A method and device for recommending functional modules in an application

Abstract

The present application discloses a method and device for recommending functional modules in an application, which constructs a heterogeneous graph corresponding to the user based on the user, the user's associated users, the user relationship data between the user and the associated users, and the user function interaction data between the user and the functional modules in the application. The heterogeneous graph is represented by a vector, and the vector representing the heterogeneous graph is input into the graph neural network model to obtain the ranking result of the correlation probability between the user and the functional module output by the graph neural network model. Among them, the higher the correlation probability, the higher the user's interest in the functional module. According to the ranking result, the functional module corresponding to the correlation probability greater than the preset threshold is recommended to the user. Since the entire implementation process can be automatically implemented by the terminal, there is no need for the user to manually search and set it, which improves the user's experience of the terminal.

Description

Technical Field

[0001] The present application relates to the field of Internet technology, and in particular to a method and device for recommending functional modules in an application. Background Art

[0002] At present, when an application program on a terminal is opened, a display page pops up with function modules recommended to the user. However, the function modules recommended to the user on the display page are basically fixed, and not all of them are function modules that the user is interested in.

[0003] Currently, if the function modules that users are interested in need to be displayed on the display page, users need to manually search for the function modules they are interested in in the application, and then add the found function modules to the display page. In this process, users may need to jump to multiple pages to search, which is cumbersome and has a poor user experience. Summary of the invention

[0004] In order to solve the above technical problems, the present application provides a method and device for recommending functional modules in an application, which are used to automatically recommend functional modules that users are interested in in the application to users.

[0005] In order to achieve the above purpose, the technical solutions provided by the embodiments of the present application are as follows:

[0006] The present application provides a method for recommending a functional module in an application, the method comprising:

[0007] Constructing a heterogeneous graph corresponding to the user; the nodes in the heterogeneous graph represent the user and the associated users of the user, and the edges in the heterogeneous graph are used to represent the user relationship data between the user and the associated users, and the user function interaction data between the user and the function modules in the application;

[0008] Inputting the vector representing the heterogeneous graph into the graph neural network model, and obtaining the ranking result of the correlation probability between the user and the functional module output by the graph neural network model; the graph neural network model is used to sort and output the correlation probability between the user and the functional module in the heterogeneous graph;

[0009] According to the ranking result, the function modules corresponding to the correlation probabilities greater than a preset threshold are recommended to the user.

[0010] Optionally, constructing a heterogeneous graph of users includes:

[0011] Build a user social network graph of users;

[0012] Constructing a user function interaction graph of the user;

[0013] A heterogeneous graph of the user is constructed according to the user social network graph of the user and the user function interaction graph of the user.

[0014] Optionally, constructing a user social network graph of the user includes:

[0015] Determine the user's associated users;

[0016] Acquiring user relationship data between the user and the associated users;

[0017] A user social network graph of the user is constructed based on the user, the associated users and the user relationship data; the nodes in the user social network graph represent the user and the associated users; and the edges of the user social network graph represent the user relationship data.

[0018] Optionally, constructing a user function interaction graph of the user includes:

[0019] Determine a first functional module used by the user and a second functional module used by an associated user of the user;

[0020] Acquire first user function interaction data between the user and the first function module, and second user function interaction data between an associated user of the user and the second function module;

[0021] A user function interaction graph of the user is constructed based on the user, the user's associated users, the first function module, the second function module, the first user function interaction data and the second user function interaction data; the nodes in the user function interaction graph represent the user, the user's associated users, the first function module and the second function module; the edges of the user represent the first user function interaction data and the second user function interaction data.

[0022] Optionally, the recommending, according to the ranking result, to the user, function modules corresponding to correlation probabilities greater than a preset threshold value, includes:

[0023] According to the ranking result, a function module whose correlation probability is greater than a preset threshold is determined as a target function module;

[0024] Labeling the target function module to obtain a label corresponding to the target function module; wherein the label corresponds to a display position in the recommendation page in a one-to-one manner;

[0025] According to the label corresponding to the target function module, the target function module is displayed at a display position corresponding to the label on the recommendation page.

[0026] Optionally, the method further comprises:

[0027] Acquire first interaction data between the user and a target functional module in the recommendation page, and second interaction data between the user and a non-recommended functional module;

[0028] The heterogeneous graph of the user is updated using the first interaction data and the second interaction data.

