Community association identification method and device, computer device, and storage medium

Abstract

This disclosure relates to the field of computer technology and discloses a method, apparatus, computer device, and storage medium for identifying associated communities. The method includes: acquiring each node in a social network and the atomic relationships between them; for any target node among the nodes, determining the graph embedding information corresponding to the target node based on the atomic relationships; constructing a social graph between the target node and all other nodes according to the graph embedding information; and identifying the target associated community from the social graph based on the node relationships in the social graph. By implementing the technical solution of this disclosure, multi-path information between nodes can be considered, facilitating better capture of complex relationships and features between nodes, and fully utilizing the graph data structure information of the social graph to achieve the identification of associated communities.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and specifically to a method, apparatus, computer device, and storage medium for identifying associated communities. Background Technology

[0002] With the rapid development of the internet and mobile internet, the scale and number of social connections have exploded, making the relationships between social connections more complex and diverse. Community detection algorithms in related technologies typically only consider direct connections between nodes. When processing large-scale social graphs, these algorithms suffer from high computational complexity, significant time and space overhead, and difficulty in effectively identifying and characterizing associated communities. Summary of the Invention

[0003] In view of this, the present disclosure provides a method, apparatus, computer equipment and storage medium for identifying associated communities, in order to solve the problem that community discovery algorithms have difficulty in identifying and characterizing associated communities.

[0004] In a first aspect, this disclosure provides a method for identifying associated communities, including: obtaining each node in a social network and the atomic relationships between each node; for any target node among the nodes, determining the graph embedding information corresponding to the target node based on the atomic relationships; constructing a social graph between the target node and each other node according to the graph embedding information; and identifying the target associated community from the social graph based on the node relationships in the social graph.

[0005] The method for identifying associated communities provided in this disclosure obtains the atomic relationships between nodes in a social network, and determines the graph embedding information corresponding to each node in the social network one by one based on the atomic relationships. This allows for consideration of multi-path information between nodes, facilitating better capture of complex relationships and features between nodes. Subsequently, the social graph between nodes is reconstructed using the graph embedding information to identify the target associated community from the social graph. This fully utilizes the graph data structure information of the social graph to achieve the identification of associated communities.

[0006] Secondly, this disclosure provides a device for identifying associated communities, comprising: an acquisition module for acquiring various nodes in a social network and the atomic relationships between them; a graph embedding determination module for determining the graph embedding information corresponding to any target node based on the atomic relationships; a social graph construction module for constructing a social graph between the target node and other nodes according to the graph embedding information; and an associated community identification module for identifying the target associated community from the social graph based on the node relationships in the social graph.

[0007] Thirdly, this disclosure provides a computer device, including: a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the method for identifying the associated community described in the first aspect or any corresponding embodiment.

[0008] Fourthly, this disclosure provides a computer-readable storage medium storing computer instructions for causing a computer to perform the association community identification method of the first aspect or any corresponding embodiment described above. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the specific embodiments of this disclosure or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0010] Figure 1 This is a flowchart illustrating a method for identifying associated communities according to an embodiment of this disclosure;

[0011] Figure 2 This is a flowchart illustrating another method for identifying associated communities according to an embodiment of this disclosure;

[0012] Figure 3 This is a flowchart illustrating another method for identifying associated communities according to an embodiment of the present disclosure;

[0013] Figure 4 This is a structural block diagram of an associated community identification device according to an embodiment of the present disclosure;

[0014] Figure 5 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present disclosure. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0016] With the rapid development of the internet and mobile internet, the scale and number of social connections have exploded, making the relationships between them more complex and diverse. Community detection algorithms in related technologies suffer from high computational complexity, significant time and space overhead, when processing large-scale social graphs. Furthermore, they typically only consider direct connections between nodes, neglecting multi-path information. Additionally, the relationships and characteristics between nodes in social graphs are often heterogeneous and non-homogeneous; for example, social connections exhibit significant differences in interests and behaviors. Community detection algorithms struggle to handle this non-homogeneity and heterogeneity, making it difficult to effectively identify and characterize interconnected communities.

[0017] Based on this, the technical solution disclosed herein uses atomic relations to determine the graph embedding information between nodes, and combines graph embedding to reconstruct the social graph for mining related communities. It can utilize multi-path information to efficiently process large-scale graphs and handle non-uniformity and heterogeneity, thereby achieving more accurate community identification.

