User behavior identification method, device, computer equipment and storage medium

Abstract

The present application relates to a method, apparatus, computer equipment, storage medium and computer program product for identifying user behavior, and relates to the field of computer technology, and can be used in the field of financial technology or other related fields. The method includes: determining at least one behavioral feature of each user behavior based on multiple user behaviors obtained from multiple channels; constructing a relationship topology diagram between each user behavior based on at least one behavioral feature of each user behavior; each main node in the relationship topology diagram corresponds to a user behavior, and each main node is associated with at least one child node corresponding to the behavioral feature of the user behavior; traversing each child node in the relationship topology diagram, and obtaining multiple target behavior feature sets based on the connection relationship between each child node; and confirming the user behavior corresponding to the behavioral feature in each target behavior feature set as the user behavior of the same public account. The use of this method can improve the recognition accuracy of user behavior.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, apparatus, computer device, storage medium, and computer program product for identifying user behavior. Background Art

[0002] A public account is an account for an enterprise or institution. The account holder is a single enterprise or institution. However, a public account can be associated with multiple users, and different users exhibit different user behaviors. To provide better services to public account users, it's necessary to identify public accounts based on user behavior.

[0003] In related technologies, relational databases are typically used to match user behaviors, identifying user behaviors with the same characteristics as belonging to the same public account. However, the user behaviors captured before and after a public account login vary, and the user behaviors captured through different login channels also vary. Therefore, relational databases struggle to strictly match user behaviors, resulting in low accuracy in user behavior recognition based on relational databases. Summary of the Invention

[0004] Based on this, it is necessary to provide a user behavior recognition method, device, computer equipment, computer-readable storage medium and computer program product that can improve the recognition accuracy of user behavior in response to the above technical problems.

[0005] In a first aspect, the present application provides a method for identifying user behavior. The method comprises:

[0006] Determining at least one behavior feature of each user behavior based on multiple user behaviors obtained from multiple channels;

[0007] Based on at least one behavioral feature of each user behavior, a relationship topology diagram between the user behaviors is constructed; the relationship topology diagram includes a plurality of main nodes, each main node corresponds to a user behavior, and each main node is associated with at least one child node, each child node corresponds to a behavioral feature of the associated user behavior;

[0008] Traversing each sub-node in the relationship topology graph, and obtaining a plurality of target behavior feature sets based on the connection relationships between the sub-nodes;

[0009] The user behavior corresponding to the behavioral features in each target behavioral feature set is confirmed as the user behavior of the same public account.

[0010] In one embodiment, traversing each sub-node in the relationship topology graph and obtaining multiple target behavior feature sets based on the connection relationships between the sub-nodes include:

[0011] Each time, a child node in the relationship topology graph is used as a starting node, and based on the connection relationship between the starting node and other nodes, a neighbor node of the starting node is determined;

[0012] Starting from the starting node, performing a depth-first traversal on the neighboring nodes to obtain a traversal path of the starting node;

[0013] Based on the behavior features corresponding to the child nodes on the traversal path of the starting node, a target behavior feature set of the starting node is obtained.

[0014] In one embodiment, traversing each sub-node in the relationship topology graph and obtaining multiple target behavior feature sets based on the connection relationships between the sub-nodes further includes:

[0015] Traversing each sub-node in the relationship topology graph, and obtaining a plurality of initial behavior feature sets based on the connection relationships between the sub-nodes;

[0016] Screening is performed on each initial behavior feature set to obtain the multiple target behavior feature sets.

[0017] In one embodiment, the screening of the initial behavior feature sets to obtain the multiple target behavior feature sets includes:

[0018] For each initial behavioral feature set, determine the weight and number of occurrences corresponding to the connection edge connecting each behavioral feature in the initial behavioral feature set; the number of occurrences represents the cumulative number of associations between the two behavioral features connected by the connection edge within a preset time period;

[0019] Based on the weights and occurrence counts corresponding to the connecting edges, target behavior features are screened out from the initial behavior feature set to obtain a target behavior feature set corresponding to the initial behavior feature set.

[0020] In one embodiment, determining the weight corresponding to the edge connecting each behavior feature in the initial behavior feature set includes:

[0021] Determining an account user associated with each behavioral feature in the initial behavioral feature set;

[0022] Based on the preset priorities of the account users and the account users associated with the various behavioral features in the initial behavioral feature set, the weights corresponding to the connecting edges between the various behavioral features in the initial behavioral feature set are determined.

[0023] In one embodiment, the screening of the initial behavior feature sets to obtain the multiple target behavior feature sets further includes:

[0024] identifying, from the respective initial behavioral feature sets, a plurality of repeated behavioral feature sets including identical behavioral features;

[0025] Deduplication processing is performed on each group of repeated behavior feature sets to obtain the multiple target behavior feature sets.

[0026] In one embodiment, constructing a relationship topology graph between the user behaviors based on at least one behavior feature of the user behaviors includes:

[0027] For each user behavior, taking the user behavior as a main node and at least one behavior feature of the user behavior as a child node of the main node, a relationship point graph between the user behaviors is obtained;

[0028] In the relationship point graph, the main node and the child nodes of the main node are connected to obtain a relationship topology graph between the various user behaviors.

[0029] In a second aspect, the present application also provides a user behavior recognition device. The device includes:

[0030] A behavior feature determination module, configured to determine at least one behavior feature of each user behavior based on multiple user behaviors acquired from multiple channels;

[0031] A topology relationship building module is configured to build a relationship topology graph between the user behaviors based on at least one behavior feature of the user behaviors; the relationship topology graph includes a plurality of main nodes, each main node corresponds to a user behavior, and each main node is associated with at least one child node, each child node corresponds to a behavior feature of the associated user behavior;

[0032] A topology relationship traversal module, configured to traverse each sub-node in the relationship topology graph and obtain a plurality of target behavior feature sets based on the connection relationships between the sub-nodes;

[0033] The user behavior identification module is used to confirm the user behavior corresponding to the behavior feature in each target behavior feature set as the user behavior of the same public account.