[0029] The present application also provides a device for recommending function modules in an application, the device comprising:

[0030] A construction unit, configured to construct a heterogeneous graph of users; the nodes in the heterogeneous graph represent the users and their associated users, and the edges in the heterogeneous graph represent the user relationship data between the users and their associated users, and the user function interaction data between the users and the function modules in the application;

[0031] An input unit, used to input the vector representing the heterogeneous graph into the graph neural network model, and obtain the ranking result of the correlation probability between the user and the functional module output by the graph neural network model; the graph neural network model is used to sort and output the correlation probability between the user and the functional module in the heterogeneous graph;

[0032] A recommendation unit is used to recommend, according to the sorting result, the function modules corresponding to the correlation probabilities greater than a preset threshold to the user.

[0033] Optionally, the construction unit comprises:

[0034] A first construction subunit is used to construct a user social network graph of the user;

[0035] A second construction subunit is used to construct a user function interaction graph of the user;

[0036] The third construction subunit is used to construct a user function heterogeneous graph of the user according to the user social network graph of the user and the user function interaction graph of the user.

[0037] An embodiment of the present application also provides a device for recommending function modules in an application, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for recommending function modules in the application is implemented.

[0038] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device executes the method for recommending function modules in an application.

[0039] Through the above technical solution, it can be seen that the present application has the following beneficial effects:

[0040] An embodiment of the present application provides a method and device for recommending functional modules in an application, which constructs a heterogeneous graph corresponding to the user based on the user, the user's associated users, the user relationship data between the user and the associated users, and the user functional interaction data between the user and the functional modules in the application. The heterogeneous graph is represented by a vector, and the vector representing the heterogeneous graph is input into the graph neural network model to obtain the ranking result of the correlation probability between the user and the functional module output by the graph neural network model. Among them, the higher the correlation probability, the higher the user's interest in the functional module. According to the ranking result, the functional module corresponding to the correlation probability greater than the preset threshold is recommended to the user. Since the entire implementation process can be automatically implemented by the terminal, there is no need for the user to manually search and set it, which improves the user experience of the terminal. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 A schematic diagram of an exemplary application scenario provided for an embodiment of the present application;

[0043] Figure 2 A flowchart of a method for recommending a functional module in an application provided in an embodiment of the present application;

[0044] Figure 3a A schematic diagram of a user social network graph provided in an embodiment of the present application;

[0045] Figure 3b A schematic diagram of a user function interaction diagram provided in an embodiment of the present application;

[0046] Figure 3c A schematic diagram of a heterogeneous graph corresponding to a user provided in an embodiment of the present application;

[0047] Figure 4 A network schematic diagram of a graph neural network model provided in an embodiment of the present application;

[0048] Figure 5 A schematic diagram of the structure of a device for recommending functional modules in an application provided in an embodiment of the present application. DETAILED DESCRIPTION

[0049] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the embodiments of the present application are further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0050] In order to facilitate understanding and explanation of the technical solution provided by the embodiments of the present application, the background technology of the embodiments of the present application is first described below.

[0051] At present, the page display of banking applications on mobile phones and other terminals is basically fixed. When a user logs in to mobile banking, the display page they see when they first enter the mobile banking application is full of functional modules such as inquiry, transfer, financial management, and deposit, and the content and location are fixed.

[0052] Fixed function modules may cause users to need to go through multiple levels of menus or page jumps to reach the function modules they are interested in, which leads to poor user experience. In addition, if there are new or optimized functions in mobile banking, they cannot be recommended more accurately to users who really need them.

[0053] Based on this, an embodiment of the present application provides a method and device for recommending functional modules in an application, which constructs a heterogeneous graph corresponding to the user according to the user, the user's associated users, the user relationship data between the user and the associated users, and the user function interaction data between the user and the functional modules in the application. The heterogeneous graph is represented by a vector, and the vector representing the heterogeneous graph is input into the graph neural network model to obtain the ranking result of the correlation probability between the user and the functional module output by the graph neural network model. Among them, the higher the correlation probability, the higher the user's interest in the functional module. According to the ranking result, the functional module corresponding to the correlation probability greater than the preset threshold is recommended to the user. Since the entire implementation process can be automatically implemented by the terminal, there is no need for the user to manually search and set it, which improves the user experience of the terminal.