[0018] According to the embodiments of this disclosure, an embodiment of a method for identifying associated communities is provided. It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowcharts, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0019] It is understood that before using the technical solutions disclosed in the various embodiments of this disclosure, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this disclosure in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0020] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose whether to provide personal information to the software or hardware, such as the electronic device, application, server, or storage medium performing the operations of this disclosed technical solution, based on the prompt message.

[0021] As an optional but non-limiting implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0022] It is understood that the above notification and user authorization process are merely illustrative and do not constitute a limitation on the implementation of this disclosure. Other methods that comply with relevant laws and regulations may also be applied to the implementation of this disclosure.

[0023] This embodiment provides a method for identifying associated communities, which can be used on computer devices such as computers and tablets. Figure 1 This is a flowchart of a method for identifying associated communities according to embodiments of this disclosure, such as... Figure 1 As shown, the process includes the following steps:

[0024] Step S101: Obtain the nodes in the social network and the atomic relationships between them.

[0025] A social network is a relational network for social activities among individuals or organizations, with each node representing a social object within the network. Atomic relationships represent the connections or paths between nodes, such as following, being followed, liking, being liked, commenting, private messaging, and watching / viewing.

[0026] When social objects engage in social activities on a social networking platform, they act as nodes within the network, and their relationships with other social objects are considered atomic relationships. Specifically, the social networking platform has a corresponding social application deployed on a computer device. When social objects engage in social activities through this application, the computer device can identify each social object currently active on the platform, thus determining the nodes within the social network. The relationships between these nodes are recorded in the social networking platform's data center, and the computer device can then retrieve the atomic relationships between these nodes from this data center.

[0027] Step S102: For any target node among all nodes, determine the graph embedding information corresponding to the target node based on atomic relationships.

[0028] Graph embedding information is used to characterize the relationships and features between each node and other nodes in a social network. After identifying multiple nodes in the social network, any one node is selected as the target node. Then, based on the atomic relationships between nodes, other nodes with relationships and features with the target node are identified, thus determining the graph embedding information corresponding to the target node. Subsequently, the graph embedding information corresponding to each node in the social network can be obtained.

[0029] Step S103: Construct a social graph between the target node and all other nodes according to the graph embedding information.

[0030] A social graph is a graph data structure in a social network, consisting of nodes and the relationships between them. It describes the social relationships and behaviors among various social entities within the social network. Based on the associations and characteristics between nodes included in the graph embedding information, the similarities and differences between nodes can be determined.

[0031] By combining the similarities and differences between nodes, social paths are formed between each node. A social graph is constructed between each node according to the social paths between nodes, and the social graph describes the similarity relationship and neighbor relationship between social objects.

[0032] Step S104: Identify the target associated community from the social graph based on the node relationships in the social graph.

[0033] The target association community is a social region formed by similar social objects; node relationships are used to represent the similarity and difference relationships between nodes. After generating the social graph, multiple nodes in the social graph are divided into different groups of similar social objects according to node relationships, and the social regions constructed by the groups of similar social objects are identified as the target association community.

[0034] The method for identifying associated communities provided in this embodiment obtains the atomic relationships between nodes in a social network, and then determines the graph embedding information corresponding to each node in the social network based on the atomic relationships. This allows for consideration of multi-path information between nodes, facilitating the better capture of complex relationships and features between nodes. Subsequently, the social graph between nodes is reconstructed using the graph embedding information to identify the target associated community from the social graph. This fully utilizes the graph data structure information of the social graph to achieve the identification of associated communities.

[0035] This embodiment provides a method for identifying associated communities, which can be used on computer devices such as computers and tablets. Figure 2 This is a flowchart of a method for identifying associated communities according to embodiments of this disclosure, such as... Figure 2 As shown, the process includes the following steps:

[0036] Step S201: Obtain the individual nodes in the social network and the atomic relationships between them. For detailed explanation, please refer to the relevant descriptions in the above embodiments; they will not be repeated here.

[0037] Step S202: For any target node among all nodes, determine the graph embedding information corresponding to the target node based on atomic relationships.

[0038] Specifically, step S202 above may include:

[0039] Step S2021: Determine the social features of the target node and each other node based on atomic relationships.

