Group discovery method based on search condition dense subgraph in large symbol graph

Abstract

The invention provides a group discovery method based on a search condition dense subgraph in a large symbol graph, and relates to the technical field of information processing, and the method comprises the steps: generating an initial symbol graph based on to-be-searched social data; sequencing the vertexes in the initial symbol diagram to obtain a sequencing result, and sequentially generating total branches corresponding to the vertexes according to the sequencing result; performing at least one branching on the total branch to obtain a target sub-branch, the target sub-branch satisfying a preset condition, during each branching, determining a pivot point based on the number of second symbol neighbors of a vertex in a union of a candidate set and a necessary set of the parent branch, and generating the sub-branch based on the pivot point; and determining a target group in the to-be-searched social data based on the target sub-branch with the maximum vertex number. According to the method, the groups which have close relations and consistent relation types can be searched from a large amount of social data, so that the accuracy of content recommendation based on the groups is improved.

Description

Technical Field

[0001] The present invention relates to the field of information processing technology, and in particular to a group discovery method based on searching for conditional dense subgraphs in a large symbolic graph. Background Art

[0002] Mining related users in social platforms can achieve more accurate content recommendations. In the prior art, users who interact with each other are often regarded as related users for content recommendation, that is, the connections between users are regarded as positive relationships to find related users and make similar content recommendations. However, in social platforms, users can express positive or negative comments to other users. The prior art does not consider different types of relationships, resulting in inaccurate related users. For example, two users who interact a lot but actually have opposite opinions are found to be related users and similar content is recommended to them. Summary of the invention

[0003] The present invention provides a group discovery method based on searching conditional dense subgraphs in a large-scale symbolic graph, so as to solve the defect of inaccurate associated users found in the prior art and improve the accuracy of associated users.

[0004] The present invention provides a group discovery method based on searching for conditional dense subgraphs in a large symbolic graph, comprising: Based on the social data to be searched, an initial symbolic graph is generated, wherein vertices in the initial symbolic graph correspond to users in the social data to be searched, edges in the initial symbolic graph correspond to relationships between the users, and symbols of the edges in the initial symbolic graph reflect categories of relationships between the users; Sorting the vertices in the initial symbolic graph to obtain a sorting result, and sequentially generating total branches corresponding to each vertex according to the sorting result, wherein the total branch includes a mandatory set and a candidate set, the mandatory set included in the total branch corresponding to a target vertex includes the target vertex, and the candidate set included in the total branch corresponding to the target vertex includes an associated vertex of the target vertex, and the associated vertex of the target vertex is obtained by removing the vertex before the target vertex in the sorting result from the neighbor vertex set of the target vertex; The total branch is branched at least once to obtain a target sub-branch, and the target sub-branch satisfies a preset condition. At each branch, a pivot point is determined based on the number of second symbolic neighbors of vertices in the union of the candidate set and the required set of the parent branch, and a sub-branch is generated based on the pivot point. The preset condition is that any two vertices in the symbolic subgraph corresponding to the union of the target sub-branch are connected to each other and the number of first symbolic neighbors of any vertex is not less than , is the default value, is the number of vertices in the subgraph corresponding to the target sub-branch; Based on the target sub-branch with the largest number of vertices, a target group in the social data to be searched is determined.

[0005] According to a group discovery method based on searching for conditional dense subgraphs in a large signed graph provided by the present invention, before determining the pivot point based on the number of second signed neighbors of vertices in the union of the candidate set and the required set of the parent branch, the method includes: Obtaining an upper bound of a first symbolic degree and an upper bound of a second symbolic degree in the union, a first symbolic degree and a total degree of candidate vertices in the union; Whether to delete the candidate vertex from the candidate set is determined based on the upper bound of the first signed degree, the upper bound of the second signed degree, the first signed degree, and the total degree.

[0006] According to a group discovery method based on searching for conditional dense subgraphs in a large signed graph provided by the present invention, the obtaining of the upper bound of the first symbol degree and the upper bound of the second symbol degree in the union includes: Obtaining an upper limit of the first symbol degree based on a first preset formula, and obtaining an upper limit of the second symbol degree based on a second preset formula; The first preset formula is: , is the upper bound of the first symbol degree, S is the required set, C is the candidate set, Represents the vertices in the union The first symbol degree of ; The second preset formula is: or ,in represents the upper bound of the second symbol degree, For variables The processing function, express The number of vertices in .

