Method and device for local community discovery based on node-enhanced hypergraph

Abstract

The application relates to the technical field of data intelligent classification, in particular to a local community discovery method and device based on node-enhanced hypergraph, wherein the method comprises the following steps: inputting initial seed nodes in a hypergraph data set according to an application scene; obtaining an extended seed set based on the initial seed nodes; extracting a quality-enhanced seed node set from the extended seed set, and taking the quality-enhanced seed node set as an initial point to calculate local node diffusion, so as to output a local community discovery subgraph. Therefore, the problems that the community discovery work in the related art can only be applied to small-scale hypergraphs, the applicability of the local community discovery work is reduced, the influence of the quality of the seed nodes on the performance is great, the precision of the local community discovery work is reduced, and the demand of the local community discovery work cannot be met and the like are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data intelligent classification, and in particular relates to a local community discovery method and device based on node-enhanced hypergraph. BACKGROUND

[0002] In related technologies, starting from an initial seed node, a local node is expanded to mine a community structure of the same category as the initial seed node. For the nodes under the community, a seed set is sampled, and the seed set is expanded according to some specific rules to realize community mining. The nodes under the community can be understood as a label.

[0003] However, the community discovery work in related technologies is only applicable to small-scale hypergraphs, which reduces the applicability of the local community discovery work, and the quality of the seed node greatly affects the performance, which reduces the precision of the local community discovery work, cannot meet the needs of the local community discovery work, and needs to be solved urgently. SUMMARY

[0004] The present application is based on the following problems and realizations of the inventors:

[0005] Today is an Internet era of data explosion, and various industries are generating a large amount of data, which contains a large amount of network data, for example, relationship information in social networks, User-Item association in e-commerce recommendations, biological protein networks, and Web networks of search engines.

[0006] These network data contain a large amount of hypergraph structure data. Hypergraph is a generalization of graph. A hyperedge can connect multiple nodes (a hyperedge in a hypergraph can be called a hyperedge). High-order association information is constructed between nodes. Correspondingly, a graph is a special case of a hypergraph. The edges in the graph can be understood as hyperedges that connect only two nodes. Due to the advantages in describing high-order relationships and the structural characteristics of hypergraphs, hypergraphs have more extensive applications on set-form data, for example, modeling users who click on the same product in a recommendation system, modeling proteins containing the same structure in a biological network, and hypergraph structures have been widely used in many real-world applications such as e-commerce, computer vision tasks, gene analysis, and recommendations.

[0007] An important task of hypergraph data analysis is to mine the inherent community structure. Hypergraph networks usually exhibit heterogeneity, that is, the attributes of nodes are not completely consistent, but belong to several categories. Nodes of the same category have more connections, and nodes connected within the same hyperedge have a greater probability of belonging to the same category. Nodes and hyperedges belonging to the same category form a sub-hypergraph, which becomes a community in the hypergraph network. How to mine the community structure from the network is one of the important tasks in hypergraph network analysis.

[0008] With the data scale under the Internet environment getting larger, the global community discovery efficiency is low, which cannot meet the needs of application scenarios. In this case, researchers further study the local community discovery algorithm, that is, starting from the initial seed node, the same community structure as the initial seed node is mined by expanding the local node. These works use a local hypergraph clustering framework based on stream computing, which has a working mode similar to PageRank. The nodes (which can be understood as a label) under the community are sampled to the seed set, and the seed set is expanded according to some specific rules to realize community mining. However, this method is carried out on a small-scale community, and the scale is within a community of ten thousand nodes. In actual scenarios, the data scale is usually difficult to control. Therefore, the research on community discovery of large-scale hypergraph is more important and has practical significance.

[0009] The application provides a local community discovery method and device based on node-enhanced hypergraph, to solve the problems in the related art that the community discovery work can only be applied to small-scale hypergraph, the applicability of the local community discovery work is reduced, the quality of the seed node has a great influence on the performance, the precision of the local community discovery work is reduced, and the needs of the local community discovery work cannot be met.

[0010] The first aspect embodiment of the application provides a local community discovery method based on node-enhanced hypergraph, including the following steps: inputting an initial seed node in a hypergraph data set according to an application scenario; obtaining an expanded seed set based on the initial seed node; extracting a quality-enhanced seed node set from the expanded seed set, and taking the quality-enhanced seed node set as an initial point to calculate local node diffusion, to output a local community discovery subgraph.