[0040] Social features represent the characteristics generated when a target node forms social relationships with other nodes, such as the attention features generated when a follow relationship is formed. Since atomic relationships represent the associations or paths between nodes, the social features of each node can be determined by combining atomic relationships. For example, for a target node to be similar to any other node, it needs to have nodes it both follows: target node 1 - follows - node 2 - followed - node 3. Here, node 2 is a node that target node 1 and any node 3 both follow. Combining atomic relationships, their association can be determined as: target node 1 follows node 2, and node 2 is followed by any node 3. Therefore, target node 1 and any node 3 can be considered similar nodes, and the social features between the target node and this node can be determined as [follow, followed].

[0041] As an optional implementation, step S2021 above may include:

[0042] Step a1: Determine the node type and relationship type of the target node and all other nodes based on atomic relationships.

[0043] Node type represents the entity that a node in a social network represents, such as a social object or social organization. Relationship type represents the relationship between nodes in a social network, such as friend relationship, follow relationship, and comment relationship.

[0044] Node type and relationship type are attributes of nodes in a social network. By combining the association relationships and association paths between nodes represented by atomic relationships, the node type and relationship type corresponding to each node can be determined.

[0045] Step a2: Determine the social characteristics of the target node and each other node according to the node type and relationship type.

[0046] Path patterns are defined and generated based on node and relation types in the data to capture semantic relationships and patterns between nodes. A metapath consists of a series of node and relation types, describing the order of nodes and edges in the path. By defining and using different metapaths, specific subgraphs or substructures can be extracted from the network, enabling deeper analysis and mining.

[0047] By combining a series of node types and relationship types, the semantic relationships and path generation patterns between the target node and other nodes can be captured. Based on the semantic relationships and path generation patterns, the order of the target node, other nodes, and the connecting edges between the target node and other nodes during the path generation process can be determined. Based on the paths between the target node and each other node, the association relationships between the target node and each other node can be determined, and then the social characteristics of the target node and each other node can be determined based on these association relationships.

[0048] In the above implementation, constructing meta-paths between nodes based on node type and relationship type to determine social features can better capture complex node relationships and semantic patterns, thereby making better use of multi-path information and improving the accuracy of subsequent social graph construction. Simultaneously, by constructing meta-paths containing multiple node types and relationship types, the heterogeneity and non-homogeneity between nodes can be effectively handled, thus improving community detection performance.

[0049] Step S2022: Identify multiple other nodes that have social relationships with the target node based on social characteristics.

[0050] By combining the social characteristics of the target node and other nodes identified above, it can be determined whether there are corresponding or similar social characteristics between the target node and other nodes. When other nodes with corresponding or similar social characteristics to the target node are identified, it can be determined that there is a corresponding social relationship between the target node and those other nodes.

[0051] Step S2023: Based on the target node and multiple other nodes, determine the graph embedding information corresponding to the target node.

[0052] Based on the social relationships and characteristics between the target node and multiple other nodes, the embedding relationship between the target node and multiple other nodes is determined. The embedding data structure between the target node and multiple other nodes is determined in combination with the embedding relationship. Then, the graph embedding information between the target node and other nodes is determined based on the embedding data structure. The subsequent social graph is constructed using the graph embedding information.

[0053] As an optional implementation, the graph embedding information can be determined using a pre-trained graph embedding model, the training method of which specifically includes:

[0054] Step b1: Obtain the node sample set and the atomic relationship sample set between nodes. The atomic relationship sample set includes atomic relationship samples of various types, and the node sample set includes multiple node samples. Each node sample has a corresponding node type and relationship type.

[0055] Step b2: Using node type and relation type, determine multiple node sample pairs, and use atomic relation samples to determine the positive and negative social paths between multiple node sample pairs.

[0056] Step b3: For any target node sample in the node sample set, use a graph embedding model to perform path reasoning on multiple positive social paths and multiple negative social paths to output the positive predicted social path and negative predicted social path corresponding to the target node sample.

[0057] Step b4: Compare the positive and negative predicted social paths, and determine the error information between the positive and negative predicted social paths based on the comparison results.

[0058] Step b5: Correct the graph embedding model based on the error information.

[0059] The node sample set is a collection of pre-collected node samples existing in the social network. This set includes multiple node samples, each with a corresponding node type and relationship type. Specifically, this node sample set can be obtained through online collection or by accessing the social network platform's database; the specific method for obtaining the node sample set is not limited here.