[0007] According to a group discovery method based on searching for conditional dense subgraphs in a large signed graph provided by the present invention, determining whether to delete the candidate vertex from the candidate set based on the upper bound of the first signed degree, the upper bound of the second signed degree, the first signed degree and the total degree includes: When the candidate vertex satisfies the first condition, the second condition or the third condition, deleting the candidate vertex from the candidate set; The first condition is: , , Representation diagram Midpoint The second symbol degree of ; The second condition is: ,in, Indicates the number of vertices in the target sub-branch with the largest number of vertices currently existing; The third condition is: , Represents the vertices in the union The total number of degrees.

[0008] According to a method for discovering a group based on searching for conditional dense subgraphs in a large signed graph provided by the present invention, the pivot point is determined based on the number of second signed neighbors of vertices in the union of the candidate set and the required set of the parent branch, including: Determine a target set based on the union set, wherein the target set includes vertices in the union set whose second symbol degree is not less than an upper bound of the second symbol degree; Filtering the target set to obtain a filtered set, wherein the filtered set includes vertices in the target set that have the most second symbol neighbors; Among the vertices of the filtered set, the point with the most second symbol neighbors in the union is selected as the pivot point.

[0009] According to a group discovery method based on searching for conditional dense subgraphs in a large-scale symbolic graph provided by the present invention, the sub-branch is generated based on the pivot point, comprising: When the pivot point does not belong to the mandatory set of the parent branch, the pivot point is used as the branch point; when the pivot point belongs to the mandatory set of the parent branch, a second symbol neighbor of the pivot point in the candidate set of the parent branch is used as the branch point; Deleting the branch point from the candidate set of the parent branch, generating the candidate set of the first child branch, and using the mandatory set of the parent branch as the mandatory set of the first child branch; The branch point is added to the mandatory set of the parent branch to obtain the mandatory set of the second child branch, and the neighbor set of the branch point in the mandatory set of the parent branch is used as the candidate set of the second child branch.

[0010] The present invention also provides a group discovery device based on searching for conditional dense subgraphs in a large-scale symbolic graph, comprising: A data conversion module, configured to generate an initial symbolic graph based on the social data to be searched, wherein vertices in the initial symbolic graph correspond to users in the social data to be searched, edges in the initial symbolic graph correspond to relationships between the users, and symbols of the edges in the initial symbolic graph reflect categories of relationships between the users; A total branch generation module is used to sort the vertices in the initial symbolic graph to obtain a sorting result, and to generate total branches corresponding to each vertex in sequence according to the sorting result, wherein the total branch includes a mandatory set and a candidate set, the mandatory set included in the total branch corresponding to the target vertex includes the target vertex, and the candidate set included in the total branch corresponding to the target vertex includes an associated vertex of the target vertex, and the associated vertex of the target vertex is obtained by removing the vertex before the target vertex in the sorting result from the neighbor vertex set of the target vertex; A branching module is used to branch the total branch at least once to obtain a target sub-branch, and the target sub-branch meets a preset condition. At each branch, a pivot point is determined based on the number of second symbol neighbors of the vertices in the union of the candidate set and the required set of the parent branch, and a sub-branch is generated based on the pivot point. The preset condition is that any two vertices in the symbol subgraph corresponding to the union of the target sub-branch are connected to each other and the number of first symbol neighbors of any vertex is not less than , is the default value, is the number of vertices in the subgraph corresponding to the target sub-branch; The target group determination module is used to determine the target group in the social data to be searched based on the target sub-branch with the largest number of vertices.

[0011] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, it implements any of the above-described group discovery methods based on searching for conditional dense subgraphs in a large symbolic graph.

[0012] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for discovering a group based on searching for conditional dense subgraphs in a large symbolic graph.

[0013] The present invention also provides a computer program product, comprising a computer program, which, when executed by a processor, implements any of the above-mentioned methods for discovering a group based on searching for conditional dense subgraphs in a large symbolic graph.