[0034] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are performed:

[0035] Determining at least one behavior feature of each user behavior based on multiple user behaviors obtained from multiple channels;

[0036] Based on at least one behavioral feature of each user behavior, a relationship topology diagram between the user behaviors is constructed; the relationship topology diagram includes a plurality of main nodes, each main node corresponds to a user behavior, and each main node is associated with at least one child node, each child node corresponds to a behavioral feature of the associated user behavior;

[0037] Traversing each sub-node in the relationship topology graph, and obtaining a plurality of target behavior feature sets based on the connection relationships between the sub-nodes;

[0038] The user behavior corresponding to the behavioral features in each target behavioral feature set is confirmed as the user behavior of the same public account.

[0039] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the following steps:

[0040] Determining at least one behavior feature of each user behavior based on multiple user behaviors obtained from multiple channels;

[0041] Based on at least one behavioral feature of each user behavior, a relationship topology diagram between the user behaviors is constructed; the relationship topology diagram includes a plurality of main nodes, each main node corresponds to a user behavior, and each main node is associated with at least one child node, each child node corresponds to a behavioral feature of the associated user behavior;

[0042] Traversing each sub-node in the relationship topology graph, and obtaining a plurality of target behavior feature sets based on the connection relationships between the sub-nodes;

[0043] The user behavior corresponding to the behavioral features in each target behavioral feature set is confirmed as the user behavior of the same public account.

[0044] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the following steps:

[0045] Determining at least one behavior feature of each user behavior based on multiple user behaviors obtained from multiple channels;

[0046] Based on at least one behavioral feature of each user behavior, a relationship topology diagram between the user behaviors is constructed; the relationship topology diagram includes a plurality of main nodes, each main node corresponds to a user behavior, and each main node is associated with at least one child node, each child node corresponds to a behavioral feature of the associated user behavior;

[0047] Traversing each sub-node in the relationship topology graph, and obtaining a plurality of target behavior feature sets based on the connection relationships between the sub-nodes;

[0048] The user behavior corresponding to the behavioral features in each target behavioral feature set is confirmed as the user behavior of the same public account.

[0049] The above-mentioned user behavior identification method, apparatus, computer equipment, storage medium and computer program product first determine at least one behavioral feature of each user behavior based on multiple user behaviors obtained from multiple channels; then, based on at least one behavioral feature of each user behavior, construct a relationship topology diagram between each user behavior; the relationship topology diagram includes multiple main nodes, each main node corresponds to a user behavior, and each main node is associated with at least one child node, and each child node corresponds to a behavioral feature of the associated user behavior; then, each child node in the relationship topology diagram is traversed, and based on the connection relationship between each child node, multiple target behavior feature sets are obtained; finally, the user behavior corresponding to the behavioral feature in each target behavior feature set is confirmed as the user behavior of the same public account. In this way, based on the connection relationship between the child nodes representing the behavioral features in the relationship topology diagram, the correlation between the various behavioral features can be fully explored, thereby exploring the correlation between the various user behaviors, and then the user behavior belonging to the same public account can be identified from multiple user behaviors. Compared with the user behavior identification method based on relational database that relies on data structure, the user behavior identification method based on the above process can accurately identify each user behavior and its behavior characteristics with different data structures through the relational topology graph, thereby improving the recognition accuracy of user behavior. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 1 is a flow chart of a method for identifying user behavior in one embodiment;

[0051] Figure 2 A schematic diagram of a relationship topology diagram between various user behaviors in one embodiment;

[0052] Figure 3 1. A flowchart illustrating steps for obtaining multiple target behavior feature sets based on the connection relationship between each sub-node in one embodiment;

[0053] Figure 41. A flowchart illustrating steps of screening each initial behavior feature set to obtain multiple target behavior feature sets in one embodiment;

[0054] Figure 5 is a flowchart of a method for identifying user behavior in another embodiment;

[0055] Figure 6 1 is a flow chart of a method for identifying public accounts based on point-edge search in one embodiment;

[0056] Figure 7 is a structural block diagram of a user behavior recognition device in one embodiment;

[0057] Figure 8 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0058] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0059] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions.

[0060] It should also be noted that the user behavior identification method, device, computer equipment, storage medium and computer program product provided by this application can be applied to the field of financial technology, such as determining whether the behaviors of multiple users logging into the bank system belong to the same corporate account; it can also be applied to other related fields, such as in the field of computer technology, through the user behavior identification method provided by this application, the corporate account to which the user behavior belongs can be identified.

[0061] It's understood that a public account is an account for an enterprise or institution, and the holder of a public account is a single enterprise or institution. A public account can be associated with multiple account users, such as the legal representative, financial director, and agent, and different account users will exhibit different user behaviors. If we can accurately identify user behaviors belonging to the same public account from different user behaviors, we can provide targeted services to account users.

[0062] In an exemplary embodiment, Figure 1As shown, a method for identifying user behavior is provided. This embodiment uses the method applied to a server as an example for illustration; it is understandable that the method can also be applied to a terminal, and can also be applied to a system including a server and a terminal, and implemented through interaction between the server and the terminal. The server can be implemented as an independent server or a server cluster consisting of multiple servers; the terminal can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, etc. In this embodiment, the method includes the following steps:

[0063] Step S102: determining at least one behavior feature of each user behavior based on multiple user behaviors acquired from multiple channels.

[0064] Among them, multi-channel refers to multiple channels for using an account, such as mobile applications, computer applications, mini-programs on WeChat, and web pages.

[0065] Among them, user behavior is used to characterize the behavior generated by the account user when using the account, such as identity authentication behavior when logging into the account, browsing behavior during the use of the account, and operation behavior during the use of the account.

