Method and apparatus for acquiring character relationship based on graph learning, electronic device and medium

Abstract

The present disclosure provides a person relationship acquisition method and device based on graph learning, electronic equipment and medium, the method comprises: acquiring multiple person relationship subgraphs, each person relationship subgraph comprising a target node formed by a unique account identifier of a user; for a first person relationship subgraph in the multiple person relationship subgraphs, in response to determining that at least one node in the first person relationship subgraph and at least one node in a second person relationship subgraph have an association relationship, acquiring a target person relationship graph composed of the first person relationship subgraph and the second person relationship subgraph; acquiring the similarity between the target node of the first person relationship subgraph and the target node of the second person relationship subgraph in the target person relationship graph; determining the person relationship between the user corresponding to the first person relationship subgraph and the user corresponding to the second person relationship subgraph based on the similarity. The present scheme realizes the acquisition of person relationship in multiple business scenarios, and improves the efficiency of person relationship mining in social scenarios.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and in particular to a method, apparatus, electronic device, and medium for obtaining person-to-person relationships based on graph learning. Background Technology

[0002] In social scenarios, a user may be registered in multiple business scenarios, with each scenario assigning the user a different identity document (ID). However, current technologies for obtaining user relationships in social scenarios are inefficient. Summary of the Invention

[0003] To solve the above-mentioned technical problems, or at least partially solve them, this disclosure provides a method, apparatus, electronic device, and medium for obtaining person-relationships based on graph learning.

[0004] According to one aspect of this disclosure, a graph learning-based method for obtaining interpersonal relationships is provided, including:

[0005] Obtain multiple character relationship subgraphs, where each character relationship subgraph includes target nodes consisting of unique account identifiers for users;

[0006] For a first character relationship subgraph in the plurality of character relationship subgraphs, in response to determining that at least one node in the first character relationship subgraph is associated with at least one node in the second character relationship subgraph, a target character relationship graph composed of the first character relationship subgraph and the second character relationship subgraph is obtained, wherein the second character relationship subgraph is a character relationship subgraph in the plurality of character relationship subgraphs other than the first character relationship subgraph;

[0007] Obtain the similarity between the target nodes of the first character relationship subgraph and the target nodes of the second character relationship subgraph in the target character relationship graph;

[0008] Based on the similarity, the relationship between the user corresponding to the first person relationship subgraph and the user corresponding to the second person relationship subgraph is determined.

[0009] According to another aspect of this disclosure, a graph learning-based device for obtaining interpersonal relationships is provided, comprising:

[0010] The first acquisition module is used to acquire multiple character relationship subgraphs, wherein each character relationship subgraph includes a target node consisting of a user's unique account identifier;

[0011] The second acquisition module is used to acquire a target character relationship diagram composed of the first character relationship diagram and the second character relationship diagram in response to determining that at least one node in the first character relationship diagram has an association relationship with at least one node in the second character relationship diagram, for the first character relationship diagram in the plurality of character relationship diagrams, wherein the second character relationship diagram is a character relationship diagram in the plurality of character relationship diagrams other than the first character relationship diagram;

[0012] The third acquisition module is used to acquire the similarity between the target nodes of the first character relationship subgraph and the target nodes of the second character relationship subgraph in the target character relationship graph.

[0013] The determination module is used to determine the relationship between the user corresponding to the first person relationship subgraph and the user corresponding to the second person relationship subgraph based on the similarity.

[0014] According to another aspect of this disclosure, an electronic device is provided, comprising:

[0015] Processor; and

[0016] Stored program memory,

[0017] The program includes instructions that, when executed by the processor, cause the processor to perform the graph learning-based person relationship acquisition method according to the foregoing aspect.

[0018] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are used to cause the computer to perform the graph learning-based person relationship acquisition method according to the foregoing aspect.

[0019] According to another aspect of this disclosure, a computer program product is provided, comprising a computer program, wherein the computer program, when executed by a processor, implements the presentation conversion based on graph learning for obtaining interpersonal relationships as described in the foregoing aspect.

[0020] One or more technical solutions provided in this disclosure obtain multiple person relationship subgraphs, each including target nodes composed of unique user account identifiers. For a first person relationship subgraph among the multiple person relationship subgraphs, in response to determining that at least one node in the first person relationship subgraph is associated with at least one node in a second person relationship subgraph, a target person relationship graph composed of the first and second person relationship subgraphs is obtained. The second person relationship subgraph is the person relationship subgraph other than the first person relationship subgraph among the multiple person relationship subgraphs. Then, the similarity between the target nodes in the first and second person relationship subgraphs in the target person relationship graph is obtained. Based on the similarity, the person relationship between the user corresponding to the first person relationship subgraph and the user corresponding to the second person relationship subgraph is determined. Using the solution of this disclosure, when a relationship is determined between nodes in two person relationship subgraphs, the person relationship between the users corresponding to these two subgraphs can be determined based on the similarity between the target nodes in these two subgraphs. This achieves person relationship acquisition in multiple business scenarios, is applicable to scenarios with multiple business lines, and improves the efficiency of person relationship mining in social scenarios. Attached Figure Description

[0021] Further details, features, and advantages of this disclosure are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which:

[0022] Figure 1 A flowchart illustrating a graph learning-based method for obtaining interpersonal relationships according to an exemplary embodiment of the present disclosure is shown.