[0014] The method for group discovery based on searching conditional dense subgraphs in large symbolic graphs provided by the present invention generates an initial symbolic graph based on the social data to be searched, wherein the vertices in the initial symbolic graph correspond to the users in the social data to be searched, the edges in the initial symbolic graph correspond to the relationships between the users, and the symbols of the edges in the initial symbolic graph reflect the categories of the relationships between the users. The initial symbolic graph is processed to find the largest subgraph that meets the preset conditions, wherein the preset conditions are that any two vertices are connected to each other and the number of first symbolic neighbors of any vertex is not less than , based on the search results, the target group in the social data to be searched is determined. In this way, groups with close relationships and consistent relationship types can be found from a large amount of social data, thereby improving the accuracy of content recommendation based on the group. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 It is a flow chart of a method for discovering a group based on searching for conditional dense subgraphs in a large symbolic graph provided by the present invention.

[0017] Figure 2 It is a schematic diagram of a symbolic graph in a group discovery method based on searching conditional dense subgraphs in a large symbolic graph provided by the present invention.

[0018] Figure 3 It is a structural schematic diagram of a group discovery device based on searching conditional dense subgraphs in a large symbolic graph provided by the present invention.

[0019] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0020] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] Combine the following Figure 1-Figure 2 The present invention provides a method for discovering a group based on searching for conditional dense subgraphs in a large symbolic graph. Figure 1As shown, the group discovery method based on searching for conditional dense subgraphs in a large symbolic graph includes the following steps: S110, generating an initial symbolic graph based on the social data to be searched, where vertices in the initial symbolic graph correspond to users in the social data to be searched, edges in the initial symbolic graph correspond to relationships between users, and symbols of the edges in the initial symbolic graph reflect categories of relationships between users; S120, sorting the vertices in the initial symbolic graph to obtain a sorting result, and generating a total branch corresponding to each vertex in turn according to the sorting result, the total branch including a mandatory set and a candidate set, the mandatory set included in the total branch corresponding to the target vertex includes the target vertex, the candidate set included in the total branch corresponding to the target vertex includes an associated vertex of the target vertex, and the associated vertex of the target vertex is obtained by removing the vertices before the target vertex in the sorting result from the neighbor vertex set of the target vertex; S130, branch the total branch at least once to obtain a target sub-branch, the target sub-branch meets the preset conditions, and at each branch, the pivot point is determined based on the number of second symbolic neighbors of the vertex in the union of the candidate set and the required set of the parent branch, and a sub-branch is generated based on the pivot point, and the preset condition is that any two vertices in the symbolic subgraph corresponding to the union of the target sub-branch are connected to each other and the number of first symbolic neighbors of any vertex is not less than , is the default value, is the number of vertices in the subgraph corresponding to the target sub-branch; S140: Determine a target group in the social data to be searched based on the target sub-branch with the largest number of vertices.

[0022] Specifically, based on the social data to be searched, generating an initial symbol graph may include: Obtaining interaction data between users in the social data to be searched; Analyze the interaction data to obtain the relationship type between each user, where the relationship type includes a first relationship type and a second relationship type; The edges in the initial signed graph are obtained based on the relationship types between the users, where the first signed edge corresponds to the first type and the second signed edge corresponds to the second type.

[0023] The first relationship type may be a positive relationship, that is, the relationship between the two users is positive, for example, they have the same opinion on the same content, and the second relationship type may be a negative relationship, that is, the relationship between the two users is negative, for example, they have opposite opinions on the same content. Of course, the first relationship type may also be a negative relationship, and the second relationship type may also be a positive relationship, and there is no limitation on this.

[0024] When the relationship type between two users is the first relationship type, an edge with the first symbol is added between the vertices corresponding to the two users. When the relationship type between two users is the second relationship type, an edge with the second symbol is added between the vertices corresponding to the two users. For two users who do not have interaction data or whose relationship type cannot be determined based on interaction data, there is no connecting edge between their corresponding vertices. In this way, an initial symbolic graph can be constructed.

[0025] After obtaining the initial symbolic graph, a symbolic subgraph that meets the preset conditions is searched in the initial symbolic graph. The preset conditions are that any two vertices in the subgraph are connected to each other and the number of first symbolic neighbors of any vertex is not less than , is the default value, is the number of vertices in the subgraph corresponding to the target sub-branch. is 0.5, Taking 100 as an example, there is an interactive relationship between any two users in the group corresponding to the symbol subgraph found in this way, and any user in the group found has the same type of relationship with nearly half of the users in the group, indicating that there is a high correlation between the users in the group, and the group has strong common interests or similar social behaviors. After finding such a group, it can not only improve the accuracy of social recommendations, but also help the content recommendation system to more effectively locate the target audience, thereby achieving more efficient content delivery and user experience optimization.