[0011] Optionally, in an embodiment of the application, the obtaining of the expanded seed set based on the initial seed node includes: calculating the first-order neighbors of the seed set; calculating the node ranking on the sub-hypergraph composed of the neighbor nodes and the seed set by random walk to reserve the first preset nodes, and calculating the sub-hypergraph of the first preset nodes combined with the initial seed node; calculating the hypergraph partition on the sub-hypergraph, outputting the node set where the seed node is located, and obtaining the expanded seed set.

[0012] Optionally, in an embodiment of the application, the extracting of the quality-enhanced seed node set from the expanded seed set includes: calculating the dense node group on the expanded seed set; calculating the purification output of the dense nodes as the expanded seed set, to obtain an initial seed node set; and merging the initial seed node set to obtain the quality-enhanced seed node set.

[0013] Optionally, in an embodiment of the present application, the calculating local node diffusion based on the quality-enhanced seed node set as an initial point comprises: performing diffusion on the local subset by using a HyperLocal algorithm to obtain an expansion result of the local node.

[0014] The second aspect embodiment of the present application provides a local community discovery device based on node-enhanced hypergraph, comprising: an input module configured to input initial seed nodes in a hypergraph dataset according to an application scenario; an acquisition module configured to acquire an expanded seed set based on the initial seed nodes; and an output module configured to extract a quality-enhanced seed node set from the expanded seed set, and calculate local node diffusion based on the quality-enhanced seed node set as an initial point, to output a local community discovery subgraph.

[0015] Optionally, in an embodiment of the present application, the acquisition module comprises: a first calculation unit configured to calculate first-order neighbors of the seed set; a second calculation unit configured to calculate node ranking on a sub-hypergraph composed of neighbor nodes and the seed set by random walk, to reserve a preset number of nodes, and calculate a sub-hypergraph of the preset number of nodes combined with the initial seed nodes; and a first acquisition unit configured to calculate hypergraph partitioning on the sub-hypergraph, to output a node set in which the seed nodes are located, and to obtain the expanded seed set.

[0016] Optionally, in an embodiment of the present application, the output module comprises: a third calculation unit configured to calculate dense node groups on the expanded seed set; a second acquisition unit configured to calculate a purification output of the dense nodes as the expanded seed set, to obtain an initial seed node set; and a third acquisition unit configured to merge the initial seed node set, to obtain the quality-enhanced seed node set.

[0017] Optionally, in an embodiment of the present application, the output module is further configured to perform diffusion on the local subset by using a HyperLocal algorithm to obtain an expansion result of the local node.

[0018] The third aspect embodiment of the present application provides an electronic device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the local community discovery method based on node-enhanced hypergraph as described in the above embodiments.

[0019] The fourth aspect embodiment of the present application provides a computer readable storage medium, which stores a computer program executable by a processor to implement the local community discovery method based on node-enhanced hypergraph as described above.

[0020] The embodiment of the present application can input an initial seed node in the hypergraph data set according to an application scenario, obtain an extended seed set based on the initial seed node, extract a quality enhanced seed node set from the extended seed set, and take the quality enhanced seed node set as an initial point to calculate local node diffusion to output a local community discovery subgraph, thereby improving the applicability of the local community discovery work and improving the precision of the local community discovery work. Therefore, the problems in the related art that the community discovery work can only be applied to small-scale hypergraphs, the applicability of the local community discovery work is reduced, the quality of the seed node greatly affects the performance, the precision of the local community discovery work is reduced, and the demand of the local community discovery work cannot be met are solved.

[0021] The additional aspects and advantages of the present application will be in part apparent and in part pointed out hereinafter in the description. BRIEF DESCRIPTION OF DRAWINGS

[0022] The above and / or additional aspects and advantages of the present application will become apparent and be readily appreciated from the following description, including the accompanying drawings, wherein:

[0023] Figure 1 A flowchart of a local community discovery method based on a node enhanced hypergraph according to an embodiment of the present application is provided.

[0024] Figure 2 A schematic diagram of a local community discovery based on a node enhanced hypergraph for an embodiment of the present application is provided.

[0025] Figure 3 A structural schematic diagram of a local community discovery device based on a node enhanced hypergraph according to an embodiment of the present application is provided.