[0023] Figure 2 A schematic diagram of a character relationship subgraph in an exemplary embodiment of this disclosure is shown;

[0024] Figure 3 A schematic diagram of the target person relationship diagram in an exemplary embodiment of this disclosure is shown;

[0025] Figure 4 A schematic block diagram of a graph learning-based person relationship acquisition apparatus according to an exemplary embodiment of the present disclosure is shown;

[0026] Figure 5 A structural block diagram of an exemplary electronic device that can be used to implement embodiments of the present disclosure is shown. Detailed Implementation

[0027] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0028] It should be understood that the steps described in the method embodiments of this disclosure may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of this disclosure is not limited in this respect.

[0029] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the description below. It should be noted that the concepts of "first", "second", etc., used in this disclosure are only used to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0030] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0031] The names of messages or information exchanged between multiple devices in the embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

[0032] The following description, with reference to the accompanying drawings, outlines the graph learning-based method, apparatus, electronic device, and storage medium for obtaining interpersonal relationships provided in this disclosure.

[0033] Currently, there are two main methods for ID relationship mining. One is to collect user ID relationship data in a fixed format through a relatively strict method, and then apply the collected ID relationship data to other scenarios after simple processing. This method requires users to input relatively detailed identity information, which is not suitable for ordinary scenarios such as education where there is no right to collect users' family attribute information. The second method is to use machine learning algorithms to mine interpersonal relationships. This method uses users' social data to mine potential connections between users, which is more suitable for scenarios with rich social relationships. It is not suitable for scenarios with insufficient social relationships (such as education scenarios), and it is difficult to mine accurate interpersonal relationships.

[0034] To address the aforementioned issues, this disclosure provides a graph learning-based method for obtaining user relationships. By acquiring multiple user relationship subgraphs, each subgraph includes target nodes composed of unique user account identifiers. For a first user relationship subgraph, in response to determining that at least one node in the first subgraph is associated with at least one node in a second subgraph, a target user relationship graph is obtained, composed of the first and second subgraphs. The second subgraph is the only user relationship subgraph among the multiple subgraphs excluding the first. Then, the similarity between the target nodes in the first and second subgraphs is obtained. Based on this similarity, the user relationship between the users corresponding to the first and second subgraphs is determined. Using this method, when a relationship is determined between nodes in two subgraphs, the user relationship between the corresponding users in these two subgraphs can be determined based on the similarity between the target nodes in these two subgraphs. This enables user relationship acquisition across multiple business scenarios, is applicable to scenarios with multiple business lines, and improves the efficiency of user relationship mining in social scenarios. This solution uses graph computing to associate user identifiers across multiple business scenarios, making it suitable for complex scenarios and improving the efficiency and accuracy of user relationship association.

[0035] Figure 1 A flowchart of a graph learning-based person relationship acquisition method according to an exemplary embodiment of the present disclosure is shown. The method can be executed by a graph learning-based person relationship acquisition device, wherein the device can be implemented in software and / or hardware, and is generally integrated into an electronic device, including devices such as mobile phones, tablets, and servers.

[0036] like Figure 1 As shown, this graph learning-based method for obtaining interpersonal relationships may include the following steps:

[0037] Step 101: Obtain multiple character relationship subgraphs, wherein each character relationship subgraph includes a target node consisting of a user's unique account identifier.

[0038] In this embodiment, a person-relationship subgraph can be generated based on user registration data in different business scenarios. When mining person-relationships, the registration data obtained is the registration data of all users in each business scenario. Therefore, multiple person-relationship subgraphs are generated, including subgraphs for different users. Furthermore, since users may use different login devices, mobile phone numbers, email addresses, etc., in different business scenarios, the same user may also correspond to different subgraphs. For each generated person-relationship subgraph, the subgraph includes nodes corresponding to the user's account identifiers in different business scenarios. These nodes include a target node composed of the user's unique account identifier. The user mentioned here may be a single user or multiple users belonging to the same household; this disclosure does not impose any limitations on this.

[0039] Different business scenarios can refer to different product line scenarios, such as home appliance control applications, gaming scenarios, learning scenarios, and shopping scenarios, or different business scenarios for the same product. Account identifiers can be identity identifiers assigned to users after registration. These identifiers can be based on third-party accounts, mobile phone numbers, email addresses, device identifiers, etc., used during registration. Unique account identifiers are determined from account identifiers based on preset rules. It's important to note that unique account identifiers are unique identifiers for natural persons; each user's unique account identifier is unique. For example, a user may have multiple accounts, including account identifiers IDa, IDb, etc., but that user has only one unique account identifier.

[0040] In this embodiment of the disclosure, a person relationship subgraph can be obtained through a routine task with a preset time interval (such as daily).

[0041] For example, when the time node for task execution arrives, a routine task is executed to obtain user registration data from multiple business scenarios up to the current time node, and to determine multiple subgraphs of relationships between people at the current time node based on the user registration data.

[0042] User registration data includes, but is not limited to, usernames, registered mobile phone numbers, linked email addresses, third-party platform accounts, etc. User registration data can be obtained with the user's authorization.

[0043] In one optional embodiment of this disclosure, association rules for account identifiers can be pre-set, and user registration data for each business scenario can be analyzed and processed according to the association rules to identify account identifiers with related relationships and generate a sub-graph of user relationships.