[0026] For the convenience of description, the symbol subgraph that meets the preset conditions is called symbol -clique, when a symbolic graph When the following conditions are met, Is a symbol -clique: for Any two points and , and Connect with each other; for Any point in , .

[0027] in, Representation diagram midpoint The number of first symbol neighbors, that is, and The number of vertices between which there are first signed edges.

[0028] like Figure 2As shown, the first type of edge between two points is represented by a solid line, and the second type of edge is represented by a dotted line. The derived subgraph of point 0124 is a symbol -clique.

[0029] The method provided by the present invention is to find the maximum symbol in the initial symbol graph -clique, to achieve the discovery of target groups.

[0030] Specifically, after the initial symbolic graph is generated, the vertices therein are first sorted. Specifically, degeneracy ordering is used for sorting to obtain the sorting result. Then, through the divide-and-conquer framework, the original search problem on the initial symbolic graph is decomposed into sub-problems on its sub-graphs, and then the pivot-based branch and bound algorithm is called to solve the sub-problems.

[0031] First, decompose the initial symbolic graph into multiple subgraphs , , that is to say, Include vertices as well as The associated vertices of The associated vertex is Remove the neighbor vertices in the initial symbolic graph from the sorted result. The remaining vertices after the vertices between them. For each , a general branch can be generated correspondingly, which includes the mandatory set S and the candidate set C. The mandatory set includes , the candidate set includes The associated vertices of .

[0032] For each main branch, at least one branch is made. The child branches generated by each branch can be used as the parent branch of the next branch to continue to generate child branches. When a branch reaches the stop branch condition, it will no longer be branched. The specific stop branch condition will be explained later. Now we will explain the process of each branch in detail.

[0033] In a possible implementation, in order to improve the search efficiency, before branching the parent branch to generate a child branch, the candidate set of the parent branch is reduced to narrow the search space of the problem. That is, before determining the pivot point based on the number of second symbolic neighbors of the vertex in the union of the candidate set and the required set of the parent branch, it includes: Get the upper bound of the first symbolic degree and the upper bound of the second symbolic degree in the union, and the first symbolic degree and the total degree of the candidate vertices in the union; Whether to delete a candidate vertex from the candidate set is determined based on an upper bound of the first signed degree, an upper bound of the second signed degree, the first signed degree, and the total degree.

[0034] The first signed degree of a vertex is the number of its first signed neighbors, the second signed degree is the number of its second signed neighbors, and the total degree is the number of all neighbors of the vertex. Reflects the maximum symbol that can be found in the branch -The upper bound of the clique size allows the user to evaluate whether the size of the maximum potential solution in the current branch exceeds the set lower bound If it does not exceed, the branch can be pruned. The upper bound of the second symbol degree is Reflects the maximum second symbol degree capacity in the branch. Although the two are different, and Combining them can improve pruning efficiency.

[0035] Obtaining and concentrating an upper bound of a first symbol degree and an upper bound of a second symbol degree, including: Obtaining an upper limit of the first symbol degree based on a first preset formula, and obtaining an upper limit of the second symbol degree based on a second preset formula; The first preset formula is: , , The upper bound of the first symbol degree is S, which is a mandatory set, and C is a candidate set. Represents and collects vertices The first symbol degree of ; The second preset formula is or ,in represents the upper bound of the second symbol degree, .

[0036] Specifically, given a branch B = (S, C), where S is the selected set and C is the candidate set, the upper bound of the first symbol degree of branch B is ,in . That is, the largest symbol that can be found in B -The size of the clique will not exceed .

[0037] For a vertex set H, define its maximum second symbolic degree for: ,definition The reason is to take advantage of its monotonicity property, which can be deduced: ,in There are three vertex sets that satisfy the relationship: .

[0038] For ease of expression, define the function To represent symbols The upper bound of the maximum second symbolic degree of a -clique is of the form: .

[0039] . Then there is, ,Right now , then the upper bound of the second symbol degree of branch B is .