[0066] Among them, the behavioral characteristics of user behavior are used to characterize the characteristic information of user behavior, such as the channels used by the account user, the user ID of the account user, the mobile phone number and device ID used by the account user in the process of using the account, etc.

[0067] Specifically, the server first collects user behaviors on multiple channels using the account, and determines at least one behavior feature of each user behavior.

[0068] For example, assuming that the user behavior 1 collected by the server on the application is logging into the account through the first mobile phone number and the first user ID on the first device, then the server can use the first device, the first mobile phone number and the first user ID as behavioral features of user behavior 1.

[0069] Step S104: construct a relationship topology diagram between the user behaviors based on at least one behavior feature of the user behaviors.

[0070] The relationship topology diagram includes multiple main nodes, each main node corresponds to a user behavior, each main node is associated with at least one child node, and each child node corresponds to a behavior feature of the associated user behavior.

[0071] Specifically, the server uses each user behavior and the behavior characteristics of each user behavior as a node, and uses the relationship between each user behavior and its corresponding behavior characteristics to construct a relationship topology diagram that represents the association relationship between each user behavior.

[0072] For example, assuming that the user behaviors collected by the server are: user behavior 1 of logging into the account on the first device through the first mobile phone number and the first user identifier, user behavior 2 of logging into the account on the second device through the first mobile phone number and the second user identifier, user behavior 3 of logging into the account on the second device through the second mobile phone number and the first user identifier, user behavior 4 of logging into the account through the second mobile phone number and the third user identifier, user behavior 5 of logging into the account on the third device through the third mobile phone number and the fourth user identifier, user behavior 6 of logging into the account through the fourth mobile phone number and the fourth user identifier, and user behavior 7 of logging into four accounts on the third device through the second mobile phone number and the fourth user identifier; then the server constructs the following Figure 2 The relationship topology diagram shown.

[0073] Step S106 , traversing each child node in the relationship topology graph, and obtaining multiple target behavior feature sets based on the connection relationships between each child node.

[0074] Specifically, the server traverses each child node in the relationship topology graph, and obtains target behavior feature sets corresponding to each of the multiple child nodes based on the connection relationships between the child nodes.

[0075] It can be understood that in the relationship topology diagram, the child nodes (behavioral characteristics) are directly connected to the main node (user behavior), and different user behaviors can have the same behavioral characteristics. Therefore, the connection relationship between each child node is represented by the connection between the main node and the child node.

[0076] For an example, see Figure 2 Taking the node representing the first device as an example, the server obtains the target behavior feature set corresponding to the node representing the first device as {first device, first user identifier, first mobile phone number, second user identifier, second device} based on the connection relationship between the node representing the first device and the remaining sub-nodes; taking the node representing the fourth user identifier as an example, the server obtains the target behavior feature set corresponding to the node representing the fourth user identifier as {fourth user identifier, second mobile phone number, third mobile phone number, third device} based on the connection relationship between the node representing the fourth user identifier and the remaining sub-nodes.

[0077] Step S108 , confirming the user behavior corresponding to each behavior feature in the target behavior feature set as the user behavior of the same public account.

[0078] Among them, the user behavior corresponding to the behavior feature in the target behavior feature set refers to the user behavior composed of the various behavior features in the target behavior feature set; taking the target behavior feature set {first device, first user identifier, first mobile phone number, second user identifier, second device} of the node representing the first device as an example, in the target behavior feature set, the first device, the first mobile phone number and the first user identifier constitute user behavior 1, and the first mobile phone number, the second device and the second user identifier constitute user behavior 2; therefore, the server can determine that user behavior 1 and user behavior 2 belong to the user behavior of the same public account. It can be understood that although the first user identifier behavior feature and the second device behavior feature of user behavior 3 exist in the target behavior feature set corresponding to the node representing the first device, since the behavior feature of the second mobile phone number of user behavior 3 does not exist, user behavior 3 does not belong to the user behavior corresponding to the behavior feature in the target behavior feature set corresponding to the node representing the first device.

[0079] Specifically, for each target behavior feature set, the server determines the user behavior that can be composed of the various behavior features in the target behavior feature set based on the various behavior features in the target behavior feature set, as the user behavior corresponding to the behavior feature in the target behavior feature set, and confirms the user behavior corresponding to the behavior feature in the target behavior feature set as the user behavior of the same public account.

[0080] For example, taking the target behavior feature set {fourth user identifier, second mobile phone number, third mobile phone number, third device} of the node representing the fourth user identifier as an example, in this target behavior feature set, the fourth account, the second mobile phone number and the third device constitute user behavior 7, and the fourth user identifier, the third mobile phone number and the third device constitute user behavior 5. Therefore, the server can determine that user behavior 5 and user behavior 7 belong to user behaviors of the same public account.

[0081] Furthermore, after confirming the user behavior of different corporate accounts, the server can also determine the usage channels of different corporate accounts based on the user behavior, and establish a mapping relationship between corporate accounts, user behavior and usage channels, providing technical support for the full-journey user identification of corporate digital operations, and facilitating the provision of personalized services to different account users of corporate accounts in a targeted manner.

[0082] In the user behavior identification method provided by the above embodiment, the server first determines at least one behavior feature of each user behavior based on multiple user behaviors obtained from multiple channels; then, based on the at least one behavior feature of each user behavior, a relationship topology graph is constructed between the user behaviors; the relationship topology graph includes multiple main nodes, each main node corresponds to a user behavior, and each main node is associated with at least one child node, each child node corresponds to a behavior feature of the associated user behavior; then, each child node in the relationship topology graph is traversed, and based on the connection relationship between each child node, multiple target behavior feature sets are obtained; finally, the user behavior corresponding to the behavior feature in each target behavior feature set is confirmed as the user behavior of the same public account. In this way, based on the child nodes representing the behavior features in the relationship topology graph, the server can fully mine the correlation between the various behavior features, thereby mining the correlation between the various user behaviors, and then identifying the user behaviors belonging to the same public account from multiple user behaviors. Compared with the user behavior identification method based on relational databases that relies on data structure, the user behavior identification method based on the above process can accurately identify each user behavior and its behavior features with different data structures through the relationship topology graph, thereby improving the recognition accuracy of user behavior.