[0026] Figure 4 A structural schematic diagram of an electronic device according to an embodiment of the present application is provided. DETAILED DESCRIPTION

[0027] The embodiments of the present application are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar notations represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the accompanying drawings are exemplary and are intended to explain the present application, and cannot be understood as limiting the present application.

[0028] A node-enhanced hypergraph-based local community discovery method and device are described below with reference to the accompanying drawings. In view of the fact that the related art mentioned above can only be applied to small-scale hypergraph community discovery, which reduces the applicability of the local community discovery, and the quality of the seed nodes has a great impact on the performance, which reduces the accuracy of the local community discovery and cannot meet the needs of the local community discovery, the present application provides a node-enhanced hypergraph-based local community discovery method. In this method, the initial seed nodes in the hypergraph dataset can be input according to the application scenario, the expanded seed set can be obtained based on the initial seed nodes, the quality-enhanced seed node set can be extracted from the expanded seed set, and the local node diffusion can be calculated based on the quality-enhanced seed node set as the initial point to output the local community discovery subgraph, thereby improving the applicability of the local community discovery and the accuracy of the local community discovery. Thus, the problems of the related art that can only be applied to small-scale hypergraph community discovery, which reduces the applicability of the local community discovery, and the quality of the seed nodes has a great impact on the performance, which reduces the accuracy of the local community discovery and cannot meet the needs of the local community discovery are solved.

[0029] Specifically, Figure 1 A flowchart of a node-enhanced hypergraph-based local community discovery method provided by the present application is shown.

[0030] As Figure 1 shown, the node-enhanced hypergraph-based local community discovery method includes the following steps:

[0031] In step S101, the initial seed nodes in the hypergraph dataset are input according to the application scenario.

[0032] It can be understood that the initial seed nodes in the hypergraph dataset can be input according to the application scenario, for example, the initial web page can be selected in search engine web page recommendation, or the initial user can be selected in social networks, which effectively improves the executability of the local community discovery.

[0033] In step S102, the expanded seed set is obtained based on the initial seed nodes.

[0034] It can be understood that the expanded seed set can be obtained based on the initial seed nodes in the following steps, for example, the hypergraph structure H=(V, E) and the initial seed set Rs can be obtained, and the seed set can be preprocessed, that is, the input hypergraph H is converted into a graph G through star expansion, thereby effectively improving the quality of the seed nodes.

[0035] Optionally, in an embodiment of the present application, the extended seed set is obtained based on the initial seed nodes, comprising: calculating first-order neighbors of the seed set; ranking nodes on a sub-hypergraph composed of the neighbor nodes and the seed set by random walk to reserve a preset number of nodes, calculating a sub-hypergraph of the preset number of nodes combined with the initial seed nodes; calculating hypergraph partitioning on the sub-hypergraph, outputting a node set in which the seed nodes are located, and obtaining the extended seed set.

[0036] For example, as shown in Figure 2 , the embodiment of the present application can perform seed node expansion, can input the seed set Rs and the graph G through Personalized PageRank random walk, obtain ranking scores F_score of first-order neighbors of the seed set Rs, reserve nodes with the top K F_score values and the initial seed set to form a sub-hypergraph H_k, and can perform sweep cut on the sub-hypergraph H_k, obtain a reference set R containing the initial seed set by partitioning the sub-hypergraph, and thus the flexibility of local community discovery work is improved.

[0037] It should be noted that the preset number of nodes is set by a person skilled in the art according to actual conditions, and is not specifically limited here.

[0038] In step S103, a quality-enhanced seed node set is extracted from the extended seed set, and local node diffusion is calculated with the quality-enhanced seed node set as an initial point to output a local community discovery subgraph.

[0039] It can be understood that the embodiment of the present application can extract a quality-enhanced seed node set from the extended seed set in the following steps, and calculate local node diffusion with the quality-enhanced seed node set in the following steps as an initial point to output a local community discovery subgraph, so that quality enhancement can be achieved by expanding and purifying the initial seed nodes, and the precision and applicability of the local community discovery work are effectively improved.

[0040] Optionally, in an embodiment of the present application, the quality-enhanced seed node set is extracted from the extended seed set, comprising: calculating dense node clusters on the extended seed set; calculating the dense nodes as a purifying output of the extended seed set to obtain an initial seed node set; and merging the initial seed node set to obtain the quality-enhanced seed node set.