[0040] Due to the inclusion The calculation of is often too loose. To solve this problem, Replace with The upper bound provided and Redefine as That is, in an improved implementation, the upper limit of the second symbol degree of branch B is .

[0041] Determining whether to delete a candidate vertex from the candidate set based on an upper bound of the first signed degree, an upper bound of the second signed degree, the first signed degree, and the total degree includes: When the candidate vertex satisfies the first condition, the second condition or the third condition, the candidate vertex is deleted from the candidate set; The first condition is: , , Representation diagram Midpoint The second symbol degree of ; The second condition is: ,in, Indicates the number of vertices in the target sub-branch with the largest number of vertices currently existing; The third condition is: , Represents and collects vertices The total number of degrees.

[0042] Through the above conditions, the vertices in the candidate set of the branch can be reduced, the search space of the problem can be narrowed, and the efficiency can be improved.

[0043] After the candidate set of the branch is reduced, it is branched to generate a new child branch. The method provided by the present invention uses a branching rule based on a pivot point to generate child branches, thereby minimizing the number of generated child branches as much as possible and improving search efficiency. The pivot point is determined by the number of second symbol neighbors of the vertex in the union of the candidate set and the required set of the parent branch, including: Determine a target set based on the union, the target set including vertices in the union whose second symbol degree is not less than an upper bound of the second symbol degree; Filter the target set to obtain a filtered set, where the filtered set includes the vertices with the most second symbol neighbors in the target set; Among the vertices in the filtered set, the point with the most second symbolic neighbors in the union is selected as the pivot point.

[0044] Given a branch B = (S, C), the target set .

[0045] Then filter Get a subset of it As a filter set, only the vertices with the most second symbolic neighbors in the selected set are retained, that is: ,in Represents the maximum number of second symbolic neighbors adjacent to a vertex in the mandatory set S.

[0046] Notice that Not empty, otherwise the entire graph will be a valid symbol -clique. In construction After that, select exist A point u with the most second symbolic neighbors in is used as the pivot point, that is: .

[0047] Generate sub-branches based on the pivot point, including: When the pivot point does not belong to the mandatory set of the parent branch, the pivot point is used as the branch point. When the pivot point belongs to the mandatory set of the parent branch, a second symbol neighbor of the pivot point in the candidate set of the parent branch is used as the branch point. Delete the branch point in the candidate set of the parent branch, generate the candidate set of the first child branch, and use the mandatory set of the parent branch as the mandatory set of the first child branch; The branch point is added to the mandatory set of the parent branch to obtain the mandatory set of the second child branch, and the neighbor set of the branch point in the mandatory set of the parent branch is used as the candidate set of the second child branch.

[0048] According to the selected pivot point, corresponding branch points and branches are generated. The specific branch rules are as follows: Case 1: If the pivot point , then the branch point for In the candidate set A negative neighbor in .

[0049] Case 2: If the pivot point , then the branch point for .

[0050] For a branch point in the parent branch, two child branches can be generated. The mandatory set in one child branch is the mandatory set of the parent branch, and the candidate set is the candidate set of the parent branch with the branch point removed. The mandatory set of the other child branch is to add the branch point to the mandatory set of the parent branch, and the candidate set is the neighbor set of the branch point in the mandatory set of the parent branch.

[0051] Through the above description, the total branch can be branched layer by layer. For each generated branch, if the stop branch condition is not met, the branch will continue to be branched. Specifically, if the graph corresponding to the current branch Satisfaction symbol -clique requirement, then try to update the current optimal solution (that is, the largest symbol currently found -clique) and terminate the branch if the total number of vertices in the current branch is less than the lower bound (i.e. the current ) or the upper bound of the first symbolic neighbor is smaller than the lower bound, then it is impossible to generate a feasible solution and the current branch is terminated.

[0052] After obtaining the target sub-branch, the user corresponding to the target sub-branch with the largest number of vertices, that is, the user corresponding to the target sub-branch with the largest number of vertices in the corresponding union, is exported as the target group in the social data to be searched, so as to achieve accurate search of related users.