[0083] like Figure 3 As shown, in an exemplary embodiment, the above step S106 traverses each child node in the relationship topology graph and obtains multiple target behavior feature sets based on the connection relationship between each child node, specifically including the following steps:

[0084] Step S302 : Each time, a child node in the relationship topology graph is used as a starting node, and based on the connection relationship between the starting node and other nodes, the neighbor nodes of the starting node are determined.

[0085] Step S304 , starting from the starting node, performing a depth-first traversal on neighboring nodes to obtain a traversal path of the starting node.

[0086] Step S306 : obtaining a target behavior feature set of the starting node based on the behavior features corresponding to the child nodes on the traversal path of the starting node.

[0087] The neighbor nodes of the starting node are nodes directly connected to the starting node, that is, the master nodes directly connected to the starting node; Figure 2 Taking the node representing the first user identifier as an example, the neighboring nodes of this node are the node representing user behavior 1 and the node representing user behavior 3.

[0088] Specifically, the server first identifies any child node in the relationship topology graph as the starting node, and based on the connection relationship between the starting node and the remaining nodes, determines the main node directly connected to the starting node as the neighbor node of the starting node; then, starting from the starting node, the server performs a depth-first traversal on each neighbor node to obtain each traversal path of the starting node; then, the server obtains the target behavior feature set of the starting node based on the behavior features corresponding to each child node on each traversal path of the starting node; then, the server continues to select any child node as the new starting node from each child node except the starting node, and repeats the above steps until the target behavior feature set of all child nodes is obtained.

[0089] See Figure 2 Taking the node representing the first user identifier as an example, the server first determines that the neighbor nodes of the node representing the first user identifier are the node representing user behavior 1 and the node representing user behavior 3 based on the connection relationship between the node representing the first user identifier and each node; then, the server starts from the node representing the first user identifier and performs a depth-first traversal on the node representing user behavior 1, and then starts from the node representing the first user identifier and performs a depth-first traversal on the node representing user behavior 3, thereby obtaining each traversal path of the node representing the first user identifier; then, the server confirms the child nodes on each traversal path and their corresponding behavior features along each traversal path of the node representing the first user identifier, thereby obtaining the target behavior feature set {first mobile phone number, first device, first user identifier, second user identifier, second device} of the node representing the first user identifier.

[0090] In this embodiment, the server can obtain the target behavior feature set of each child node by depth-first traversal of the child nodes in the relationship topology graph, and thus use the behavior feature represented by each child node as the central feature to identify the behavior feature of the user behavior belonging to the same public account as the behavior feature from the remaining behavior features; it can be understood that the relationship topology graph is constructed based on the content of the behavior feature itself, so there are no strict requirements on the data structure of the behavior feature, and the relationship topology graph can establish connections between independent user behaviors through connecting edges. Therefore, the server can match the behavior features more widely through the relationship topology graph, thereby improving the accuracy of identifying user behaviors.

[0091] In an exemplary embodiment, the above step S106 traverses each child node in the relationship topology graph, and obtains multiple target behavior feature sets based on the connection relationship between each child node. Specifically, it also includes the following contents: traverses each child node in the relationship topology graph, and obtains multiple initial behavior feature sets based on the connection relationship between each child node; and screens each initial behavior feature set to obtain multiple target behavior feature sets.

[0092] Among them, the screening processing can be noise reduction processing, deduplication processing, or noise reduction processing and deduplication processing; noise reduction processing refers to filtering out behavioral features with large noise and interference in the behavioral features of each initial behavioral feature set; deduplication processing refers to retaining one of the repeated multiple initial feature sets.

[0093] Specifically, the server first traverses each child node in the relationship topology graph, and obtains the initial behavior feature set corresponding to each child node based on the connection relationship between each child node; then, the server filters each initial behavior feature set to obtain multiple filtered initial behavior feature sets as multiple target behavior feature sets.

[0094] It is understandable that if a screening method combining noise reduction and deduplication is adopted, the server can first perform noise reduction and then deduplication (which can reduce the amount of data processing during the deduplication process), or it can first deduplication and then perform noise reduction (which can ensure the accuracy of deduplication and avoid erroneous deduplication).

[0095] For an example, see Figure 2 Based on the connection relationship between the node representing the first device and its corresponding neighbor node, the server obtains the initial behavior feature set of the node representing the first device as {first device, first user identifier, second mobile phone number, fourth user identifier, third mobile phone number, third device, fourth mobile phone number, third user identifier, first mobile phone number, second user identifier, second device}. After screening and processing, the server obtains the target behavior feature set of the node representing the first device as {first device, first user identifier, first mobile phone number, second user identifier, second device}.

[0096] In this embodiment, the server can perform noise reduction and / or deduplication on the initial behavior feature sets by screening and processing each initial behavior feature set. On the one hand, it can reduce the interference of irrelevant behavior features on the user behavior recognition results. On the other hand, it can make the user behavior recognition results more concise.

[0097] like Figure 4 As shown, in an exemplary embodiment, each initial behavior feature set is screened to obtain multiple target behavior feature sets, specifically including the following steps:

[0098] Step S402: for each initial behavior feature set, determine the weight and number of occurrences corresponding to the connection edges connecting the various behavior features in the initial behavior feature set.

[0099] Step S404 : based on the weights and occurrence counts corresponding to the connecting edges, target behavior features are screened out from the initial behavior feature set to obtain a target behavior feature set corresponding to the initial behavior feature set.