[0054] In order to facilitate understanding of the recommended method of the functional module in the application provided in the embodiment of the present application, the following Figure 1 An exemplary application scenario of the method is described. Figure 1 This is a schematic diagram of an exemplary application scenario provided in an embodiment of the present application. The method can be applied in a terminal device 101.

[0055] The terminal device 101 constructs a heterogeneous graph corresponding to the user based on the user, the user's associated users, the user relationship data between the user and the associated users, and the user function interaction data between the user and the function modules in the application.

[0056] The terminal device 101 represents the heterogeneous graph with a vector, inputs the vector representing the heterogeneous graph into the graph neural network model, and obtains the ranking result of the correlation probability between the user and the functional module output by the graph neural network model. The higher the correlation probability, the higher the user's interest in the functional module.

[0057] The terminal device 101 recommends the function modules corresponding to the correlation probabilities greater than the preset threshold to the user according to the ranking results. Thus, the terminal device 101 updates the original display page on the application program to a recommendation page that meets the user's needs.

[0058] Those skilled in the art will understand that Figure 1 The framework diagram shown is only an example in which the embodiments of the present application can be implemented. The scope of application of the embodiments of the present application is not limited by any aspect of the framework.

[0059] Based on the above description, the recommended method of the functional modules in the application provided by this application will be described in detail with reference to the accompanying drawings.

[0060] See also Figure 2 , Figure 2 Flow chart of a method for recommending a functional module in an application provided in an embodiment of the present application. The method can be executed by the aforementioned terminal device 101. Figure 2 As shown, the recommended method of the functional modules in the application includes S201-S203:

[0061] S201: Construct a heterogeneous graph corresponding to the user; the nodes in the heterogeneous graph represent the user and the user's associated users, and the edges in the heterogeneous graph are used to represent the user relationship data between the user and the associated users, and the user function interaction data between the user and the function modules in the application.

[0062] A graph is a data structure composed of nodes and edges. In the heterogeneous graph corresponding to a user, nodes represent users and their associated users, and edges are used to represent user relationship data between users and associated users, as well as user function interaction data between users and function modules in the application. Heterogeneous graphs can accurately depict the relationship between users and associated users, and between users and function modules.

[0063] Among them, the functional modules are mainly constructed according to the functional attributes, and are roughly divided into basic functional modules, common functional modules and advanced functional modules. Basic functional modules include account query function, transfer function, etc., common functional modules include fund fixed investment, baby financial management, etc., and advanced functional modules include new functions launched by banks.

[0064] User relationship data mainly includes data such as transfer time, transfer amount, transfer frequency, etc. between users and their associated users.

[0065] The user-function interaction data mainly refers to the historical information of the user's interaction with the function module, including browsing time, usage frequency or purchasing habits. For example, the browsing time and usage frequency of the user's account query function. Another example is the purchasing habits in fund fixed investment.

[0066] It can be understood that the heterogeneous graph corresponding to the user indicates the association relationship between the user and the associated users, and the association relationship between the user and the functional module.

[0067] In a possible implementation, the embodiment of the present application provides a specific implementation of constructing a heterogeneous graph of users, including:

[0068] A1: Build a user social network graph of users;

[0069] A2: Build a user function interaction diagram for users;

[0070] A3: Construct a heterogeneous graph of users based on their user social network graph and their user function interaction graph.

[0071] The user social network graph is used to describe the relationship between the user and the user's associated users. The user's associated users are some users with whom the user has transaction data, for example, the user's associated users include users with whom the user has transfer transactions.

[0072] The user function interaction diagram is used to describe the user's usage relationship between functional modules.

[0073] After obtaining the user's user social network graph and the user function interaction graph, the user's user social network graph and the user's user function interaction graph can be combined to construct the user's heterogeneous graph.

[0074] In specific implementation, a user social network graph of users is constructed, including:

[0075] B1: Determine the user's associated users;

[0076] B2: Obtain user relationship data between the user and associated users;

[0077] B3: Based on the users, associated users and user relationship data, a user social network graph of the users is constructed; the nodes in the user social network graph represent users and associated users; the edges of the user social network graph represent user relationship data.