[0060] The atomic relation sample set is a pre-collected set of relationships between nodes in a social network, including multiple atomic relation samples of different types. Specifically, this atomic relation sample set can be obtained through pre-simulation or by accessing the database of the social network platform; the method of obtaining the atomic relation sample set is not specifically limited here.

[0061] Node sample pairs include positive node sample pairs and negative node sample pairs. Based on the node type and relation type corresponding to each node sample, and combined with atomic relations, multiple similar node sample pairs are determined, and these similar node sample pairs are used as positive node sample pairs. At the same time, based on atomic relations, differing node sample pairs are determined, and these differing node sample pairs are used as negative node sample pairs.

[0062] Since atomic relations can characterize the relationships between node samples, after constructing positive and negative node sample pairs, we can construct a metapath for positive social paths between positive node sample pairs by combining atomic relations. At the same time, we can simulate a positive social path between negative node sample pairs by combining atomic relations.

[0063] When using graph embedding models for inference, the same input metapath is needed to generate new node embeddings. The input metapath of a graph embedding model is typically a sequence of multiple atomic relations used to describe the multivariate path information between node samples.

[0064] Before using Metapath to generate node embeddings, a self-supervised learning method (such as InfoNCE) can be used to train the graph embedding model, resulting in more accurate and comprehensive node embeddings. During training, positive and negative node sample pairs are used to construct the loss function of the self-supervised learning method, enabling the graph embedding model to learn the similarities and differences between node samples.

[0065] Specifically, let there be five types of atomic relations: like, follow, comment, social interaction, and watch, defined as r1, r2, r3, r4, and r5 respectively. For any input node sample x... i ∈{x1,x2,……,x N}, r1(x i x j The expression indicates that node sample i liked a submission from node sample j.

[0066] Taking the GraphSAGE model as a graph embedding model as an example, the graph of the GraphSAGE model can be composed of eight subgraphs: [r1, r2, r1], [r2, r1, r2], [r3, r4, r3], [r4, r3, r4], [r5, r1], [r5, r2], [r5, r3], [r5, r4].

[0067] Where [r1, r2, r1] represents a given center sample x i Its corresponding first-order neighbor sample x i,(0,0) Sample x at the same center node i For a like relationship to exist, r1(x) i x i(0,0) ), and its corresponding second-order neighbor x i(0,0),(1,0) First-order neighbor x i(0,0) There is a like relationship r2(x) i,(0,0) ,x i(0,0),(1,0) The corresponding third-order neighbors have a like relationship with the same-order neighbors.

[0068] Taking a subgraph with only two relations as an example, the node sample x can be obtained through a lookup table. i The corresponding 3-dimensional embedding vector i ∈R deThen, for the sampled k2 second-order neighbors, we can obtain e i(0,0),(1) ∈R k2*de .

[0069] For a predefined aggregation function a(·), the samples of first-order neighbor nodes can be obtained by aggregating the samples of second-order neighbor nodes, which can be represented as o. i,(0,0) =a(o i,(0,0),(1) Similarly, for the sampled k1 first-order neighbor nodes, we can obtain o. i,(0) ∈R k1*de Finally, the center node sample of this subgraph can be represented as o. i =a(o i,(0) ).

[0070] After obtaining the embedding vectors of each subgraph, the embedding vectors of each subgraph are concatenated, and a multilayer perceptron (MLP) is used to perform non-linear processing to obtain the corresponding features. The obtained features are then normalized to complete the training of the GrpahSAGE model.

[0071] Meanwhile, during the training of the GrpahSAGE model, the GrpahSAGE model can combine atomic relations to infer positive social paths and multiple negative social paths, and output positive and negative predicted social paths for target node samples.

[0072] A self-supervised (semi-supervised) loss function is constructed using positive and negative social path prediction to predict whether the current node sample has common associations with other node samples, such as nodes with shared interests. Therefore, the GrpahSAGE model is iteratively trained using this self-supervised (semi-supervised) loss function to correct the GrpahSAGE model and make the graph embedding information output by the GrpahSAGE model more accurate.

[0073] In the above implementation, positive node sample pairs and negative node sample pairs are constructed to determine the corresponding social paths. Then, the graph embedding model is trained using a self-supervised learning method, thereby obtaining more accurate and comprehensive graph embedding information and improving the node classification effect.

[0074] As an optional implementation, the above method may further include:

[0075] Step c1: Detect whether there are any new nodes in the social network according to a preset period.

