application recommendation device may be divided into different modules according to needs to complete all or part of the functions of the application recommendation device.

FIG. 9 is a schematic diagram of an internal structure of a server (or cloud, etc.) in one embodiment. As shown in fig. 9, the server includes a processor and a memory connected through a system bus. Wherein the processor is configured to provide computing and control capabilities to support operation of the entire electronic device. The memory may include a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The computer program is executable by a processor for implementing an application recommendation method provided by the following embodiments. The internal memory provides a cached operating environment for operating system computer programs in the non-volatile storage medium. The server may be implemented as a stand-alone server or as a server cluster composed of a plurality of servers. It will be appreciated by those skilled in the art that the structure shown in fig. 9 is merely a block diagram of a portion of the structure associated with the present application and is not limiting of the server to which the present application is applied, and that a particular server may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.

In one embodiment of the present application, a server is provided, the server comprising a memory and a processor, the memory having stored therein a computer program, the processor, when executing the computer program, performing the steps of:

receiving an acquisition request for an application recommendation page sent by a target terminal, wherein the acquisition request carries a terminal identifier of the target terminal; acquiring user characteristic data corresponding to the target terminal, terminal characteristic data of the target terminal and application characteristic data of each application program to be recommended according to the terminal identification, and acquiring a predicted download probability value corresponding to each application program according to the user characteristic data, the terminal characteristic data and each application characteristic data; and determining a target application program to be displayed in an application recommendation page and the position of the target application program in the application recommendation page based on the predicted download probability value corresponding to each application program, generating the application recommendation page based on the target application program and the position, and sending the application recommendation interface to the target terminal.

In one embodiment, the processor when executing the computer program further performs the steps of:

And for each application program, splicing the user characteristic data, the terminal characteristic data and the application characteristic data of the application program to obtain spliced characteristic data, and inputting the spliced characteristic data into a download probability prediction model to obtain a predicted download probability value of the application program.

acquiring training adjustment samples according to the download data of different terminals to different application programs, wherein the training adjustment samples comprise sample user characteristic data, sample terminal characteristic data, sample application characteristic data of the application programs and sample download probability values; and adjusting model parameters of the download probability prediction model based on the training adjustment sample.

In one embodiment, the user characteristic data includes at least one of application download data, application click data, application browse data, and application search data detected based on the target terminal; the terminal characteristic data comprises at least one of model data and network type data of the target terminal; the application characteristic data of each application program at least comprises application grade data of the application program.

obtaining the number of the predicted resource value transitions corresponding to each application program; for each application program, correcting the predicted download probability value of the application program according to the resource value transfer quantity of the application program to obtain a first value of the application program; and determining a target application program to be displayed in an application recommendation page and the position of the target application program in the application recommendation page based on the first value of each application program.

obtaining conversion rates corresponding to the application programs, wherein the conversion rates are used for representing the ratio of the historical registration quantity of the application programs to the historical download quantity of the application programs; for each application program, correcting the first value of the application program according to the conversion rate of the application program to obtain the second value of the application program; the second values of the application programs are arranged in a descending order to obtain a sorting result; and determining the target application program to be displayed in the application recommendation page from the application programs according to the sorting result, and determining the position of the target application program in the application recommendation page.

acquiring a plurality of candidate application programs, and determining recommended conversion values corresponding to the candidate application programs respectively, wherein the recommended conversion values comprise at least one of application person average download data, application exposure and application resource numerical transfer quantity; and screening candidate application programs of which recommendation conversion values do not reach preset recommendation index values from the candidate application programs, and taking the screened candidate application programs as application programs to be recommended.

Embodiments of the present application also provide a computer-readable storage medium. One or more non-transitory computer-readable storage media containing computer-executable instructions that, when executed by one or more processors, cause the processors to perform the steps of:

In one embodiment, the computer-executable instructions, when executed by one or more processors, cause the processors to further perform the steps of:

The present embodiments also provide a computer program product containing instructions which, when run on a computer, cause the computer to perform the steps of:

In one embodiment, when it is run on a computer, the computer is caused to further perform the steps of:

Any reference to memory, storage, database, or other medium used herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM), which acts as external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), memory bus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

The above examples only represent a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the present application. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application is to be determined by the claims appended hereto.

Claims

1. An application recommendation method, the method comprising:

obtaining the number of the predicted resource value transitions corresponding to each application program;

For each application program, correcting the predicted download probability value of the application program according to the resource value transfer quantity of the application program to obtain a first value of the application program;

and determining a target application program to be displayed in an application recommendation page and the position of the target application program in the application recommendation page based on the first value of each application program, generating the application recommendation page based on the target application program and the position, and sending the application recommendation page to the target terminal.

2. The method of claim 1, wherein the obtaining a predicted download probability value corresponding to each application program according to the user feature data, the terminal feature data, and each application feature data comprises:

3. The method according to claim 2, wherein the method further comprises:

Acquiring training adjustment samples according to the download data of different terminals to different application programs, wherein the training adjustment samples comprise sample user characteristic data, sample terminal characteristic data, sample application characteristic data of the application programs and sample download probability values;

and adjusting model parameters of the download probability prediction model based on the training adjustment sample.

4. The method of claim 1, wherein the user characteristic data includes at least one of application download data, application click data, application browse data, and application search data detected based on the target terminal; the terminal characteristic data comprises at least one of model data and network type data of the target terminal; the application characteristic data of each application program at least comprises application grade data of the application program.

5. The method of claim 1, wherein determining the target application to be presented in the application recommendation page and the location of the target application in the application recommendation page based on the first value of each of the applications comprises:

Obtaining conversion rates corresponding to the application programs, wherein the conversion rates are used for representing the ratio of the historical registration quantity of the application programs to the historical download quantity of the application programs;

for each application program, correcting the first value of the application program according to the conversion rate of the application program to obtain the second value of the application program;

the second values of the application programs are arranged in a descending order to obtain a sorting result;

and determining the target application program to be displayed in the application recommendation page from the application programs according to the sorting result, and determining the position of the target application program in the application recommendation page.

6. The method according to claim 1, wherein before the obtaining, according to the terminal identifier, the user feature data corresponding to the target terminal, the terminal feature data of the target terminal, and the application feature data of each application program to be recommended, the method further includes:

acquiring a plurality of candidate application programs, and determining recommended conversion values corresponding to the candidate application programs respectively, wherein the recommended conversion values comprise at least one of application person average download data, application exposure and application resource numerical transfer quantity;

And screening candidate application programs of which recommendation conversion values do not reach preset recommendation index values from the candidate application programs, and taking the screened candidate application programs as application programs to be recommended.

7. An application recommendation device, the device comprising:

the recommendation module is used for determining a target application program to be displayed in an application recommendation page and the position of the target application program in the application recommendation page based on the predicted download probability value corresponding to each application program, generating the application recommendation page based on the target application program and the position, and sending the application recommendation page to the target terminal;

Wherein, the recommendation module includes:

8. The apparatus of claim 7, wherein the obtaining module is specifically configured to splice, for each application program, the user feature data, the terminal feature data, and application feature data of the application program to obtain spliced feature data, and input the spliced feature data into a download probability prediction model to obtain a predicted download probability value of the application program.

9. A server comprising a memory and a processor, the memory having stored therein a computer program which, when executed by the processor, causes the processor to perform the steps of the method of any of claims 1 to 6.

10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any one of claims 1 to 6.