[0053] The group discovery device based on the search condition dense subgraph in a large symbolic graph provided by the present invention is described below. The group discovery device based on the search condition dense subgraph in a large symbolic graph described below and the group discovery method based on the search condition dense subgraph in a large symbolic graph described above can be referred to each other. Figure 3 As shown, the group discovery device based on searching conditional dense subgraphs in a large signed graph provided by the present invention comprises: A data conversion module 310 is used to generate an initial symbolic graph based on the social data to be searched, where the vertices in the initial symbolic graph correspond to users in the social data to be searched, the edges in the initial symbolic graph correspond to the relationships between users, and the symbols of the edges in the initial symbolic graph reflect the relationship categories between users; A total branch generation module 320 is used to sort the vertices in the initial symbolic graph to obtain a sorting result, and to generate a total branch corresponding to each vertex in turn according to the sorting result, wherein the total branch includes a mandatory set and a candidate set, wherein the mandatory set included in the total branch corresponding to the target vertex includes the target vertex, and the candidate set included in the total branch corresponding to the target vertex includes an associated vertex of the target vertex, wherein the associated vertex of the target vertex is obtained by removing the vertices before the target vertex in the sorting result from the neighbor vertex set of the target vertex; The branch module 330 is used to branch the total branch at least once to obtain a target sub-branch, and the target sub-branch meets the preset conditions. At each branch, the pivot point is determined based on the number of second symbol neighbors of the vertex in the union of the candidate set and the required set of the parent branch, and the sub-branch is generated based on the pivot point. The preset condition is that any two vertices in the symbol subgraph corresponding to the union of the target sub-branch are connected to each other and the number of first symbol neighbors of any vertex is not less than , is the default value, is the number of vertices in the subgraph corresponding to the target sub-branch; The target group determination module 340 is used to determine the target group in the social data to be searched based on the target sub-branch with the largest number of vertices.

[0054] Figure 4 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 4 As shown, the electronic device may include: a processor 410, a communications interface 420, a memory 430 and a communication bus 440, wherein the processor 410, the communications interface 420 and the memory 430 communicate with each other through the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute a group discovery method based on searching for conditional dense subgraphs in a large symbolic graph, the method comprising: generating an initial symbolic graph based on the social data to be searched, the vertices in the initial symbolic graph correspond to users in the social data to be searched, the edges in the initial symbolic graph correspond to the relationships between users, and the symbols of the edges in the initial symbolic graph reflect the categories of the relationships between users; sorting the vertices in the initial symbolic graph to obtain the sorting results, and sequentially generating the total branches corresponding to each vertex according to the sorting results, the total branches including the mandatory set and the candidate set, and the mandatory set included in the total branch corresponding to the target vertex The target vertex is included in the candidate set included in the total branch corresponding to the target vertex, and the associated vertex of the target vertex is obtained by removing the vertices before the target vertex in the sorting result from the neighbor vertex set of the target vertex; the total branch is branched at least once to obtain a target sub-branch, and the target sub-branch meets the preset conditions. At each branch, the pivot point is determined based on the number of second symbolic neighbors of the vertex in the union of the candidate set and the required set of the parent branch, and the sub-branch is generated based on the pivot point. The preset condition is that any two vertices in the symbolic subgraph corresponding to the union of the target sub-branch are connected to each other and the number of first symbolic neighbors of any vertex is not less than , is the default value, is the number of vertices in the subgraph corresponding to the target sub-branch; based on the target sub-branch with the largest number of vertices, the target group in the social data to be searched is determined.

[0055] In addition, the logic instructions in the above-mentioned memory 430 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0056] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the group discovery method based on searching conditional dense subgraphs in a large symbolic graph provided by the above-mentioned methods, and the method includes: generating an initial symbolic graph based on social data to be searched, the vertices in the initial symbolic graph correspond to users in the social data to be searched, the edges in the initial symbolic graph correspond to the relationships between users, and the symbols of the edges in the initial symbolic graph reflect the relationship categories between users; sorting the vertices in the initial symbolic graph to obtain the sorting results, and generating the total branches corresponding to each vertex in turn according to the sorting results. , the total branch includes a mandatory set and a candidate set, the mandatory set included in the total branch corresponding to the target vertex includes the target vertex, the candidate set included in the total branch corresponding to the target vertex includes the associated vertex of the target vertex, and the associated vertex of the target vertex is obtained by removing the vertices before the target vertex in the sorting result from the neighbor vertex set of the target vertex; the total branch is branched at least once to obtain a target sub-branch, and the target sub-branch meets the preset conditions. At each branch, the pivot point is determined based on the number of second symbolic neighbors of the vertex in the union of the candidate set and the mandatory set of the parent branch, and a sub-branch is generated based on the pivot point. The preset condition is that any two vertices in the symbolic subgraph corresponding to the union of the target sub-branch are connected to each other and the number of first symbolic neighbors of any vertex is not less than , is the default value, is the number of vertices in the subgraph corresponding to the target sub-branch; based on the target sub-branch with the largest number of vertices, the target group in the social data to be searched is determined.