[0041] For example, as shown in Figure 2 , the embodiment of the present application can perform seed node purification, can construct a sub-hypergraph about the reference set R according to the reference set R, that is:

[0042] H_R=(R, E(R)),

[0043] Wherein, R is the reference set.

[0044] The sub-hypergraph of the reference set R is grown from the seed set Rs, and the information contained in H_R is the complex high-order association between the seed nodes and other nodes.

[0045] Next, it is assumed that the intersection size of R and the real community T is within a reasonable range, i.e.

[0046] μ|Rs|<=|R∩T|,

[0047] where Rs is the seed set, T is the real community, and μ>0 is a constant, which is set as the expansion parameter of the seed set.

[0048] Thus, the part of R∩T can be retrieved in R, and the subset R_ex that can reflect the real community T to the greatest extent is selected, and R_ex is used as the supplementary node set of the seed set Rs, so that the original node set Rs is expanded into an expanded node set that can reflect the cluster T, i.e.

[0049] Rs’=Rs∪R_ex,

[0050] where R_ex is a subset of the real community T.

[0051] Therefore, the embodiment of the application can regard the problem as screening a connected subset from the node set R, and thus the modularity of the hypergraph can be used to partition the internal part of the hypergraph H_R, so as to obtain a classification result in the reference set R, where the modularity of the hypergraph is defined as:

[0052]

[0053] where z is a vector representing the node label, Ω is an affinity function, d is a vector storing the degree of each node, and p is a defined partition vector.

[0054] The affinity function can be used to control the probability of placing a hyperedge in a certain node set, the degree of a node is the number of all edges connected to the node, and the partition vector is used to store the number of clusters in the node set rather than the specific distribution.

[0055] In order to make the algorithm faster, the AON (All-Or-Nothing) affinity function can be used, i.e.

[0056]

[0057] The AON affinity function only considers whether the given edge is completely contained in a single cluster, which is very important for scalability calculation.

[0058] In the reference set R, the seed set Rs spreads to the four around, the whole diffusion process is carried out through random walk, it is assumed that most nodes in R intersection T have the same label z, therefore, the AON affinity function is suitable for local community discovery method.

[0059] Further, the AON affinity function can be substituted into the equation modularity definition, and the variant of modularity can be obtained, that is:

[0060]

[0061] Wherein, J(w) is the part irrelevant to the partition z, and k is the size of the hyperedge.

[0062] The modularity of the hypergraph can be evolved from a kind of Poisson degree correction hypergraph random block model, and it is applied to the Louvain type algorithm to solve the internal partition task of the reference set R.

[0063] Wherein, the Louvain type algorithm can be divided into two stages, in the first stage, all nodes in the graph or hypergraph can be regarded as independent communities, at this time, the number of communities is equal to the number of nodes, for each node i, the community of the adjacent node is tried to move to, the community that makes the modularity maximized after moving is reserved, when no movement occurs, the label of i is unchanged, the process is repeated until the label of all nodes remains stable, in the second stage, the nodes in each cluster can be merged into a super node, the weight of the super node is converted into the sum of the weights of all merged nodes, the weight between communities is converted into the weight of the edge between super nodes, then the two stages are repeated until the label of the node does not change.

[0064] Further, the embodiments of the application can input the reference set R into the Louvain algorithm to obtain the internal clustering in the reference set, and several clusters with larger size in R are taken as the output node set Ro, wherein the size of Ro depends on the seed set expansion parameter mu, that is:

[0065] |Ro|>=4μ|Rs|,

[0066] Wherein, Ro is the output node set.

[0067] Then, the node R_ex with higher degree in Ro is taken as the supplementary node set of the seed set Rs, and | can be obtained:

[0068] R_ex|=μ|Rs|,

[0069] Wherein, R is obtained by random walk of the seed set Rs, most of the nodes in R have stronger relevance with the seed set Rs, therefore, the cluster with large size in R contains nodes with high relevance with Rs, the probability of these nodes belonging to the nodes in the real community T is higher, and selecting the nodes R_ex with high confidence is more conducive to the diffusion of the seed set to the four directions, so that the range that the seed set can explore is larger.