[0078] First, obtain the associated users with whom the user has transactions. After determining the associated users of the user. Obtain the user relationship data such as the transfer time, transfer amount, and transfer frequency of transactions between the user and the user's associated users. Use the user and the user's associated users as nodes of the graph, and the user relationship data as the edges of the graph to build the user's user social network graph. See Figure 3a , Figure 3a A schematic diagram of a user social network diagram provided in an embodiment of the present application. Figure 3a As shown in FIG, the nodes in the user social network graph represent users and their associated users, and the edges represent user relationship data.

[0079] During the specific implementation, a user function interaction diagram is constructed, including:

[0080] C1: Determine the first functional module used by the user and the second functional module used by the user's associated users;

[0081] C2: Acquire first user function interaction data between the user and the first function module, and second user function interaction data between the associated user of the user and the second function module;

[0082] C3: Construct a user function interaction graph of the user according to the user, the user's associated users, the first function module, the second function module, the first user function interaction data and the second user function interaction data; the nodes in the user function interaction graph represent the user, the user's associated users, the first function module and the second function module; the edges of the user represent the first user function interaction data and the second user function interaction data.

[0083] First, determine the first function module used by the user in the application and the second function module used by the user's associated users. After determining the first function module used by the user and the second function module used by the user's associated users, obtain the first user function interaction data such as the user's browsing time, usage frequency and purchasing habits on the first function module, and obtain the second user function interaction data such as the user's associated users on the second function module. Use the user, the user's associated users, the first function module and the second function module as the nodes of the graph, and use the first user function interaction data and the second user function interaction data as the edges of the graph to construct the user's user function interaction graph. See Figure 3b , Figure 3b A schematic diagram of a user function interaction diagram provided in an embodiment of the present application. Figure 3b As shown, the circular nodes in the user function interaction graph represent users and their associated users, and the square nodes represent the first function module and the second function module. The edges between the circular nodes and the square nodes represent the first user function interaction data and the second user function interaction data.

[0084] See also Figure 3c , Figure 3c A schematic diagram of a heterogeneous graph corresponding to a user provided in an embodiment of the present application. Figure 3cAs shown, the figure is a heterogeneous graph corresponding to the user, which is composed of the user social network graph of the user and the user function interaction graph of the user. In the figure, the circular nodes represent the user and the associated user of the user, and the square nodes represent the first function module and the second function module. The edges between the circular nodes represent the user relationship data. The edges between the circular nodes and the square nodes represent the first user function interaction data and the second user function interaction data.

[0085] S202: Input the vector representing the heterogeneous graph into the graph neural network model to obtain the ranking result of the correlation probabilities between users and functional modules output by the graph neural network model; the graph neural network model is used to sort and output the correlation probabilities between users and functional modules in the heterogeneous graph.

[0086] After obtaining the heterogeneous graph, the heterogeneous graph is represented by a vector. Specifically, the vector consists of two parts: one is a matrix composed of node information of each node, and the other is an adjacency matrix composed of edge relationships. The vector representing the heterogeneous graph is input into the graph neural network model to obtain the ranking structure of the correlation probability between users and functional modules.

[0087] Specifically, the graph neural network model is a model that runs directly on the graph structure. It is a deep learning-based method that runs on the graph structure. Using the graph neural network model, the node features of the node corresponding to the user and the node features of the neighboring nodes are iteratively aggregated to represent the node corresponding to the user. After iterative aggregation, the relationship between the node corresponding to the user and the functional module is calibrated in the form of correlation probability.

[0088] See also Figure 4 , Figure 4 A network diagram of a graph neural network model provided in an embodiment of the present application. Figure 4 As shown in the figure, the vector represented by the heterogeneous graph is input into the graph neural network model, and after the operation of the activation function in the double hidden layer, the correlation probability between the user and the functional module is finally output. The edges with correlation probabilities greater than the preset threshold are marked with five-pointed stars. It can be understood that Figure 4 As an example only, the network structure of the graph neural network model is not limited to Figure 4 In the example, the identification method of the edge with a correlation probability greater than the preset threshold is not limited to Figure 4 Examples in .

[0089] It should be noted that different graph neural network model network structures and algorithms can be selected as needed. This application does not limit the network structure of the graph neural network model, as long as the graph neural network can output the correlation probability between users and functional modules. For example, an encoder-based method or a relational graph convolutional network R-GCN can be selected.

[0090] The embodiment of the present application is based on a heterogeneous graph and a graph neural network model, which can fully learn the node information and edge association information of the graph and the topological structure information of the graph, avoiding the large errors in the recommendation results caused by manual intervention and manual parameter adjustment in the traditional recommendation algorithm, and more accurately recommends functional modules to users.