[0044] In one optional embodiment of this disclosure, a large number of training samples can be collected. These training samples include the relationships between different account identifiers. A maximum connected subgraph model is trained using these training samples to predict the relationships between account identifiers, thereby outputting a person-relationship subgraph. In this embodiment, user registration data from various business scenarios can be obtained. The obtained user registration data is input into the trained maximum connected subgraph model, which then outputs multiple person-relationship subgraphs.

[0045] Considering that user registration data may contain letters, underscores, or other characters, which are not supported by graph computation, when obtaining the person relationship subgraph, for the user registration data, we can first perform auto-increment mapping on the account data involved in the registration data, mapping the original account to a pure numeric account identifier, and then mine the person relationship subgraph.

[0046] For example, an auto-incrementing mapping can be performed starting from 1, mapping the original account data to positive integers. For instance, user_id_1 is mapped to the number 1, user_id_2 is mapped to the number 2, phone_number1 is mapped to the number 3, and so on, completing a purely digitized auto-incrementing mapping.

[0047] Step 102: For the first character relationship subgraph in the plurality of character relationship subgraphs, in response to determining that at least one node in the first character relationship subgraph is associated with at least one node in the second character relationship subgraph, obtain a target character relationship graph composed of the first character relationship subgraph and the second character relationship subgraph, wherein the second character relationship subgraph is the character relationship subgraph in the plurality of character relationship subgraphs other than the first character relationship subgraph.

[0048] The first character relationship subgraph is any one of the multiple character relationship subgraphs obtained, and the second character relationship subgraph is any character relationship subgraph other than the first character relationship subgraph. The second character relationship subgraph can be one or more.

[0049] As time goes by, users' usage data in various business scenarios becomes richer and richer. The new user data may create relationships between two or more previously isolated subgraphs of user relationships. These subgraphs can then be linked together through associated nodes to form a new subgraph of user relationships.

[0050] For example, Figure 2 This illustration shows a schematic diagram of a character relationship subgraph in an exemplary embodiment of the present disclosure. Figure 2In the diagram, the left side of the arrow represents two subgraphs representing user relationships generated based on user registration data. On day T, the user logged into a third-party application on the device corresponding to id3 (device id) using id2 (social account 1) (e.g., login account 1, id1). Based on this data, it can be determined that there are edges between id1 and id2, and between id3, and also between id2 and id3. Figure 2 The first person relationship subgraph in the graph; the user also registered login account 2 (id7) on day T, and at the time of registration, he / she also bound social account 2 (id5), mobile phone number (id6) and email address (id4). The graph is generated based on this data. Figure 2 The second subgraph of the relationship between people is shown below. Assuming the first subgraph uses `id1` as the unique account identifier for the user corresponding to that subgraph, and the second subgraph uses `id7` as the unique account identifier for the user corresponding to that subgraph, then the node corresponding to `id1` is the target node in the first subgraph, and the node corresponding to `id7` is the target node in the second subgraph. On day T+1, the user logs into social account 2 on the device corresponding to `id3` (device ID). Therefore, an edge relationship is formed between `id3` in the first subgraph and `id5` in the second subgraph. This confirms that the node corresponding to `id3` in the first subgraph and the node corresponding to `id5` in the second subgraph are associated, and the two subgraphs are connected, generating the following... Figure 2 The relationship diagram of the target characters shown to the right of the middle arrow.

[0051] Step 103: Obtain the similarity between the target nodes of the first person relationship subgraph and the target nodes of the second person relationship subgraph in the target person relationship graph.

[0052] In this embodiment of the disclosure, after obtaining the target person relationship graph, the target nodes in the original first person relationship subgraph and the original second person relationship subgraph can be determined from the target person relationship graph, and the similarity between the target nodes in the first person relationship subgraph and the target nodes in the second person relationship subgraph can be calculated. Specifically, when there are multiple second person relationship subgraphs, the similarity between the target nodes in the first person relationship subgraph and the target nodes in each of the second person relationship subgraphs is calculated.

[0053] For example, for multiple generated subgraphs of interpersonal relationships, the unique account identifier corresponding to the target node of each subgraph can be recorded. After obtaining the target interpersonal relationship graph, based on the recorded unique account identifiers corresponding to the target nodes of each subgraph, the unique account identifiers corresponding to the target nodes of all subgraphs constituting the target interpersonal relationship graph are found, thus finding all target nodes in the target interpersonal relationship graph. Then, with user authorization, feature data corresponding to each unique account identifier can be extracted from the business data accumulated in the business scenario to which each unique account identifier belongs. This yields the feature data of the user corresponding to the first interpersonal relationship subgraph and the feature data of the user corresponding to the second interpersonal relationship subgraph. The feature data may include, but is not limited to, geographic location feature data, real name feature data, and contact information feature data. Geographic location feature data may include, for example, province, city, and neighborhood; contact information feature data may include, for example, phone number, email address, and third-party account. Next, based on a preset similarity algorithm, similarity calculation can be performed using the feature data of the user corresponding to the first interpersonal relationship subgraph and the feature data of the user corresponding to the second interpersonal relationship subgraph to obtain the similarity between the target nodes of the first interpersonal relationship subgraph and the target nodes of the second interpersonal relationship subgraph.

[0054] The preset similarity algorithm can be, but is not limited to, any one of the following: cosine similarity algorithm, Jaccard similarity coefficient algorithm, and Pearson correlation coefficient algorithm.