[0070] Further, the embodiment of the present application can combine the seed set Rs and the enhanced seed set R_ex, that is:

[0071] Rs' = Rs U R_ex,

[0072] Therefore, the embodiment of the present application can form an initial seed set with enhanced quality, improve the applicability of the local community discovery work, and improve the precision of the local community discovery work.

[0073] Optionally, in an embodiment of the present application, the local node diffusion is calculated with the quality-enhanced seed node set as the initial point, including: using the HyperLocal algorithm to diffuse the local subset to obtain the expansion result of the local node.

[0074] For example, as shown in the figure, Figure 2 the embodiment of the present application can realize local community discovery through node diffusion, and can use the HyperLocal algorithm to diffuse the local subset, wherein the HyperLocal algorithm is a local algorithm based on maximum flow and minimum cut to solve the HLC (Hypergraph Local Conductance, hypergraph local conductance) minimization problem, and can be divided into two stages, wherein the first stage can regard the reference set R and its first-order neighbor node N(R) as a local sub-hypergraph, that is:

[0075] HL = (R U N(R), E(R U N(R)),

[0076] Then, the linear threshold expansion method is used to convert the local hypergraph H_L into a graph G_L, and a source node s and a sink node t are introduced, in G_L, for each r E Rs', a weight is added. Infinite edge (s, r), for each u E R, a weight is added. αd(u) edge (s, u), for each v E R^bar (R^bar is the threshold expansion node set), a weight is added. εαd(v) edge (v, t), wherein the parameter α E (0, 1) is related to HLC, after using the maximum flow and minimum cut in G_L to obtain the node set S after partitioning, the single-hop neighbor nodes of S are expanded into H_L, and the process is repeated until H_L is no longer expanded.

[0077] Further, in the second stage, the hypergraph local conductance HLC(S*) of the node set S* outputted in the last stage can be calculated as a new a, and then the two stages are repeated until HLC converges.

[0078] Therefore, in the diffusion process, the embodiments of the present application strengthen the connection between the expanded seed set Rs' and the source node s, make full use of the information of Rs', and have a large number of nodes in the node set S obtained by using the maximum flow minimum cut that are highly associated with Rs'. Since R_ex is obtained in the reference set R by the Louvain algorithm and has a large amount of information of the true community T, R_ex can better find the nodes in the true community in the diffusion process.

[0079] In summary, the embodiments of the present application can sort the nodes near the seed set by random walk, and select the smallest conductive subset for segmentation according to the sorting. Since the segmentation method of random walk depends on the selection of the initial seed set, the internal connection of the clustering result is not ideal, and therefore the nodes with good internal connection can be adaptively selected by the random block model, and the nodes with good connection are fully utilized in the second diffusion process. By minimizing the local conductivity, the community structure close to the true value is mined. The embodiments of the present application are strongly local, and the running time only depends on the size of the input set, and the entire hypergraph is not explored, so that the embodiments of the present application can be effectively applied to large-scale hypergraph.

[0080] According to the local community discovery method based on node-enhanced hypergraph proposed in the embodiments of the present application, the initial seed node in the input hypergraph data set can be input according to the application scene, the expanded seed set is obtained based on the initial seed node, the quality-enhanced seed node set is extracted from the expanded seed set, and the local node diffusion is calculated based on the quality-enhanced seed node set as the initial point to output the local community discovery subgraph. Therefore, the applicability of the local community discovery work is improved, and the precision of the local community discovery work is improved. Thus, the community discovery work in the related art can only be applied to small-scale hypergraph, the applicability of the local community discovery work is reduced, the quality of the seed node greatly affects the performance, the precision of the local community discovery work is reduced, and the demand of the local community discovery work cannot be met.

[0081] Secondly, the local community discovery device based on node-enhanced hypergraph according to the embodiments of the present application is described with reference to the accompanying drawings.

[0082] Figure 3 is a block schematic diagram of the local community discovery device based on node-enhanced hypergraph according to the embodiments of the present application.

[0083] As Figure 3As shown, the local community discovery device 10 based on node-enhanced hypergraph includes an input module 100, an acquisition module 200 and an output module 300.

[0084] Specifically, the input module 100 is configured to input an initial seed node in a hypergraph dataset according to an application scenario.

[0085] The acquisition module 200 is configured to acquire an extended seed set based on the initial seed node.