[0076] Step c2: When a new node is added to the social network, update the graph embedding information corresponding to the target node.

[0077] After training the graph embedding model, it was deployed to the production environment of a social network platform to infer the diverse social paths between nodes. Specifically, during deployment, considering factors such as the performance and efficiency of the graph embedding model, distributed computing frameworks and GPU acceleration technologies were used to deploy the model, facilitating the effective processing of large-scale graphs and enabling the identification of more accurate community structures.

[0078] Since social networking platforms are constantly generating new social data, a preset period can be set to perform timed inference on the new social data, thereby achieving timed updates of graph embedded information. This preset period is a pre-defined detection period, such as 10 minutes, 30 minutes, 1 hour, etc., and is not specifically limited here; those skilled in the art can determine it according to actual needs.

[0079] During the process of computer equipment detecting social data generated by social networking platforms according to a preset cycle, it can detect whether there are new nodes (i.e., newly registered social objects on the social networking platform) or new social data in the social network. When it is determined that there are new nodes or new social data in the social network, the new nodes can be embedded into the current social network through a graph embedding model. The association relationships and association features between nodes are updated for the new nodes or new social data, thereby realizing the timed updating of graph embedding information to facilitate similarity measurement when delineating related communities.

[0080] It should be noted that when using graph embedding models for timed inference, considering the scale and increment of social data, as well as the inference efficiency and accuracy of the model, incremental computing and caching techniques can be used to achieve timed inference.

[0081] In the above implementation, by periodically detecting new nodes or new social data in the social network, the graph embedding information corresponding to the target node is updated periodically, so as to realize the periodic update of the social graph and further improve the accuracy of the division of related communities.

[0082] Step S203: Construct a social graph between the target node and all other nodes according to the graph embedding information. For detailed explanation, please refer to the relevant descriptions in the above embodiments; they will not be repeated here.

[0083] Step S204: Based on the node relationships in the social graph, identify the target associated community from the social graph. For detailed explanation, please refer to the relevant descriptions in the above embodiments; they will not be repeated here.

[0084] The method for identifying associated communities provided in this embodiment combines the social characteristics between the target node and other nodes to determine multiple other nodes that have social relationships with the target node, thereby determining the corresponding graph embedding information for constructing a social graph. This allows for the full utilization of diverse community path information to construct the social graph, significantly improving the community discovery effect.

[0085] This embodiment provides a method for identifying associated communities, which can be used on computer devices such as computers and tablets. Figure 3 This is a flowchart of a method for identifying associated communities according to embodiments of this disclosure, such as... Figure 3 As shown, the process includes the following steps:

[0086] Step S301: Obtain the individual nodes in the social network and the atomic relationships between them. For detailed explanation, please refer to the relevant descriptions in the above embodiments; they will not be repeated here.

[0087] Step S302: For any target node among all nodes, determine the graph embedding information corresponding to the target node based on atomic relationships. For detailed explanation, please refer to the relevant descriptions in the above embodiments; they will not be repeated here.

[0088] Step S303: Construct a social graph between the target node and all other nodes according to the graph embedding information.

[0089] Specifically, step S303 above may include:

[0090] Step S3031: Determine the multi-dimensional social paths between the target node and each other node based on graph embedding information.

[0091] Multi-dimensional social paths are the social paths generated by a target node engaging in social activities with multiple other nodes. These multi-dimensional social paths can represent the social connections generated by the interaction between the target node and other nodes.

[0092] As mentioned above, graph embedding information contains the relationships and characteristics between nodes. These relationships and characteristics allow us to construct edges between the target node and other nodes; these edges constitute the multi-dimensional social paths corresponding to the target node. Furthermore, by sequentially treating each node in the social network as the target node, we can determine the multi-dimensional social paths corresponding to each node in the social network.

[0093] As an optional implementation, step S3031 above may include:

[0094] Step d1: Based on graph embedding information, perform feature measurement on the target node and each other node to determine the feature similarity between the target node and each other node.

[0095] Step d2: Construct multi-dimensional social paths between the target node and other nodes based on feature similarity.

[0096] The process involves identifying nodes from graph embedding information. For any target node, other nodes with which it has a relationship are determined. A graph embedding algorithm is then used to generate embedding vectors corresponding to the target node and each of these other nodes. These embedding vectors are used to measure features between the target node and other nodes to determine the feature similarities and differences between them. The graph embedding algorithm can be either the GraphSAGE graph neural network algorithm or the DeepWalk algorithm; no specific limitation is made here.