[0055] Step 104: Based on the similarity, determine the relationship between the user corresponding to the first person relationship subgraph and the user corresponding to the second person relationship subgraph.

[0056] In this embodiment of the disclosure, after obtaining the similarity between the target nodes of the first person relationship subgraph and the target nodes of the second person relationship subgraph, the person relationship between the user corresponding to the first person relationship subgraph and the user corresponding to the second person relationship subgraph can be determined based on the calculated similarity.

[0057] For example, if the similarity between the target node in the first person relationship subgraph and the target node in the second person relationship subgraph is greater than a preset similarity threshold, it can be determined that there is a relationship between the users corresponding to these two subgraphs. These two users may be the same user or users from the same family. If the similarity is not greater than the preset similarity threshold, it can be determined that there is no relationship between the two users.

[0058] The similarity threshold can be set according to actual needs, such as setting the similarity threshold to 0.3, 0.5, 0.6, etc.

[0059] The graph learning-based method for obtaining user relationships in this disclosure involves acquiring multiple user relationship subgraphs, each including target nodes composed of unique user account identifiers. For a first user relationship subgraph, in response to determining that at least one node in the first subgraph is associated with at least one node in a second subgraph, a target user relationship graph is obtained, composed of the first and second subgraphs. The second subgraph is the only user relationship subgraph in the multiple subgraphs excluding the first. Then, the similarity between the target nodes in the first and second subgraphs is obtained. Based on this similarity, the user relationship between the user corresponding to the first and second subgraphs is determined. This method allows for the determination of user relationships between users in two subgraphs when a relationship is found between nodes in those subgraphs, based on the similarity between the target nodes. This enables user relationship acquisition across multiple business scenarios, is applicable to scenarios with multiple business lines, and improves the efficiency of user relationship mining in social scenarios.

[0060] Typically, when a user performs a new login operation, it may lead to new associations between multiple person-relationship subgraphs. Therefore, in one optional embodiment of this disclosure, upon receiving a user's login operation, multiple account identifiers corresponding to the login operation can be obtained in response to the login operation. Then, it is determined whether the multiple account identifiers exist in at least two person-relationship subgraphs respectively. In response to the existence of a first node composed of the first account identifier among the multiple account identifiers in the first person-relationship subgraph, and a second node composed of the second account identifier among the multiple account identifiers in the second person-relationship subgraph, it is determined that there is an association between the first node and the second node.

[0061] Among them, multiple accounts are identified as having at least two identifiers.

[0062] Still with Figure 2 For example, when a user logs into social account 2 through the device corresponding to id3, the two account identifiers corresponding to this login operation can be obtained as device id (id3) and social account 2 (id5), respectively. Figure 2 As shown in the two subgraphs to the left of the middle arrow, the first subgraph contains nodes with ID3, and the second subgraph contains nodes with ID5. Therefore, it can be determined that there is a relationship between the nodes with ID3 and the nodes with ID5. Figure 2 The dotted line in the target character relationship diagram to the right of the middle arrow is shown.

[0063] In one optional embodiment of this disclosure, when obtaining the similarity between the target node of the first subgraph of the target person relationship graph and the target node of the second subgraph of the target person relationship graph, the first feature data of the user corresponding to the first subgraph of the target person relationship graph and the second feature data of the user corresponding to the second subgraph of the target person relationship graph can be obtained first. Based on the locality-sensitive hashing algorithm, features of the same dimension in the first feature data and the second feature data are mapped into the same group to obtain multiple groups. Then, for each group, based on the Jaccard similarity coefficient algorithm, the Jaccard similarity coefficient between the target node of the first subgraph of the target person relationship graph and the target node of the second subgraph of the target person relationship graph on the corresponding dimension of the group is calculated. Finally, based on the average value of the Jaccard similarity coefficients between the target node of the first subgraph of the target person relationship graph and the target node of the second subgraph of the target person relationship graph on the corresponding dimensions of the multiple groups, the similarity between the target node of the first subgraph of the target person relationship graph and the target node of the second subgraph of the target person relationship graph is determined.

[0064] The feature data may include, but is not limited to, geographic location feature data, real name feature data, and contact information feature data. Geographic location feature data may include, for example, the province, city, and neighborhood. Contact information feature data may include, for example, a phone number, email address, or third-party account.

[0065] In this embodiment of the disclosure, with user authorization, the first feature data of the user corresponding to the first person-relationship subgraph can be extracted from the business data accumulated under the business scenarios to which each account identifier belongs in the first person-relationship subgraph, and the second feature data of the user corresponding to the second person-relationship subgraph can be extracted from the business data accumulated under the business scenarios to which each account identifier belongs in the second person-relationship subgraph. Then, based on the locality-sensitive hashing algorithm, the features in the first and second feature data are mapped according to the same dimension, mapping features belonging to the same dimension into the same group, thus obtaining multiple groups. The number of groups matches the number of dimensions of the features contained in the feature data.

[0066] For example, a locality-sensitive hashing algorithm based on the Jaccard similarity coefficient can be used to calculate the similarity between the acquired first feature data and the second feature data, wherein the larger the Jaccard similarity coefficient value, the higher the similarity between the users corresponding to the two identifiers.