[0086] The output module 300 is configured to extract a quality-enhanced seed node set from the extended seed set, and calculate local node diffusion with the quality-enhanced seed node set as an initial point, to output a local community discovery subgraph.

[0087] Optionally, in an embodiment of the present application, the acquisition module 200 includes a first calculation unit, a second calculation unit and a first acquisition unit.

[0088] The first calculation unit is configured to calculate first-order neighbors of the seed set.

[0089] The second calculation unit is configured to calculate node ranking on a sub-hypergraph composed of the neighbor nodes and the seed set by random walk, to reserve a preset number of nodes, and calculate a sub-hypergraph of the preset number of nodes combined with the initial seed node.

[0090] The first acquisition unit is configured to calculate hypergraph partitioning on the sub-hypergraph, output a node set where the seed node is located, and obtain the extended seed set.

[0091] Optionally, in an embodiment of the present application, the output module 300 includes a third calculation unit, a second acquisition unit and a third acquisition unit.

[0092] The third calculation unit is configured to calculate dense node groups on the extended seed set.

[0093] The second acquisition unit is configured to calculate a purification output of the dense nodes as the extended seed set, and obtain an initial seed node set.

[0094] The third acquisition unit is configured to merge the initial seed node set, to obtain a quality-enhanced seed node set.

[0095] Optionally, in an embodiment of the present application, the output module is further configured to diffuse the local subset by using a HyperLocal algorithm, to obtain an extended result of the local nodes.

[0096] It should be noted that the foregoing explanation and description of the embodiment of the local community discovery method based on node-enhanced hypergraph also applies to the embodiment of the local community discovery device based on node-enhanced hypergraph, which will not be described herein.

[0097] According to the node-enhanced hypergraph-based local community discovery device provided in the embodiments of the present application, the initial seed nodes in the hypergraph data set can be input according to the application scene, the expanded seed set is obtained based on the initial seed nodes, the quality-enhanced seed node set is extracted from the expanded seed set, and the local node diffusion is calculated with the quality-enhanced seed node set as the initial point, so as to output the local community discovery subgraph, thereby improving the applicability of the local community discovery work and improving the precision of the local community discovery work. Thus, the problems in the prior art that the community discovery work can only be applied to small-scale hypergraphs, the applicability of the local community discovery work is reduced, the quality of the seed nodes greatly affects the performance, the precision of the local community discovery work is reduced, and the local community discovery work cannot meet the demand of the local community discovery work are solved.

[0098] Figure 4 The structure schematic diagram of the electronic device provided in the embodiments of the present application is provided. The electronic device can include:

[0099] The memory 401, the processor 402, and the computer program stored in the memory 401 and executable on the processor 402.

[0100] The processor 402 implements the node-enhanced hypergraph-based local community discovery method provided in the above embodiments when executing the program.

[0101] Further, the electronic device further includes:

[0102] The communication interface 403 is used for communication between the memory 401 and the processor 402.

[0103] The memory 401 is used to store the computer program executable on the processor 402.

[0104] The memory 401 can include a high-speed RAM memory, and can also include a non-volatile memory, for example, at least one disk memory.

[0105] If the memory 401, the processor 402, and the communication interface 403 are independently implemented, the communication interface 403, the memory 401, and the processor 402 can be connected to each other through a bus and complete the communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, Figure 4Only one bus or only one type of bus can exist, however.

[0106] Optionally, in a specific implementation, if the memory 401, the processor 402 and the communication interface 403 are integrated on a chip, the memory 401, the processor 402 and the communication interface 403 can complete the communication with each other through an internal interface.

[0107] The processor 402 can be a central processing unit (CPU) or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present application.

[0108] The embodiments of the present application further provide a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the node-enhanced hypergraph based local community discovery method as above.

[0109] In the description of the present specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in any one or N embodiments or examples. In addition, the person skilled in the art can combine and combine the different embodiments or examples described in the present specification and the features of the different embodiments or examples without contradiction.

[0110] In addition, the terms "first", "second" are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "N" is at least two, for example, two, three, etc., unless otherwise explicitly and specifically limited.

[0111] Any processes or methods described in the flowcharts or otherwise described herein can be understood as representing a module, segment, or portion of code that includes one or N steps for implementing the specified logical functions or processes. The scope of a preferred embodiment of the present application encompasses combinations of one or more steps of the described processes, even if not explicitly described in the flowcharts or elsewhere herein, and includes additional implementations that can be performed in an order different from the order shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those having ordinary skill in the art to which the embodiments of the present application pertain.