[0057] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the group discovery method based on searching conditional dense subgraphs in a large symbolic graph provided by the above-mentioned methods, the method comprising: generating an initial symbolic graph based on social data to be searched, the vertices in the initial symbolic graph corresponding to users in the social data to be searched, the edges in the initial symbolic graph corresponding to the relationships between users, and the symbols of the edges in the initial symbolic graph reflecting the relationship categories between users; sorting the vertices in the initial symbolic graph to obtain the sorting results, and generating the total branches corresponding to the respective vertices in turn according to the sorting results, the total branches comprising a required set and a candidate set, and the target set. The required set included in the total branch corresponding to the target vertex includes the target vertex, the candidate set included in the total branch corresponding to the target vertex includes the associated vertex of the target vertex, and the associated vertex of the target vertex is obtained by removing the vertices before the target vertex in the sorting result from the neighbor vertex set of the target vertex; the total branch is branched at least once to obtain a target sub-branch, and the target sub-branch meets the preset conditions. At each branch, the pivot point is determined based on the number of second symbol neighbors of the vertex in the union of the candidate set and the required set of the parent branch, and a sub-branch is generated based on the pivot point. The preset condition is that any two vertices in the symbol subgraph corresponding to the union of the target sub-branch are connected to each other and the number of first symbol neighbors of any vertex is not less than , is the default value, is the number of vertices in the subgraph corresponding to the target sub-branch; based on the target sub-branch with the largest number of vertices, the target group in the social data to be searched is determined.

[0058] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0059] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0060] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for discovering groups based on searching for conditionally dense subgraphs in large symbolic graphs, characterized in that: include: Based on the social data to be searched, an initial symbolic graph is generated, wherein vertices in the initial symbolic graph correspond to users in the social data to be searched, edges in the initial symbolic graph correspond to relationships between the users, and symbols of the edges in the initial symbolic graph reflect categories of relationships between the users; Sorting the vertices in the initial symbolic graph to obtain a sorting result, and sequentially generating total branches corresponding to each vertex according to the sorting result, wherein the total branch includes a mandatory set and a candidate set, the mandatory set included in the total branch corresponding to a target vertex includes the target vertex, and the candidate set included in the total branch corresponding to the target vertex includes an associated vertex of the target vertex, and the associated vertex of the target vertex is obtained by removing the vertex before the target vertex in the sorting result from the neighbor vertex set of the target vertex; The total branch is branched at least once to obtain a target sub-branch, and the target sub-branch satisfies a preset condition. At each branch, a pivot point is determined based on the number of second symbolic neighbors of vertices in the union of the candidate set and the required set of the parent branch, and a sub-branch is generated based on the pivot point. The preset condition is that any two vertices in the symbolic subgraph corresponding to the union of the target sub-branch are connected to each other and the number of first symbolic neighbors of any vertex is not less than , is the default value, is the number of vertices in the subgraph corresponding to the target sub-branch; Based on the target sub-branch with the largest number of vertices, a target group in the to-be-searched social data is determined.

2. The method for discovering a group based on searching for conditional dense subgraphs in a large symbolic graph according to claim 1, characterized in that: Before determining the pivot point based on the number of second symbolic neighbors of vertices in the union of the candidate set and the mandatory set of the parent branch, the method includes: Obtaining an upper bound of a first symbolic degree and an upper bound of a second symbolic degree in the union, a first symbolic degree and a total degree of candidate vertices in the union; Whether to delete the candidate vertex from the candidate set is determined based on the upper bound of the first signed degree, the upper bound of the second signed degree, the first signed degree, and the total degree.