[0100] The number of occurrences indicates the cumulative number of associations between the two behavioral features connected by the edge within a preset time period; the preset time period is the historical time period before the current user behavior collection, such as 180 days before the current user behavior collection; the cumulative number of associations refers to the total number of associations between the behavioral feature and the other behavioral features of the user behavior to which it belongs within the preset time period; see Figure 2 Taking the first mobile phone number and the second user ID as an example, there is one association between the first mobile phone number and the second user ID through user behavior 2; taking the third device and the fourth user ID as an example, there are two associations between the third device and the fourth user ID through user behavior 5 and user behavior 7 respectively.

[0101] The weight represents the priority of the corresponding behavioral feature, specifically the preset priority of the account user associated with the behavioral feature. The preset priority of the account user is used to indicate the degree of association between each account user and the public account. The higher the weight, the higher the preset priority of the account user associated with the behavioral feature, and thus the greater the likelihood that the behavioral feature represents the behavior generated by the account user when using the public account. The priority of a behavioral feature can also be the priority of the behavioral feature itself, such as the priority between different devices or the priority between different sub-accounts.

[0102] It is understandable that since the connection between each sub-node is represented by the connection between the main node and the sub-node, the edge involved in the path between the sub-node-main node-sub-node can be regarded as the connecting edge between the behavioral features corresponding to the two sub-nodes; for example, see Figure 2 The weight corresponding to the connection edge between the first device and the first mobile phone number is 0.9.

[0103] Specifically, for each initial behavior feature set, the server determines the weight and number of occurrences corresponding to the connecting edges between each behavior feature; then, based on the weight and number of occurrences of the connecting edges, the server filters out the target behavior features along the traversal path, whose connecting edge weights are greater than the weight threshold and whose connecting edge occurrences are greater than the number threshold, from among the various behavior features in the initial behavior feature set, to obtain the target behavior feature set corresponding to the initial behavior feature set.

[0104] For example, assume the weight threshold is 0.65 and the number threshold is 2; see Figure 2 Taking the node representing the first device as an example, the initial behavior feature set of the node representing the first device is {first device, first user identifier, second mobile phone number, fourth user identifier, third mobile phone number, third device, fourth mobile phone number, third user identifier, first mobile phone number, second user identifier, second device}; according to Figure 2 It can be seen that along the traversal path of the node representing the first device, the weight of the connection edge between the first user identifier and the second mobile phone number is 0.35. Therefore, the server needs to filter out the second mobile phone number from the initial behavior feature set of the node representing the first device, as well as the behavior features corresponding to each sub-node after the node representing the second mobile phone number in the traversal path of the node representing the first device, such as the third user identifier, the fourth user identifier, the fourth mobile phone number, the third mobile phone number, and the third device; at the same time, the server confirms the cumulative number of associations of each behavior feature based on the user behavior recognition results of the previous 180 days. For example, if the total number of associations of the second user identifier with the first mobile phone number and the second device in the previous 180 days is less than the number threshold of 2, then the server filters out the second user identifier from the initial behavior feature set of the node representing the first device, and finally obtains the target behavior feature set of the node representing the first device as {first device, first user identifier, second device, first mobile phone number}; if the cumulative number of associations of each behavior feature is greater than the number threshold, then the server obtains the target behavior feature set of the node representing the first device as {first device, first user identifier, second device, first mobile phone number, second user identifier}.

[0105] For another example, see Figure 2 Taking the node representing the fourth user identifier as an example, since along the traversal path of the node representing the fourth user identifier, the weight of the connection edge between the fourth user identifier and the second mobile phone number is 0.8, the weight of the connection edge between the second mobile phone number and the first user identifier is 0.35, and the weight of the connection edge between the second mobile phone number and the third user identifier is 0.2, the server needs to filter out the behavior features corresponding to each child node after the node representing the second mobile phone number in the traversal path of the node representing the fourth user identifier in the initial behavior feature set of the node representing the fourth user identifier, so as to obtain the target behavior feature set of the node representing the fourth user identifier.

[0106] It is understandable that the same public account can be associated with multiple account users, and these account users may perform operations on public devices or other people's computers, and the family members of these account users may also perform operations on other accounts through the account user's devices; therefore, in the initial behavioral feature set, there is actually a large amount of interference information; in this embodiment, the server can filter out behavioral features with interference information from multiple behavioral features in the initial behavioral feature set through the weights and number of occurrences corresponding to the connection edges, thereby avoiding the impact of behaviors such as account users performing operations on public devices or other people's computers, or account users' family members performing operations on other accounts through the account user's devices on user behavior recognition, thereby improving the accuracy of user behavior recognition.

[0107] In an exemplary embodiment, determining the weights corresponding to the various behavioral features in the initial behavioral feature set specifically includes the following: determining the account users associated with the various behavioral features in the initial behavioral feature set; and determining the weights corresponding to the connection edges between the various behavioral features in the initial behavioral feature set based on the preset priorities of the account users and the account users associated with the various behavioral features in the initial behavioral feature set.

[0108] Among them, the preset priority of the account user represents the degree of correlation between the use of each account and the public account; the preset priority of the account user can be set according to the importance of the account user to the public account, for example: U-Shield binding person>Messenger agreement binding person>Legal representative>Handling person>Financial person in charge; It can be understood that the preset priority of the account user can also be determined according to the frequency of use of the account by each account user in the past period of time, and the higher the frequency, the higher the priority.

[0109] Specifically, the server first determines the preset priority of the account users associated with each mobile phone number behavior feature in the initial behavior feature set; then, based on the preset priority order of the account users and the account users associated with each mobile phone number behavior feature, the server determines the priority of connecting each mobile phone number behavior feature in the initial behavior feature set, and then, based on the priority of each mobile phone number behavior feature, determines the weight corresponding to the connection edge between each mobile phone number behavior feature and the remaining behavior features.