[0097] Subsequently, by combining the feature similarities between the target node and other nodes, connections are constructed between the target node and all other nodes, forming the multi-dimensional social paths corresponding to the target node. By treating each node as a target node, the multi-dimensional social paths possessed by each node in the social network can be determined.

[0098] In the above implementation, by using a graph neural network algorithm to determine the multi-dimensional social paths between nodes, the processing speed of multi-dimensional social paths is improved, and efficient processing of large-scale data is achieved.

[0099] Step S3032: Construct a social graph according to multiple social paths.

[0100] After obtaining the multiple social paths, the corresponding social graph can be constructed between the nodes using an approximate nearest neighbor algorithm to determine the similarity and neighbor relationships between nodes in the social network. The approximate nearest neighbor algorithm can be either Locality Sensitive Hashing (LSH) or Hierarchical Navigable Small Word (HNSW), but no specific limitation is made here.

[0101] Step S304: Identify the target associated community from the social graph based on the node relationships in the social graph.

[0102] Specifically, step S304 above may include:

[0103] Step S3041: Obtain the node relationships between the nodes in the social graph.

[0104] Node relationships characterize whether there are similarity or difference relationships between nodes. Specifically, nodes with similar relationships in a social graph have corresponding social paths. Here, by identifying the social paths between nodes in the social graph, the node relationships between each node can be determined. If there is no corresponding social path between nodes, it indicates that the node relationship between the two nodes is a difference relationship.

[0105] Step S3042: Based on node relationships, the nodes in the social graph are divided to obtain multiple similar node groups.

[0106] A similarity node group is a set of nodes that have similar features. By identifying nodes with similar characteristics in a social graph through node relationships, similarity node groups can be formed.

[0107] In one specific implementation, a clustering algorithm can be used to divide the nodes in the social graph into different groups of similar nodes. The clustering algorithm can be K-means clustering (KMeans), hierarchical agglomerative clustering (HAC), Gaussian mixture model (GMM), or any other clustering algorithm; no specific limitation is made here.

[0108] In one specific implementation, a community detection algorithm can also be used to mine the characteristic similarities between nodes in the social graph, so as to divide each node into different similar node groups. The community detection algorithm can be the unsupervised heuristic algorithm Louvain, or the community classification algorithm Infomap, etc., without specific limitations here.

[0109] Step S3043: Divide the social structure of the social graph through multiple similar node groups to obtain the target associated communities corresponding to multiple similar node groups.

[0110] The social graph is divided into social structures based on similarity node groups, resulting in social network regions formed by each similar node group. These social network regions are the target associated communities.

[0111] Step S305: Store the target associated community in a preset location, and perform community data analysis based on the stored target associated community.

[0112] The preset location is a pre-defined storage location for storing the target associated communities, such as a database or file system corresponding to a social networking platform. Specifically, after generating the target associated communities, they are stored in the database or file system to facilitate subsequent applications and community data analysis targeting the target associated communities. Considering factors such as the scale and structure of the data, as well as the efficiency of data access and querying, the storage method can be selected according to actual needs; no specific limitations are made here.

[0113] The method for identifying associated communities provided in this embodiment, during the construction of a social graph using multiple social paths, can more accurately measure the similarity between nodes by performing feature similarity measurements, thus facilitating better community segmentation. By segmenting nodes through relationships within the social graph to obtain multiple similar node groups, and then using these similar node groups to segment the social structure of the social graph, a more accurate associated community structure can be identified, achieving accurate mining of associated communities.

[0114] This embodiment also provides an identification device for associated communities, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0115] This embodiment provides a device for identifying associated communities, such as... Figure 4 As shown, it includes:

[0116] The acquisition module 401 is used to acquire the nodes in the social network and the atomic relationships between them.

[0117] The graph embedding determination module 402 is used to determine the graph embedding information corresponding to any target node among all nodes based on atomic relationships.

[0118] The social graph construction module 403 is used to construct a social graph between the target node and each other node according to the graph embedding information.

[0119] The associated community identification module 404 is used to identify the target associated community from the social graph based on the node relationships in the social graph.

[0120] In some alternative implementations, the graph embedding determination module 402 described above may include:

[0121] The social feature determination unit is used to determine the social features of the target node and other nodes based on atomic relationships.