[0112] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor- containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions, or a combination thereof. For purposes of this specification, a "computer-readable medium" can be any apparatus that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. The computer-readable medium can be a product of "tangible" non-transitory computer-readable medium. More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electronic connection having one or more wires (electronic devices), a portable computer diskette (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber (optical devices), and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via optical scanning of the paper or other medium, then compiled, interpreted, or otherwise processed in a suitable manner, if necessary, and stored in a computer memory.

[0113] It should be understood that aspects of the present application can be implemented in hardware, software, firmware, or combinations thereof. In the above embodiments, the N steps or methods can be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented in hardware and in another embodiment, any of the following technologies, or combinations thereof, can be used: discrete logic circuitry having logic gates for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGA), field programmable gate arrays (FPGA), and the like.

[0114] Those skilled in the art of the present technology can understand that all or part of the steps carried out by the above-mentioned embodiment method can be completed by programs instructing related hardware, and the programs can be stored in a computer readable storage medium. When the program is executed, it includes one of the steps of the method embodiment or a combination thereof.

[0115] In addition, each functional unit in each embodiment of the present application can be integrated into one processing module, or each unit can exist physically alone, or two or more units can be integrated into one module. The integrated module can be realized in the form of hardware or in the form of a software functional module. When the integrated module is realized in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer readable storage medium.

[0116] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. A local community discovery method based on node-enhanced hypergraph, characterized in that: The following steps are involved: Input the initial seed nodes in the hypergraph dataset according to the application scenario; Acquire an extended seed set based on the initial seed node; as well as Extracting a quality enhancement seed node set from the expanded seed set, and using the quality enhancement seed node set as an initial point, calculating local node diffusion to output a local community discovery subgraph; The acquiring of an extended seed set based on the initial seed node includes: Calculate the first-order nearest neighbors of the seed set; Calculating node order by random walk on the sub-hypergraph formed by the neighboring nodes and the seed set to retain the first preset nodes, and calculating the sub-hypergraph of the preset nodes combined with the initial seed nodes; Calculating hypergraph partitioning on the sub-hypergraph, outputting a node set where a seed node is located, and obtaining the extended seed set; The extracting the quality enhancement seed node set from the extended seed set includes: computing dense node clusters on the expanded seed set; Calculating the dense nodes as a purified output of the expanded seed set to obtain an initial seed node set; Merging the initial seed node sets to obtain the quality enhancement seed node set; The calculating of local node diffusion using the quality enhancement seed node set as an initial point includes: The HyperLocal algorithm is used to diffuse the local subset and obtain the expansion result of the local node.

2. A local community discovery device based on node-enhanced hypergraph, characterized in that: include: An input module, used to input the initial seed nodes in the hypergraph dataset according to the application scenario; An acquisition module, configured to acquire an extended seed set based on the initial seed node; as well as an output module, configured to extract a quality enhancement seed node set from the expanded seed set, and calculate local node diffusion using the quality enhancement seed node set as an initial point to output a local community discovery subgraph; The acquisition module includes: A first calculation unit is used to calculate the first-order nearest neighbors of the seed set; A second computing unit is configured to calculate node ordering by random walk on a sub-hypergraph consisting of neighboring nodes and the seed set, so as to retain a preset number of nodes and calculate a sub-hypergraph of the preset number of nodes combined with the initial seed node; A first acquisition unit is configured to calculate hypergraph partitioning on the sub-hypergraph, output a node set where a seed node is located, and obtain the extended seed set; The output module includes: a third computing unit, configured to compute dense node clusters on the extended seed set; A second acquisition unit is configured to calculate the dense nodes as a purified output of the expanded seed set to obtain an initial seed node set; a third acquiring unit, configured to merge the initial seed node set to obtain the quality enhancement seed node set; The output module is further configured to diffuse the local subset using a HyperLocal algorithm to obtain an expansion result of the local node.

3. An electronic device, characterized in that: include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the local community discovery method based on node-enhanced hypergraph according to claim 1.

4. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the local community discovery method based on node-enhanced hypergraph as claimed in claim 1.

Images