[0110] For an example, see Figure 2 , assuming that the first mobile phone number is the mobile phone number bound to the U-Shield of public account 1, the second mobile phone number is the mobile phone number of the financial person in charge of public account 1, and the second mobile phone number is the mobile phone number bound to the messenger protocol of public account 2, and the third mobile phone number is the mobile phone number of the legal representative of public account 2; according to the preset priority of U-Shield binder>Messenger protocol binder>Legal representative>Handler>Financial person in charge, the server can obtain the following Figure 2The weights corresponding to the connecting edges between each mobile phone number behavior feature and the remaining behavior features are shown.

[0111] Furthermore, the server can also set priorities based on different devices used during account usage, or based on different accounts used during account usage. Taking the device priority setting as an example, the server can prioritize the internal devices of the organization or enterprise to which the public account belongs as higher priority devices, and the personal devices of the account user associated with the public account as lower priority devices. Similarly, the server can also determine the priority of devices based on the frequency of use of the public account on each device over a period of time.

[0112] In this embodiment, the server determines the priority between each behavioral feature through the priority of the account users associated with different behavioral features, and then determines the corresponding weights for the connecting edges between each behavioral feature, so that the server can quickly screen out behavioral features that are not highly correlated with public accounts from the initial behavioral feature set, eliminate the negative impact of interfering behavioral features on user behavior identification, and thereby improve the accuracy of user behavior identification.

[0113] In an exemplary embodiment, each initial behavior feature set is screened to obtain multiple target behavior feature sets, which specifically includes the following contents: from each initial behavior feature set, multiple groups of repeated behavior feature sets with exactly the same behavior features are identified; each group of repeated behavior feature sets is deduplicated to obtain multiple target behavior feature sets.

[0114] Specifically, the server compares the behavioral features in each initial behavioral feature set, identifies multiple groups of repeated behavioral feature sets that include exactly the same behavioral features from each initial behavioral feature set, and deduplicates each group of repeated behavioral feature sets (retaining any one behavioral feature set in each group of repeated behavioral feature sets) to obtain multiple deduplicated initial behavioral feature sets; then, the server identifies the multiple deduplicated initial behavioral feature sets, as well as the initial behavioral feature sets whose behavioral features are not exactly the same as those of the other initial behavioral feature sets, as target behavioral feature sets.

[0115] In this embodiment, the server can make the user behavior recognition result more concise and reduce the repetitiveness of the user behavior recognition result by deduplicating each initial behavior feature set.

[0116] In an exemplary embodiment, the above step S104 constructs a relationship topology diagram between each user behavior based on at least one behavioral feature of each user behavior, which specifically includes the following contents: for each user behavior, the user behavior is used as the main node and at least one behavioral feature of the user behavior is used as the child node of the main node to obtain a relationship point diagram between each user behavior; in the relationship point diagram, the main node and the child nodes of the main node are connected to obtain a relationship topology diagram between each user behavior.

[0117] Specifically, for each user behavior, the server takes the user behavior as the main node and the behavior characteristics of the user behavior as the sub-nodes of the main node to obtain the relationship point diagram between each user behavior; then, according to the correlation between the user behavior and the behavior characteristics, the server connects the corresponding main node and sub-node to obtain the following: Figure 2 The relationship topology diagram between various user behaviors is shown.

[0118] In this embodiment, the server constructs the main node and sub-nodes based on the correlation between user behavior and behavioral characteristics, and at the same time associates each independent user behavior with the connection relationship between the main node and the sub-node, so as to facilitate the subsequent better mining of the relationship between each user behavior and provide a basis for matching user behaviors.

[0119] In an exemplary embodiment, Figure 5 As shown, another user behavior identification method is provided, which is described by taking the application of this method to a server as an example, and includes the following steps:

[0120] Step S501: determining at least one behavior feature of each user behavior based on multiple user behaviors acquired from multiple channels.

[0121] Step S502 : For each user behavior, a relationship point diagram between each user behavior is obtained by taking the user behavior as a main node and at least one behavior feature of the user behavior as a sub-node of the main node.

[0122] Step S503: In the relationship point graph, the main node and the child nodes of the main node are connected to obtain a relationship topology graph between various user behaviors.

[0123] Step S504 , each time taking a child node in the relationship topology graph as a starting node, and determining neighbor nodes of the starting node based on the connection relationship between the starting node and other nodes.

[0124] Step S505 , starting from the starting node, performing a depth-first traversal on neighboring nodes to obtain a traversal path of the starting node.

[0125] Step S506 : obtaining an initial behavior feature set of the starting node based on the behavior features corresponding to the child nodes on the traversal path of the starting node.

[0126] Step S507 : for each initial behavior feature set, determine the weight and the number of occurrences corresponding to the connection edges connecting the various behavior features in the initial behavior feature set.

[0127] Step S508 : Based on the weights and occurrence counts corresponding to the connecting edges, target behavior features are screened out from the initial behavior feature set to obtain a target behavior feature set corresponding to the initial behavior feature set.

[0128] After step S506, the following steps may also be included:

[0129] Step S509 : identifying, from each initial behavior feature set, a plurality of repeated behavior feature sets including identical behavior features.

[0130] In step S510 , duplicate behavior feature sets of each group are removed to obtain multiple target behavior feature sets.

[0131] After step S508 and / or step S510, the method further includes:

[0132] Step S511 : confirming the user behavior corresponding to each behavior feature in the target behavior feature set as the user behavior of the same public account.