[0122] The associated node determination unit is used to determine multiple other nodes that have social relationships with the target node based on social features.

[0123] The graph embedding information determination unit is used to determine the graph embedding information corresponding to the target node based on the target node and multiple other nodes.

[0124] In some optional implementations, the social feature determination unit described above may include:

[0125] The node information determination sub-unit is used to determine the node type and relationship type of the target node and other nodes based on atomic relationships.

[0126] The feature determination subunit is used to determine the social features of the target node and each other node according to the node type and relationship type.

[0127] In some alternative implementations, the graph embedding determination module 402 described above may further include:

[0128] The model training unit is used to train the graph embedding model to determine graph embedding information through the pre-trained graph embedding model.

[0129] Specifically, the aforementioned model training unit is used to: acquire a set of node samples and a set of atomic relation samples between nodes, wherein the set of atomic relation samples includes multiple types of atomic relation samples, and the set of node samples includes multiple node samples, each with a corresponding node type and relation type; using the node type and relation type, determine multiple node sample pairs, and using the atomic relation samples, determine the positive and negative social paths between the multiple node sample pairs; for any target node sample in the set of node samples, use a graph embedding model to perform path reasoning on multiple positive and multiple negative social paths to output the positive and negative predicted social paths corresponding to the target node sample; compare the positive and negative predicted social paths, and determine the error information between the positive and negative predicted social paths based on the comparison results; and correct the graph embedding model based on the error information.

[0130] In some alternative embodiments, the above-described apparatus may include:

[0131] A new detection module has been added to detect whether new nodes exist in the social network according to a preset period.

[0132] The update module is used to update the graph embedding information of the target node when a new node is added to the social network.

[0133] In some alternative implementations, the social graph construction module 403 described above may include:

[0134] The social path determination unit is used to determine the multi-dimensional social paths between the target node and each other node based on graph embedding information.

[0135] Building units are used to construct social graphs according to multiple social paths.

[0136] In some optional implementations, the social path determination unit described above may include:

[0137] The feature similarity determination subunit is used to measure the features of the target node and each other node based on graph embedding information, and to determine the feature similarity between the target node and each other node.

[0138] The multi-dimensional social path construction subunit is used to construct multi-dimensional social paths between the target node and other nodes based on feature similarity.

[0139] In some optional implementations, the aforementioned associated community identification module 404 may include:

[0140] The node relationship acquisition unit is used to acquire the node relationships that exist between various nodes in the social graph.

[0141] The partitioning unit is used to partition nodes in a social graph based on node relationships, resulting in multiple similar node groups.

[0142] The associated community determination unit is used to divide the social structure of the social graph through multiple similar node groups to obtain the target associated communities corresponding to multiple similar node groups.

[0143] In some alternative embodiments, the above-described apparatus may further include:

[0144] The storage module is used to store the target associated communities in a preset location and to perform community data analysis based on the stored target associated communities.

[0145] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0146] In this embodiment, the identification device for associated communities is presented in the form of a functional unit. Here, a unit refers to an ASIC circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above-mentioned functions.

[0147] The associated community identification device provided in this disclosure acquires the atomic relationships between nodes in a social network, and determines the graph embedding information corresponding to each node in the social network one by one based on the atomic relationships. This allows for consideration of multi-path information between nodes, facilitating better capture of complex relationships and features between nodes. Subsequently, the social graph between nodes is reconstructed through the graph embedding information to identify the target associated community from the social graph. This fully utilizes the graph data structure information of the social graph to achieve associated community identification.

[0148] This disclosure also provides a computer device having the above-described features. Figure 4 The identification device for the associated community is shown.

[0149] Please see Figure 5 , Figure 5 This is a schematic diagram of the structure of a computer device provided in an optional embodiment of this disclosure, such as... Figure 5 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system). Figure 5 Take a processor 10 as an example.

[0150] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0151] The memory 20 stores instructions executable by at least one processor 10 to cause the at least one processor 10 to perform the method shown in the above embodiments.