3. The method for discovering a group based on searching for conditional dense subgraphs in a large symbolic graph according to claim 2, characterized in that: The obtaining of the upper bound of the first symbol degree and the upper bound of the second symbol degree in the union includes: Obtaining an upper limit of the first symbol degree based on a first preset formula, and obtaining an upper limit of the second symbol degree based on a second preset formula; The first preset formula is: , is the upper bound of the first symbol degree, S is the required set, C is the candidate set, Represents the vertices in the union The first symbol degree of ; The second preset formula is: or ,in represents the upper bound of the second symbol degree, For variables The processing function, express The number of vertices in .

4. The method for discovering a group based on searching for conditional dense subgraphs in a large symbolic graph according to claim 3, characterized in that: The determining whether to delete the candidate vertex from the candidate set based on the upper bound of the first signed degree, the upper bound of the second signed degree, the first signed degree and the total degree comprises: When the candidate vertex satisfies the first condition, the second condition or the third condition, deleting the candidate vertex from the candidate set; The first condition is: , , Representation diagram Midpoint The second symbol degree of ; The second condition is: ,in, Indicates the number of vertices in the target sub-branch with the largest number of vertices currently existing; The third condition is: , Represents the vertices in the union The total number of degrees.

5. The method for discovering a group based on searching for conditional dense subgraphs in a large symbolic graph according to claim 1, characterized in that: The determining of the pivot point based on the number of second symbolic neighbors of vertices in the union of the candidate set and the mandatory set of the parent branch comprises: Determine a target set based on the union set, wherein the target set includes vertices in the union set whose second symbol degree is not less than an upper bound of the second symbol degree; Filtering the target set to obtain a filtered set, wherein the filtered set includes vertices in the target set that have the most second symbol neighbors; Among the vertices of the filtered set, the point with the most second symbol neighbors in the union is selected as the pivot point.

6. The method for discovering a group based on searching for conditional dense subgraphs in a large symbolic graph according to claim 1, characterized in that: The generating of sub-branches based on the pivot point comprises: When the pivot point does not belong to the mandatory set of the parent branch, the pivot point is used as the branch point; when the pivot point belongs to the mandatory set of the parent branch, a second symbol neighbor of the pivot point in the candidate set of the parent branch is used as the branch point; Deleting the branch point from the candidate set of the parent branch, generating the candidate set of the first child branch, and using the mandatory set of the parent branch as the mandatory set of the first child branch; The branch point is added to the mandatory set of the parent branch to obtain the mandatory set of the second child branch, and the neighbor set of the branch point in the mandatory set of the parent branch is used as the candidate set of the second child branch.

7. A group discovery device based on searching for conditional dense subgraphs in large signed graphs, characterized in that: include: A data conversion module, configured to generate an initial symbolic graph based on the social data to be searched, wherein vertices in the initial symbolic graph correspond to users in the social data to be searched, edges in the initial symbolic graph correspond to relationships between the users, and symbols of the edges in the initial symbolic graph reflect categories of relationships between the users; A total branch generation module is used to sort the vertices in the initial symbolic graph to obtain a sorting result, and to generate total branches corresponding to each vertex in sequence according to the sorting result, wherein the total branch includes a mandatory set and a candidate set, the mandatory set included in the total branch corresponding to the target vertex includes the target vertex, and the candidate set included in the total branch corresponding to the target vertex includes an associated vertex of the target vertex, and the associated vertex of the target vertex is obtained by removing the vertex before the target vertex in the sorting result from the neighbor vertex set of the target vertex; A branching module is used to branch the total branch at least once to obtain a target sub-branch, and the target sub-branch meets a preset condition. At each branch, a pivot point is determined based on the number of second symbol neighbors of the vertices in the union of the candidate set and the required set of the parent branch, and a sub-branch is generated based on the pivot point. The preset condition is that any two vertices in the symbol subgraph corresponding to the union of the target sub-branch are connected to each other and the number of first symbol neighbors of any vertex is not less than , is the default value, is the number of vertices in the subgraph corresponding to the target sub-branch; The target group determination module is used to determine the target group in the social data to be searched based on the target sub-branch with the largest number of vertices.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the method for discovering a group based on searching for conditional dense subgraphs in a large symbolic graph is implemented as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for discovering a group based on searching for conditional dense subgraphs in a large symbolic graph is implemented as claimed in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for discovering a group based on searching for conditional dense subgraphs in a large symbolic graph is implemented as claimed in any one of claims 1 to 6.

Images