[0133] In this embodiment, first, the server constructs the main node and the child node based on the association between the user behavior and the behavior feature, and at the same time associates each independent user behavior with the connection relationship between the main node and the child node; second, the relationship topology graph is constructed based on the content of the behavior feature itself, and there is no strict requirement for the data structure of the behavior feature, and the relationship topology graph can establish the connection between each independent user behavior through the connection edge. Therefore, the server can match the behavior feature more widely through the relationship topology graph; third, the server can obtain the target behavior feature set of each child node by depth-first traversal of the child nodes in the relationship topology graph, and thus use the behavior feature represented by each child node as the central feature to identify the behavior feature of the user behavior belonging to the same public account as the behavior feature from the remaining behavior features; fourth, the server can reduce the noise and / or deduplicate the initial behavior feature set by screening each initial behavior feature set. On the one hand, it can reduce the interference of irrelevant behavior features on the user behavior recognition results, and on the other hand, it can make the user behavior recognition results more concise. Compared with the user behavior identification method based on relational database that relies on data structure, the user behavior identification method based on the above process can accurately identify each user behavior and its behavior characteristics with different data structures through the relational topology graph, thereby improving the recognition accuracy of user behavior.

[0134] In order to more clearly illustrate the user behavior identification method provided by the embodiment of the present application, the user behavior identification method is specifically described below with a specific embodiment, but it should be understood that the embodiment of the present application is not limited thereto. Figure 6 As shown, in an exemplary embodiment, the present application also provides a method for identifying public accounts based on point-edge search, which specifically includes the following steps:

[0135] Step 1: Acquisition of user behavior and its behavioral characteristics.

[0136] The server collects the embedded information from various channels to confirm the user behavior of different users, as well as the characteristic information that represents the identities of different users, as the behavioral characteristics of user behavior.

[0137] Step 2: Construction of relationship topology diagram.

[0138] The server abstracts user behavior and its behavioral characteristics into nodes, and forms connection edges between each node based on the correlation between user behavior and behavioral characteristics. The server also determines the weight and threshold of each connection edge based on the priority of different users to obtain a relationship topology graph.

[0139] Step 3: Traverse the relationship topology graph.

[0140] The server adopts a depth-first traversal scheme to scan each node in the relationship topology graph and obtain the first behavior feature set of each node.

[0141] Step 4: Integration of result sets.

[0142] For each first behavior feature, the server first filters out the target behavior feature from each first behavior feature set based on the weight and threshold of the connecting edge to obtain the second behavior feature set of each node; then, the server deduplicates each second behavior feature set to obtain multiple third behavior feature sets.

[0143] Step 5: Identification of public accounts.

[0144] The user behaviors constituted by the behavioral features in each third behavioral feature set are confirmed as user behaviors of the same public account, and account numbers are assigned to the user behaviors of each public account, and a mapping relationship is established between the account numbers and the channels for collecting user behaviors.

[0145] It is understandable that due to the large differences in the buried information collected by different channels, if the traditional corporate account identification method based on a structured relational database is used to match user behavior, it will lead to problems such as poor efficiency and long time consumption; but in this embodiment, the server introduces the calculation idea of ​​graph computing, and solves the problem of low efficiency caused by the inconsistency of the buried elements collected by each channel through point-edge search, thereby realizing the unified authentication of multi-roles, multi-channels, and before and after login of corporate accounts, and improving the accuracy of corporate account identification; in addition, through the setting of edge weights and thresholds, abnormal data and occasional data are filtered, thereby improving the credibility of corporate account identification. The corporate account identification method based on the above process breaks through the information islands of each channel and provides technical support for the full-journey identification of users in corporate digital operations.

[0146] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0147] Based on the same inventive concept, the embodiments of the present application also provide a user behavior recognition device for implementing the user behavior recognition method involved above. The implementation solution provided by this device is similar to the implementation solution described in the above method. Therefore, the specific limitations of one or more user behavior recognition device embodiments provided below can be referred to the limitations of the user behavior recognition method above and will not be repeated here.

[0148] In an exemplary embodiment, Figure 7 As shown, a user behavior recognition device is provided, including: a behavior feature determination module 702, a topology relationship construction module 704, a topology relationship traversal module 706 and a user behavior recognition module 708, wherein:

[0149] The behavior feature determination module 702 is configured to determine at least one behavior feature of each user behavior based on multiple user behaviors acquired from multiple channels.

[0150] The topological relationship construction module 704 is used to construct a relationship topology diagram between each user behavior based on at least one behavioral feature of each user behavior; the relationship topology diagram includes multiple main nodes, each main node corresponds to a user behavior, and each main node is associated with at least one child node, and each child node corresponds to a behavioral feature of the associated user behavior.

[0151] The topology relationship traversal module 706 is used to traverse each sub-node in the relationship topology graph and obtain multiple target behavior feature sets based on the connection relationship between each sub-node.

[0152] The user behavior identification module 708 is used to identify the user behavior corresponding to the behavior feature in each target behavior feature set as the user behavior of the same public account.

[0153] In an exemplary embodiment, the topological relationship traversal module 706 is also used to take a child node in the relationship topology graph as the starting node each time, and determine the neighbor nodes of the starting node based on the connection relationship between the starting node and other nodes; starting from the starting node, perform a depth-first traversal on the neighbor nodes to obtain the traversal path of the starting node; based on the behavioral characteristics corresponding to each child node on the traversal path of the starting node, obtain the target behavioral feature set of the starting node.

[0154] In an exemplary embodiment, the topology relationship traversal module 706 is also used to traverse each child node in the relationship topology graph, and obtain multiple initial behavior feature sets based on the connection relationship between each child node; and filter each initial behavior feature set to obtain multiple target behavior feature sets.

[0155] In an exemplary embodiment, the topological relationship traversal module 706 is also used to determine, for each initial behavior feature set, the weight and number of occurrences corresponding to the connecting edge between the various behavior features in the initial behavior feature set; the number of occurrences represents the cumulative number of associations between the two behavior features connected by the connecting edge within a preset time period; based on the weight and number of occurrences corresponding to the connecting edge, the target behavior feature is screened out from the initial behavior feature set to obtain the target behavior feature set corresponding to the initial behavior feature set.