[0067] Locality Sensitive Hashing (LSH) is used to solve the problem of finding similar nodes in high-dimensional space. When the acquired feature data has many dimensions, performing a linear search directly in the high-dimensional space will lead to the curse of dimensionality and low efficiency. The role of LSH is to map the points in the original high-dimensional space to different positions in one or more hash tables. These positions are called buckets. The mapping principle is that points that are very close in the original high-dimensional space will be mapped to the same bucket with a high probability. Points in the same bucket are also very likely to be similar in the original high-dimensional space. In this way, you can directly search for elements in this bucket, which greatly improves the search efficiency.

[0068] In this embodiment of the disclosure, the acquired first feature data and second feature data can be hashed according to dimensions, and data of the same dimension can be mapped into the same bucket (i.e., group). One bucket is one group. In each bucket, the Jaccard similarity coefficient algorithm is used to calculate the Jaccard similarity coefficient between the target node of the first person relationship subgraph and the target node of the second person relationship subgraph in the corresponding dimension of the bucket. Then, the average Jaccard similarity coefficient between the target node of the first person relationship subgraph and the target node of the second person relationship subgraph is calculated in all dimensions. The average Jaccard similarity coefficient is determined as the similarity between the target node of the first person relationship subgraph and the target node of the second person relationship subgraph in the target person relationship graph. The formula for calculating the Jaccard similarity coefficient is shown in formula (1).

[0069]

[0070] Here, A and B represent the sets of feature data of users corresponding to the same dimension in two subgraphs within the same bucket.

[0071] In this embodiment of the disclosure, by calculating the similarity between the target nodes of the first subgraph of the target person relationship graph and the target nodes of the second subgraph of the target person relationship graph, data support is provided for subsequently determining the relationship between users corresponding to the two subgraphs.

[0072] In one optional embodiment of this disclosure, the calculated similarity can be used to determine the relationship between users corresponding to the first person-relationship subgraph and users corresponding to the second person-relationship subgraph. Specifically, the calculated similarity can be compared with a preset similarity threshold. If the similarity is greater than the threshold, a relationship is determined to exist between the users corresponding to the first person-relationship subgraph and the users corresponding to the second person-relationship subgraph; otherwise, no relationship is determined to exist between them.

[0073] The similarity threshold can be preset according to actual needs, such as setting the similarity threshold to 0.3, 0.5, 0.6, etc.

[0074] Furthermore, if it is determined that there is a relationship between the user corresponding to the first person relationship subgraph and the user corresponding to the second person relationship subgraph, a set of unique account identifiers can be generated based on the unique account identifiers corresponding to the target nodes of the first person relationship subgraph and the target nodes of the second person relationship subgraph.

[0075] In other words, the unique account identifier set consists of the unique account identifiers of users with whom there is a relationship, and the unique account identifier set includes at least two unique account identifiers.

[0076] In this embodiment of the disclosure, by aggregating the unique account identifiers of users with interpersonal relationships to generate a unique account identifier set, the associated users are linked together, thereby realizing the mining of interpersonal relationships in multiple business scenarios.

[0077] To facilitate subsequent queries and retrieval, in this embodiment of the disclosure, after determining the set of unique account identifiers, a unique account identifier can be selected from the set as the set identifier corresponding to that set. The set identifier can be viewed as a family ID, and the set of unique account identifiers can be viewed as a family. Users in the set of unique account identifiers are either the same user or users belonging to the same family. This disclosure provides different schemes for determining the set identifier, which will be explained below.

[0078] In one optional embodiment of this disclosure, the business identifier of the business to which each unique account identifier belongs in the unique account identifier set can be obtained. Then, the priority order of different business identifiers is queried, and the target business identifier with the highest priority is determined from the business identifiers. Then, the unique account identifier of the business corresponding to the target business identifier in the unique account identifier set is determined as the set identifier corresponding to the unique account identifier set.

[0079] The priority order of different business identifiers can be preset according to actual needs. For example, in scenarios involving different businesses, the priority of the corresponding business identifier can be set according to the size of the business department. The size of the department can be reflected by the number of registered users. The more registered users, the larger the department, and the higher the priority of the corresponding business identifier. The business identifier can be the department identifier or the department name, etc.

[0080] In this embodiment of the disclosure, the business identifier of the business to which each unique account identifier belongs in the unique account identifier set can be obtained from the business data of the associated business scenario. For each unique account identifier obtained, by querying the priority order of different business identifiers, the priority order of the business identifier of the business to which each unique account identifier belongs in the unique account identifier set can be determined, and then the target business identifier with the highest priority can be determined. The business identifier of the business to which the target business identifier belongs is the unique account identifier of the target business identifier and is determined as the set identifier corresponding to the unique account identifier set.

[0081] For example, suppose the priorities of different business identifiers are arranged from high to low as business c > business a > business b > business d. In the set of unique account identifiers, the business identifier of the business to which unique account identifier ID1 belongs is business a, the business identifier of the business to which unique account identifier ID2 belongs is business c, and the business identifier of the business to which unique account identifier ID3 belongs is business d. By querying the priority order of the different business identifiers, it can be determined that the business to which ID2 belongs has the highest priority in the set of unique account identifiers. That is, business c is the target business identifier, and ID2 is determined as the set identifier corresponding to the set of unique account identifiers.

[0082] In one optional implementation of this disclosure, the generation time of each unique account identifier in the unique account identifier set can be obtained, and the unique account identifier with the earliest generation time can be determined as the set identifier corresponding to the unique account identifier set.