[0091] S203: According to the ranking result, the function modules corresponding to the correlation probabilities greater than a preset threshold are recommended to the user.

[0092] After the ranking result is determined, the ranking result of the function modules commonly used by the user is determined. A preset threshold is set, and the function modules corresponding to the correlation probability greater than the preset threshold are recommended to the user. Among them, the preset threshold is set according to the actual situation and is not limited to the preset threshold.

[0093] In a possible implementation, the embodiment of the present application provides a specific implementation of recommending function modules corresponding to correlation probabilities greater than a preset threshold to a user according to the sorting results, including:

[0094] According to the sorting results, the function modules whose correlation probability is greater than a preset threshold are determined as target function modules;

[0095] Label the target functional module and obtain the label corresponding to the target functional module; the label corresponds to the display position in the recommendation page one by one;

[0096] According to the label corresponding to the target function module, the target function module is displayed at the display position corresponding to the label on the recommendation page.

[0097] It is understandable that the target function module is a function module that meets the user's interest requirements. The target function module needs to be displayed on the display page. A label can be added to each target function module by adding a label to the target function module. The label and the display position of the display page correspond one to one, and their relationship needs to be set in advance. For example, if the label of one of the target function modules is 1, then the target function module is added to the display position corresponding to "1".

[0098] Based on the above description, the functional modules that meet the user's interest needs can be displayed to the user in a customized manner, making the display of the functional modules in the application more flexible.

[0099] An embodiment of the present application provides a method for recommending functional modules in an application, which constructs a heterogeneous graph corresponding to the user based on the user, the user's associated users, the user relationship data between the user and the associated users, and the user function interaction data between the user and the functional modules in the application. The heterogeneous graph is represented by a vector, and the vector representing the heterogeneous graph is input into the graph neural network model to obtain the ranking result of the correlation probability between the user and the functional module output by the graph neural network model. Among them, the higher the correlation probability, the higher the user's interest in the functional module. According to the ranking result, the functional module corresponding to the correlation probability greater than the preset threshold is recommended to the user. Since the entire implementation process can be automatically implemented by the terminal, there is no need for the user to manually search and set it, which improves the user experience of the terminal.

[0100] In a possible implementation, the present application embodiment further provides another function module recommendation in an application, the method further comprising:

[0101] Acquire first interaction data between the user and the target functional module in the recommended page, and second interaction data between the user and the non-recommended functional module;

[0102] The user's heterogeneous graph is updated using the first interaction data and the second interaction data.

[0103] It is understandable that after the recommended function modules are displayed according to the user's interest needs, when the user faces the recommendation page (i.e., the display page), he may choose to interact with the function modules on the recommendation page, or he may choose to interact with the non-recommended function modules on the non-recommended page. Based on this, the first interaction data between the user and the target function module in the recommendation page and the second interaction data between the user and the non-recommended function module can be recorded. The first interaction data and the second interaction data enrich the original data related to the user, and the user's original data and the added first interaction data and second interaction data are used to update the user's heterogeneous graph, thereby obtaining a more accurate graph structure that describes the user. Based on this graph structure, the graph neural network model can be used to re-evaluate the function modules that meet the user's interest needs, and based on this, the recommendation page for the user can be updated.

[0104] Based on the above description, when the user's associated users change, there are new functional modules in the application, or the interaction between the user and the functional module changes, the user data is updated synchronously in real time, and then the heterogeneous graph corresponding to the user is rebuilt, and the graph neural network module is reused to compound the functional modules of the user's interest needs, and then the recommendation page is updated again according to the recommended functional modules. In this way, the user's demand for functional modules is met and the user's experience of using the application is improved.

[0105] Based on the method for recommending function modules in an application provided by the above method embodiment, the device for recommending function modules in an application provided by an embodiment of the present application will be described below in conjunction with the accompanying drawings.