[0152] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0153] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0154] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0155] This disclosure also provides a computer-readable storage medium in which the methods described in this disclosure can be implemented in hardware or firmware, or implemented as recordable on a storage medium, or implemented as computer code originally stored on a remote storage medium or a non-transitory machine-readable storage medium and subsequently stored on a local storage medium after being downloaded over a network. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium may be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium may also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code that, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0156] Although embodiments of the present disclosure have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present disclosure, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for identifying associated communities, characterized in that, include: Obtain the individual nodes in a social network and the atomic relationships between them; For any target node among the nodes, the graph embedding information corresponding to the target node is determined based on the atomic relationship; Construct a social graph between the target node and all other nodes according to the graph embedding information; Based on the node relationships in the social graph, the target associated community is identified from the social graph; The step of determining the graph embedding information corresponding to the target node based on the atomic relationships includes: determining the social features corresponding to the target node and each other node based on the atomic relationships; determining multiple other nodes that have social relationships with the target node based on the social features; and determining the graph embedding information corresponding to the target node based on the target node and the multiple other nodes. The step of determining the social features corresponding to the target node and each other node based on atomic relationships includes: determining the node type and relationship type of the target node and each other node based on the atomic relationships; and determining the social features corresponding to the target node and each other node according to the node type and relationship type.

2. The method according to claim 1, characterized in that, The graph embedding information is determined through a pre-trained graph embedding model, and the training method of the graph embedding model includes: Obtain a set of node samples and a set of atomic relationship samples between nodes. The set of atomic relationship samples includes multiple types of atomic relationship samples, and the set of node samples includes multiple node samples. Each node sample has a corresponding node type and relationship type. Using the node type and the relationship type, determine the plurality of positive node sample pairs and negative node sample pairs, and use the atomic relationship samples to determine the positive social path and negative social path between the plurality of node sample pairs; For any target node sample in the node sample set, the graph embedding model is used to perform path reasoning on multiple social paths and multiple negative social paths to output the positive predicted social path and negative predicted social path corresponding to the target node sample. By comparing the positive predicted social paths and the negative predicted social paths, error information between the positive predicted social paths and the negative predicted social paths is determined based on the comparison results; The graph embedding model is corrected based on the error information.

3. The method according to claim 1 or 2, characterized in that, Also includes: The system detects whether any new nodes have been added to the social network according to a preset period. When a new node is added to the social network, the graph embedding information corresponding to the target node is updated.

4. The method according to claim 1, characterized in that, The construction of the social graph between the target node and all other nodes according to the graph embedding information includes: Based on the graph embedding information, determine the multi-dimensional social paths between the target node and each other node; The social graph is constructed according to the described multiple social paths.

5. The method according to claim 4, characterized in that, The step of determining the multi-dimensional social paths between the target node and other nodes based on the graph embedding information includes: Based on the graph embedding information, feature measurement is performed on the target node and each other node to determine the feature similarity between the target node and each other node; The multi-dimensional social paths between the target node and each other node are constructed based on the aforementioned feature similarity.

6. The method according to claim 1, characterized in that, The step of identifying target associated communities from the social graph based on node relationships in the social graph includes: Obtain the node relationships between the nodes in the social graph; Based on the node relationships, the nodes in the social graph are divided to obtain multiple similar node groups; The social structure of the social graph is divided by multiple similar node groups to obtain the target associated communities corresponding to the multiple similar node groups.

7. The method according to claim 1, characterized in that, Also includes: The target associated community is stored in a preset location, and community data analysis is performed based on the stored target associated community.

8. A device for identifying communities, characterized in that, include: The acquisition module is used to acquire the individual nodes in the social network and the atomic relationships between them. The graph embedding determination module is used to determine the graph embedding information corresponding to any target node among the nodes based on the atomic relationship. A social graph construction module is used to construct a social graph between the target node and all other nodes according to the graph embedding information. The associated community identification module is used to identify target associated communities from the social graph based on the node relationships in the social graph; The graph embedding determination module includes: a social feature determination unit, used to determine the social features corresponding to the target node and each other node based on atomic relationships; an associated node determination unit, used to determine multiple other nodes that have social relationships with the target node based on the social features; and a graph embedding information determination unit, used to determine the graph embedding information corresponding to the target node based on the target node and the multiple other nodes. The social feature determination unit includes: a node information determination subunit, used to determine the node type and relationship type of the target node and each other node based on the atomic relationship; and a feature determination subunit, used to determine the social features corresponding to the target node and each other node according to the node type and relationship type.

9. A computer device, characterized in that, include: A memory and a processor are communicatively connected, the memory storing computer instructions, and the processor executing the computer instructions to perform the method for identifying associated communities as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing a computer to perform the method for identifying the associated community as described in any one of claims 1 to 7.

Images