[0083] For example, the generation time of each unique account identifier can be obtained from the business data of the business scenarios associated with each unique account identifier in the unique account identifier set, and then the unique account identifier with the earliest generation time can be determined as the set identifier corresponding to the unique account identifier set.

[0084] Typically, a natural person's unique account identifier is unique. After a user registers in one business scenario, the unique account identifier will not change when registering in other scenarios. Therefore, in this embodiment of the disclosure, the unique account identifier with the earliest generation time can be determined as the set identifier based on the generation time of the unique account identifier, which has high accuracy.

[0085] Furthermore, in one optional embodiment of this disclosure, after determining the set identifier, the correspondence between the set identifier and other unique account identifiers in the unique account identifier set can be stored to facilitate subsequent querying and retrieval.

[0086] In one optional embodiment of this disclosure, the correspondence between set identifiers and account identifiers can also be stored for easy querying and retrieval. Therefore, in this embodiment, the method further includes: obtaining the remaining account identifiers in the target person relationship graph excluding the set identifier; and storing the association between the set identifier and the remaining account identifiers for easy querying of account identifiers under the same set identifier.

[0087] For example, after generating the target person relationship graph, the account identifiers contained in the target person relationship graph can be stored. Since the set identifiers contained in the unique account identifier set are determined from the unique account identifiers with existing person relationships, and these unique account identifiers all exist in the target person relationship graph, the set identifier is an account identifier contained in the target person relationship graph. Therefore, in this embodiment of the disclosure, after determining the set identifier corresponding to the unique account identifier set, the remaining account identifiers other than the set identifier can be determined from the target person relationship graph. The remaining account identifiers and the set identifier belong to the same family, so the association relationship between the set identifier and the remaining account identifiers can be stored, so that other account identifiers in the target person relationship graph are associated with the determined set identifier, so as to facilitate subsequent querying and retrieval, providing convenience for subsequent refined operation.

[0088] Figure 3 This illustration shows a schematic diagram of a target person relationship diagram in an exemplary embodiment of the present disclosure, to... Figure 3 For example, assuming the final set of unique account identifiers includes id1 and id7, and id7 is determined as the set identifier, then... Figure 3 The remaining account identifiers in the target person relationship diagram shown include id1 to id6 and id8. Associating id7 with id1 to id6 and id8 greatly facilitates subsequent querying and retrieval.

[0089] For example, by querying the relationship between a certain set identifier and other account identifiers in the storage, two users can be identified from the account identifiers associated with that set identifier. One user experienced the service of business 1, and the other user experienced the service of business 2. The operations staff can then query the data associated with the set identifier and combine it with user profile tag data to make targeted push notifications and access.

[0090] The scheme disclosed herein employs a graph algorithm, which can mine implicit relationships between users from implicit user data. It is applicable to the mining of interpersonal relationships in multiple scenarios. It is the first to propose to mine relationships by constructing unique account identifiers for user IDs in different business scenarios of educational institutions, and to associate relationships between users by constructing set identifiers, thereby improving the efficiency and accuracy of account identifier association.

[0091] This exemplary embodiment also provides a graph learning-based apparatus for obtaining interpersonal relationships. Figure 4 A schematic block diagram of a graph learning-based interpersonal relationship acquisition apparatus according to an exemplary embodiment of the present disclosure is shown, such as... Figure 4 As shown, the graph learning-based character relationship acquisition device 40 includes: a first acquisition module 410, a second acquisition module 420, a third acquisition module 430, and a determination module 440.

[0092] The first acquisition module 410 is used to acquire multiple character relationship subgraphs, wherein each character relationship subgraph includes a target node consisting of a user's unique account identifier.

[0093] The second acquisition module 420 is used to acquire a target character relationship diagram composed of the first character relationship diagram and the second character relationship diagram in response to determining that at least one node in the first character relationship diagram has an association relationship with at least one node in the second character relationship diagram, for the first character relationship diagram in the plurality of character relationship diagrams, wherein the second character relationship diagram is a character relationship diagram in the plurality of character relationship diagrams other than the first character relationship diagram;

[0094] The third acquisition module 430 is used to acquire the similarity between the target node of the first person relationship subgraph and the target node of the second person relationship subgraph in the target person relationship graph.

[0095] The determination module 440 is used to determine the relationship between the user corresponding to the first person relationship subgraph and the user corresponding to the second person relationship subgraph based on the similarity.

[0096] Optionally, the second acquisition module 420 is further configured to:

[0097] In response to receiving a user's login operation, obtain multiple account identifiers corresponding to the login operation;

[0098] In response to the existence of a first node composed of a first account identifier among the plurality of account identifiers in the first character relationship subgraph, and a second node composed of a second account identifier among the plurality of account identifiers in the second character relationship subgraph, it is determined that the first node and the second node are associated.

[0099] Optionally, the third acquisition module 430 is further configured to:

[0100] Obtain the first feature data of the user corresponding to the first person relationship subgraph and the second feature data of the user corresponding to the second person relationship subgraph;

[0101] Based on the locality-sensitive hashing algorithm, features of the same dimension in the first feature data and the second feature data are mapped into the same group, resulting in multiple groups;

[0102] For each group, based on the Jaccard similarity coefficient algorithm, calculate the Jaccard similarity coefficient between the target node of the first person relationship subgraph and the target node of the second person relationship subgraph in the corresponding dimension of that group;

[0103] Based on the mean value of the Jaccard similarity coefficients of the target nodes of the first character relationship subgraph and the second character relationship subgraph across the multiple grouping dimensions, the similarity between the target nodes of the first character relationship subgraph and the target nodes of the second character relationship subgraph in the target character relationship graph is determined.