[0106] See also Figure 5 , Figure 5 This is a schematic diagram of a structure of a device for recommending a functional module in an application provided by an embodiment of the present application. Figure 5 As shown, the recommended devices of the functional modules in this application include:

[0107] A construction unit 501 is used to construct a heterogeneous graph of users; the nodes in the heterogeneous graph represent the users and the associated users of the users, and the edges in the heterogeneous graph represent the user relationship data between the users and the associated users, and the user function interaction data between the users and the function modules in the application;

[0108] An input unit 502 is used to input the vector representing the heterogeneous graph into the graph neural network model, and obtain the ranking result of the correlation probability between the user and the functional module output by the graph neural network model; the graph neural network model is used to sort and output the correlation probability between the user and the functional module in the heterogeneous graph;

[0109] The recommendation unit 503 is used to recommend the function modules corresponding to the correlation probabilities greater than a preset threshold to the user according to the sorting results.

[0110] Optionally, in some implementations of the embodiments of the present application, the construction unit 501 includes:

[0111] A first construction subunit is used to construct a user social network graph of the user;

[0112] A second construction subunit is used to construct a user function interaction graph of the user;

[0113] The third construction subunit is used to construct a user function heterogeneous graph of the user according to the user social network graph of the user and the user function interaction graph of the user.

[0114] Optionally, in some implementations of the embodiments of the present application, the first construction subunit includes:

[0115] A first determining subunit, used to determine associated users of a user;

[0116] A first acquisition subunit, used to acquire user relationship data between the user and the associated user;

[0117] The fourth construction subunit is used to construct a user social network graph of the user based on the user, the associated users and the user relationship data; the nodes in the user social network graph represent the user and the associated users; the edges of the user social network graph represent the user relationship data.

[0118] Optionally, in some implementations of the embodiments of the present application, the second construction subunit includes:

[0119] A second determining subunit is used to determine a first functional module used by the user and a second functional module used by an associated user of the user;

[0120] A second acquisition subunit is used to acquire first user function interaction data between the user and the first function module, and second user function interaction data between an associated user of the user and the second function module;

[0121] The fifth construction subunit is used to construct a user function interaction graph of the user based on the user, the user's associated users, the first function module, the second function module, the first user function interaction data and the second user function interaction data; the nodes in the user function interaction graph represent the user, the user's associated users, the first function module and the second function module; the edges of the user represent the first user function interaction data and the second user function interaction data.

[0122] Optionally, in some implementations of the embodiments of the present application, the recommendation unit 503 includes:

[0123] A third determining subunit is used to determine, according to the sorting result, a function module whose correlation probability is greater than a preset threshold as a target function module;

[0124] A third acquisition subunit is used to label the target function module and obtain a label corresponding to the target function module; the label corresponds to a display position in the recommendation page one by one;

[0125] The display subunit is used to display the target function module at a display position corresponding to the tag on the recommendation page according to the tag corresponding to the target function module.

[0126] Optionally, in some implementations of the embodiments of the present application, the device further includes:

[0127] An acquisition unit, configured to acquire first interaction data between the user and a target functional module in the recommendation page, and second interaction data between the user and a non-recommended functional module;

[0128] An updating unit is used to update the heterogeneous graph of the user by using the first interaction data and the second interaction data.

[0129] In addition, an embodiment of the present application also provides a device for recommending functional modules in an application, characterized in that it includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, and when the processor executes the computer program, it implements the method for recommending functional modules in an application as described in any of the above embodiments.

[0130] In addition, an embodiment of the present application also provides a computer-readable storage medium, characterized in that instructions are stored in the computer-readable storage medium, and when the instructions are executed on a terminal device, the terminal device executes the function module recommendation method in the application as described in any of the above embodiments.

[0131] An embodiment of the present application provides a device for recommending functional modules in an application, which constructs a heterogeneous graph corresponding to the user based on the user, the user's associated users, the user relationship data between the user and the associated users, and the user function interaction data between the user and the functional modules in the application. The heterogeneous graph is represented by a vector, and the vector representing the heterogeneous graph is input into the graph neural network model to obtain the ranking result of the correlation probability between the user and the functional module output by the graph neural network model. Among them, the higher the correlation probability, the higher the user's interest in the functional module. According to the ranking result, the functional module corresponding to the correlation probability greater than the preset threshold is recommended to the user. Since the entire implementation process can be automatically implemented by the terminal, there is no need for the user to manually search and set it, which improves the user experience of the terminal.

[0132] It can be known from the description of the above implementation mode that those skilled in the art can clearly understand that all or part of the steps in the above-mentioned embodiment method can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product can be stored in a storage medium such as ROM / RAM, a disk, an optical disk, etc., including several instructions for enabling a computer device (which can be a personal computer, a server, or a network communication device such as a media gateway, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.