[0156] In an exemplary embodiment, the topological relationship traversal module 706 is also used to determine the account users associated with each behavioral feature in the initial behavioral feature set; based on the preset priority of the account users and the account users associated with each behavioral feature in the initial behavioral feature set, determine the weights corresponding to the connecting edges between each behavioral feature in the initial behavioral feature set.

[0157] In an exemplary embodiment, the topological relationship traversal module 706 is further used to identify multiple groups of repeated behavior feature sets including exactly the same behavior features from each initial behavior feature set; and perform deduplication processing on each group of repeated behavior feature sets to obtain multiple target behavior feature sets.

[0158] In an exemplary embodiment, the topological relationship construction module 704 is used to obtain a relationship point graph between each user behavior, with the user behavior as the main node and at least one behavioral feature of the user behavior as the child node of the main node; in the relationship point graph, the main node and the child nodes of the main node are connected to obtain a relationship topology graph between each user behavior.

[0159] Each module in the user behavior recognition device described above can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0160] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 8As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O) and a communication interface. The processor, memory and input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store data such as user behavior and its behavior characteristics obtained from multiple channels, user behavior identification results for public accounts, etc. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a user behavior identification method is implemented.

[0161] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0162] In an exemplary embodiment, a computer device is further provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0163] In an exemplary embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0164] In an exemplary embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0165] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processor involved in the various embodiments provided herein may be, but are not limited to, a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, and the like.

[0166] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0167] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A user behavior identification method, characterized in that: include: Determining at least one behavior feature of each user behavior based on multiple user behaviors obtained from multiple channels; constructing a relationship topology graph between the user behaviors according to at least one behavior feature of the user behaviors; The relationship topology graph includes multiple main nodes, each main node corresponds to a user behavior, each main node is associated with at least one child node, and each child node corresponds to a behavior feature of the associated user behavior; Traversing each sub-node in the relationship topology graph, and obtaining a plurality of target behavior feature sets based on the connection relationships between the sub-nodes; The user behavior corresponding to the behavioral features in each target behavioral feature set is confirmed as the user behavior of the same public account, wherein the user behavior corresponding to the behavioral features in the target behavioral feature set refers to the user behavior composed of the various behavioral features in the target behavioral feature set.

2. The method according to claim 1, characterized in that The traversing of each sub-node in the relationship topology graph and obtaining multiple target behavior feature sets based on the connection relationships between the sub-nodes include: Each time, a child node in the relationship topology graph is used as a starting node, and based on the connection relationship between the starting node and other nodes, a neighbor node of the starting node is determined; Starting from the starting node, performing a depth-first traversal on the neighboring nodes to obtain a traversal path of the starting node; Based on the behavior features corresponding to the child nodes on the traversal path of the starting node, a target behavior feature set of the starting node is obtained.

3. The method according to claim 1, characterized in that The traversing each sub-node in the relationship topology graph and obtaining a plurality of target behavior feature sets based on the connection relationships between the sub-nodes further includes: Traversing each sub-node in the relationship topology graph, and obtaining a plurality of initial behavior feature sets based on the connection relationships between the sub-nodes; Screening is performed on each initial behavior feature set to obtain the multiple target behavior feature sets.

4. The method according to claim 3, characterized in that The screening process of each initial behavior feature set to obtain the multiple target behavior feature sets includes: For each initial behavioral feature set, determine the weight and number of occurrences corresponding to the connection edge connecting each behavioral feature in the initial behavioral feature set; the number of occurrences represents the cumulative number of associations between the two behavioral features connected by the connection edge within a preset time period; Based on the weights and occurrence counts corresponding to the connecting edges, target behavior features are screened out from the initial behavior feature set to obtain a target behavior feature set corresponding to the initial behavior feature set.

5. The method according to claim 4, characterized in that The determining of the weights corresponding to the connecting edges between the behavioral features in the initial behavioral feature set includes: Determining an account user associated with each behavioral feature in the initial behavioral feature set; Based on the preset priorities of the account users and the account users associated with the respective behavioral features in the initial behavioral feature set, weights corresponding to the connecting edges between the respective behavioral features in the initial behavioral feature set are determined.

6. The method according to claim 3, characterized in that The screening of the initial behavior feature sets to obtain the plurality of target behavior feature sets further includes: identifying, from the respective initial behavioral feature sets, a plurality of repeated behavioral feature sets including identical behavioral features; Deduplication processing is performed on each group of repeated behavior feature sets to obtain the multiple target behavior feature sets.

7. The method according to any one of claims 1 to 6, characterized in that The constructing a relationship topology diagram between the user behaviors according to at least one behavior feature of the user behaviors includes: For each user behavior, taking the user behavior as a main node and at least one behavior feature of the user behavior as a child node of the main node, a relationship point graph between the user behaviors is obtained; In the relationship point graph, the main node and the child nodes of the main node are connected to obtain a relationship topology graph between the various user behaviors.

8. A user behavior recognition device, characterized in that: The device comprises: A behavior feature determination module, configured to determine at least one behavior feature of each user behavior based on multiple user behaviors acquired from multiple channels; A topology relationship building module is configured to build a relationship topology graph between the user behaviors based on at least one behavior feature of the user behaviors; the relationship topology graph includes a plurality of main nodes, each main node corresponds to a user behavior, and each main node is associated with at least one child node, each child node corresponds to a behavior feature of the associated user behavior; A topology relationship traversal module, configured to traverse each sub-node in the relationship topology graph and obtain a plurality of target behavior feature sets based on the connection relationships between the sub-nodes; The user behavior identification module is used to confirm the user behavior corresponding to the behavior feature in each target behavior feature set as the user behavior of the same public account, wherein the user behavior corresponding to the behavior feature in the target behavior feature set refers to the user behavior composed of each behavior feature in the target behavior feature set.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

11. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images