[0104] Optionally, the determining module 440 is further configured to:

[0105] The similarity is compared with a preset similarity threshold;

[0106] In response to the similarity being greater than the similarity threshold, it is determined that there is a relationship between the user corresponding to the first person relationship subgraph and the user corresponding to the second person relationship subgraph;

[0107] Furthermore, the graph learning-based interpersonal relationship acquisition device 40 also includes:

[0108] The generation module is used to generate a set of unique account identifiers based on the unique account identifiers corresponding to the target nodes of the first person relationship subgraph and the target nodes of the second person relationship subgraph, respectively.

[0109] Optionally, the graph learning-based interpersonal relationship acquisition device 40 further includes:

[0110] The fourth acquisition module is used to acquire the business identifier of the business to which each unique account identifier belongs in the set of unique account identifiers;

[0111] The query module is used to query the priority order of different preset business identifiers and determine the target business identifier with the highest priority from the business identifiers;

[0112] The first identifier determination module is used to determine the unique account identifier belonging to the business corresponding to the target business identifier in the unique account identifier set as the set identifier corresponding to the unique account identifier set.

[0113] Optionally, the graph learning-based interpersonal relationship acquisition device 40 further includes:

[0114] The fifth acquisition module is used to acquire the generation time of each unique account identifier in the set of unique account identifiers;

[0115] The second identifier determination module is used to determine the unique account identifier with the earliest generation time as the set identifier corresponding to the unique account identifier set.

[0116] Optionally, the graph learning-based interpersonal relationship acquisition device 40 further includes:

[0117] The sixth acquisition module is used to acquire the remaining account identifiers in the target person relationship diagram, excluding the set identifiers.

[0118] The storage module is used to store the association between the set identifier and the remaining account identifiers, so as to facilitate querying account identifiers under the same set identifier.

[0119] The graph learning-based person relationship acquisition device provided in this disclosure can execute any graph learning-based person relationship acquisition method applicable to electronic devices provided in this disclosure, and has the corresponding functional modules and beneficial effects of the method execution. Content not described in detail in the device embodiments of this disclosure can be referred to the description in any method embodiment of this disclosure.

[0120] Exemplary embodiments of this disclosure also provide an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor. The memory stores a computer program executable by the at least one processor, the computer program, when executed by the at least one processor, causing the electronic device to perform a graph learning-based person-relationship acquisition method according to embodiments of this disclosure.

[0121] Exemplary embodiments of this disclosure also provide a non-transitory computer-readable storage medium storing a computer program, wherein the computer program, when executed by a computer's processor, is used to cause the computer to perform a graph learning-based person relationship acquisition method according to embodiments of this disclosure.

[0122] Exemplary embodiments of this disclosure also provide a computer program product, including a computer program, wherein, when executed by a computer's processor, the computer program is used to cause the computer to perform a graph learning-based person relationship acquisition method according to embodiments of this disclosure.

[0123] refer to Figure 5The present invention describes a structural block diagram of an electronic device 1100 that can serve as a server or client of the present disclosure, which is an example of a hardware device that can be applied to various aspects of the present disclosure. The electronic device is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0124] like Figure 5 As shown, the electronic device 1100 includes a computing unit 1101, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 1102 or a computer program loaded from a storage unit 1108 into a random access memory (RAM) 1103. The RAM 1103 may also store various programs and data required for the operation of the device 1100. The computing unit 1101, ROM 1102, and RAM 1103 are interconnected via a bus 1104. An input / output (I / O) interface 1105 is also connected to the bus 1104.

[0125] Multiple components in electronic device 1100 are connected to I / O interface 1105, including: input unit 1106, output unit 1107, storage unit 1108, and communication unit 1109. Input unit 1106 can be any type of device capable of inputting information to electronic device 1100. Input unit 1106 can receive input digital or character information and generate key signal inputs related to user settings and / or function control of electronic device. Output unit 1107 can be any type of device capable of presenting information and may include, but is not limited to, a display, speaker, video / audio output terminal, vibrator, and / or printer. Storage unit 1108 may include, but is not limited to, disk and optical disk. Communication unit 1109 allows electronic device 1100 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks, and may include, but is not limited to, modems, network cards, infrared communication devices, wireless communication transceivers, and / or chipsets, such as Bluetooth™ devices, WiFi devices, WiMax devices, cellular communication devices, and / or the like.

[0126] The computing unit 1101 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1101 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1101 performs the various methods and processes described above. For example, in some embodiments, the graph learning-based person-relationship acquisition method can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 1108. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 1100 via ROM 1102 and / or communication unit 1109. In some embodiments, the computing unit 1101 can be configured to perform the graph learning-based person-relationship acquisition method by any other suitable means (e.g., by means of firmware).

[0127] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0128] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0129] As used in this disclosure, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, device, and / or apparatus (e.g., disk, optical disk, memory, programmable logic device (PLD)) for providing machine instructions and / or data to a programmable processor, including machine-readable media that receive machine instructions as machine-readable signals. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor.

[0130] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0131] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0132] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other.