[0133] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the method disclosed in the embodiment, since it corresponds to the system disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the system part description.

[0134] It should also be noted that, in this article, the term "includes", "comprising" or any other variation thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of more restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device including the element.

[0135] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for recommending functional modules in an application, characterized in that: The method comprises: Constructing a user social network graph of the user; the nodes in the user social network graph represent the user and associated users; the edges of the user social network graph represent user relationship data; the associated users are users who have transactions with the user; Constructing a user function interaction graph of the user; the nodes in the user function interaction graph represent the user, the user's associated users, the first function module and the second function module; the edges of the user represent the first user function interaction data and the second user function interaction data; Constructing a heterogeneous graph of the user according to the user social network graph of the user and the user function interaction graph of the user; the nodes in the heterogeneous graph represent the user and the associated users of the user, and the edges in the heterogeneous graph are used to represent the user relationship data between the user and the associated users, and the user function interaction data between the user and the function modules in the application; The vector representing the heterogeneous graph is input into the graph neural network model, and the node features of the node corresponding to the user and the node features of the neighboring nodes are iteratively aggregated to obtain the sorting result of the correlation probability between the user and the functional module; the graph neural network model is used to sort and output the correlation probability between the user and the functional module in the heterogeneous graph; the vector includes a matrix composed of node information of each node and an adjacency matrix composed of edge relationships; According to the ranking result, the function modules corresponding to the correlation probabilities greater than a preset threshold are recommended to the user.

2. The method according to claim 1, characterized in that The step of constructing a user social network graph of a user includes: Determine the user's associated users; Acquiring user relationship data between the user and the associated users; A user social network graph of the user is constructed according to the user, the associated users and the user relationship data.

3. The method according to claim 1, characterized in that The constructing the user function interaction graph of the user includes: Determine a first functional module used by the user and a second functional module used by an associated user of the user; Acquire first user function interaction data between the user and the first function module, and second user function interaction data between an associated user of the user and the second function module; A user function interaction graph of the user is constructed according to the user, the associated users of the user, the first function module, the second function module, the first user function interaction data and the second user function interaction data.

4. The method according to claim 1, characterized in that: The recommending, according to the ranking result, to the user, function modules corresponding to correlation probabilities greater than a preset threshold value, comprises: According to the ranking result, a function module whose correlation probability is greater than a preset threshold is determined as a target function module; Labeling the target function module to obtain a label corresponding to the target function module; wherein the label corresponds to a display position in the recommendation page in a one-to-one manner; According to the label corresponding to the target function module, the target function module is displayed at a display position corresponding to the label on the recommendation page.

5. The method according to claim 4, characterized in that The method further comprises: Acquire first interaction data between the user and a target functional module in the recommendation page, and second interaction data between the user and a non-recommended functional module; The heterogeneous graph of the user is updated using the first interaction data and the second interaction data.

6. A device for recommending function modules in an application, characterized in that: The device comprises: A construction unit is used to construct a user social network graph of a user; the nodes in the user social network graph represent the user and associated users; the edges of the user social network graph represent the user relationship data; the associated users are users who have transactions with the user; a user function interaction graph of the user is constructed; the nodes in the user function interaction graph represent the user, the associated users of the user, the first function module and the second function module; the edges of the user represent the first user function interaction data and the second user function interaction data; a heterogeneous graph of the user is constructed according to the user social network graph of the user and the user function interaction graph of the user, wherein the edges in the heterogeneous graph are used to represent the user relationship data between the user and the associated users, and the user function interaction data between the user and the function modules in the application; An input unit, used to input the vector representing the heterogeneous graph into the graph neural network model, iteratively aggregate the node features of the node corresponding to the user and the node features of the neighboring nodes, and obtain the sorting result of the correlation probability between the user and the functional module; the graph neural network model is used to sort and output the correlation probability between the user and the functional module in the heterogeneous graph; the vector includes a matrix composed of node information of each node, and an adjacency matrix composed of edge relationships; A recommendation unit is used to recommend, according to the sorting result, the function modules corresponding to the correlation probabilities greater than a preset threshold to the user.

7. A function module recommendation device in an application, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for recommending function modules in an application as described in any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed on a terminal device, the terminal device executes the method for recommending function modules in an application as described in any one of claims 1 to 5.

Images