Claims

1. A graph learning-based method for obtaining interpersonal relationships, wherein, The method includes: Obtain multiple character relationship subgraphs, where each character relationship subgraph includes target nodes consisting of unique account identifiers for users; For a first character relationship subgraph in the plurality of character relationship subgraphs, in response to determining that at least one node in the first character relationship subgraph is associated with at least one node in the second character relationship subgraph, a target character relationship graph composed of the first character relationship subgraph and the second character relationship subgraph is obtained, wherein the second character relationship subgraph is a character relationship subgraph in the plurality of character relationship subgraphs other than the first character relationship subgraph; Obtain the similarity between the target nodes of the first character relationship subgraph and the target nodes of the second character relationship subgraph in the target character relationship graph; Based on the similarity, the relationship between the user corresponding to the first person relationship subgraph and the user corresponding to the second person relationship subgraph is determined. The step of determining that at least one node in the first person relationship subgraph is associated with at least one node in the second person relationship subgraph includes: In response to receiving a user's login operation, obtain multiple account identifiers corresponding to the login operation; In response to the existence of a first node composed of a first account identifier among the plurality of account identifiers in the first character relationship subgraph, and a second node composed of a second account identifier among the plurality of account identifiers in the second character relationship subgraph, it is determined that the first node and the second node are associated.

2. The graph learning-based method for obtaining interpersonal relationships as described in claim 1, wherein, The step of obtaining the similarity between the target nodes of the first person relationship subgraph and the target nodes of the second person relationship subgraph in the target person relationship graph includes: Obtain the first feature data of the user corresponding to the first person relationship subgraph and the second feature data of the user corresponding to the second person relationship subgraph; Based on the locality-sensitive hashing algorithm, features of the same dimension in the first feature data and the second feature data are mapped into the same group, resulting in multiple groups; For each group, based on the Jaccard similarity coefficient algorithm, calculate the Jaccard similarity coefficient between the target node of the first person relationship subgraph and the target node of the second person relationship subgraph in the corresponding dimension of that group; Based on the mean value of the Jaccard similarity coefficients of the target nodes of the first character relationship subgraph and the second character relationship subgraph across the multiple grouping dimensions, the similarity between the target nodes of the first character relationship subgraph and the target nodes of the second character relationship subgraph in the target character relationship graph is determined.

3. The graph learning-based method for obtaining interpersonal relationships as described in any one of claims 1-2, wherein, The step of determining the relationship between the user corresponding to the first person relationship subgraph and the user corresponding to the second person relationship subgraph based on the similarity includes: The similarity is compared with a preset similarity threshold; In response to the similarity being greater than the similarity threshold, it is determined that there is a relationship between the user corresponding to the first person relationship subgraph and the user corresponding to the second person relationship subgraph; Furthermore, the method further includes: A set of unique account identifiers is generated based on the unique account identifiers corresponding to the target nodes of the first and second character relationship subgraphs, respectively.

4. The graph learning-based method for obtaining interpersonal relationships as described in claim 3, wherein, The method further includes: Obtain the business identifier of the business to which each unique account identifier belongs in the set of unique account identifiers; Query the priority order of different preset service identifiers, and determine the target service identifier with the highest priority from the service identifiers; The unique account identifier belonging to the business corresponding to the target business identifier in the unique account identifier set is determined as the set identifier corresponding to the unique account identifier set.

5. The graph learning-based method for obtaining interpersonal relationships as described in claim 3, wherein, The method further includes: Obtain the generation time of each unique account identifier in the set of unique account identifiers; The unique account identifier with the earliest generation time is determined as the set identifier corresponding to the set of unique account identifiers.

6. The graph learning-based method for obtaining interpersonal relationships as described in claim 4 or 5, wherein, The method further includes: Obtain the remaining account identifiers in the target person relationship graph, excluding the set identifiers; Store the association between the set identifier and the remaining account identifiers to facilitate querying account identifiers under the same set identifier.

7. A graph learning-based device for acquiring interpersonal relationships, wherein, The device includes: The first acquisition module is used to acquire multiple character relationship subgraphs, wherein each character relationship subgraph includes a target node consisting of a user's unique account identifier; The second acquisition module is used to acquire a target character relationship diagram composed of the first character relationship diagram and the second character relationship diagram in response to determining that at least one node in the first character relationship diagram has an association relationship with at least one node in the second character relationship diagram, for the first character relationship diagram in the plurality of character relationship diagrams, wherein the second character relationship diagram is a character relationship diagram in the plurality of character relationship diagrams other than the first character relationship diagram; The third acquisition module is used to acquire the similarity between the target nodes of the first character relationship subgraph and the target nodes of the second character relationship subgraph in the target character relationship graph. The determination module is used to determine the relationship between the user corresponding to the first person relationship subgraph and the user corresponding to the second person relationship subgraph based on the similarity. The second acquisition module is further configured to: In response to receiving a user's login operation, obtain multiple account identifiers corresponding to the login operation; In response to the existence of a first node composed of a first account identifier among the plurality of account identifiers in the first character relationship subgraph, and a second node composed of a second account identifier among the plurality of account identifiers in the second character relationship subgraph, it is determined that the first node and the second node are associated.

8. An electronic device, comprising: processor; as well as Stored program memory, The program includes instructions that, when executed by the processor, cause the processor to perform the graph learning-based person relationship acquisition method according to any one of claims 1-6.

9. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to execute the graph learning-based method for obtaining person relationships according to any one of claims 1-6.

Images