Community structure-based symbol network propagation control method, apparatus and device

Abstract

The invention provides a communication control method, device and equipment of a symbol network based on a community structure, and relates to the technical field of communication networks. The symbol network is divided to obtain a plurality of communities, the polar relation between the community structure and each node is considered, the influence score of each node is calculated, then the seed node of each community is determined, the influence factor of each community is calculated according to the connection relation of each node, and then the activation probability of each seed node is determined. Therefore, the information propagation model is improved by considering the activation probability of the polar relationship between the community structure and each node, accurate evaluation of the net positive influence of the seed node is realized, and the accuracy of the node influence in the symbol network is improved.

Description

Technical Field

[0001] The present invention relates to the field of communication network technology, and in particular to a method, device and equipment for controlling propagation of a symbolic network based on a community structure. Background Art

[0002] The rise of the Internet and the booming development of social networks have completely changed the way information is disseminated and how people interact. Online social networking platforms, such as Facebook, Twitter, Instagram, and LinkedIn, have become powerful tools for connecting people, sharing insights, and promoting products.

[0003] In social networks, the problem of maximizing influence can be summarized as: how to select a set of starting nodes (seed set) so that their information, opinions or influence can be spread to the greatest extent in the entire network, that is, the number of people ultimately affected in the network is the largest.

[0004] At present, most of the research on the above issues is conducted in unsigned networks, ignoring the polarity information between users. The polarity information between users includes positive relationships (friends, trust, likes) and negative relationships (enemies, distrust, likes). Users with positive relationships will exert positive influence during the information dissemination process, while users with negative relationships will exert negative influence during the information dissemination. These polarity information plays a vital role in real-world social networks. The current analysis method of unsigned networks does not consider the impact of user relationships, resulting in inaccurate influence assessment. Summary of the invention

[0005] The present invention provides a method, device and equipment for controlling the propagation of a symbolic network based on a community structure, which can improve the accuracy of the influence of nodes in the symbolic network.

[0006] In a first aspect, the present invention provides a method for controlling the spread of a symbolic network based on a community structure, the method comprising: obtaining a connection relationship between nodes in the symbolic network and a polarity relationship between nodes; based on the connection relationship between nodes, using to perform community division to obtain multiple communities; screening bridge nodes of multiple communities, and based on the bridge nodes of each community and the polarity relationship between nodes, calculating the influence score of each node in each community; based on the influence score of each node in each community, determining the seed node of each community; based on the connection relationship between nodes in each community, calculating the influence factor of each community seed node; based on the influence factor of each community seed node, calculating the activation probability of each seed node, the activation probability including the activation probability within the community and the activation probability between communities; based on the activation probability of each seed node and an information propagation model, determining the net positive influence of various sub-nodes in each community, the activation probability being a model parameter of the information propagation model.

[0007] In a possible implementation, the influence factors include a positive behavior score, a negative behavior score, a community density, and a connection strength between communities; based on the connection relationship between each node in each community, the influence factors of each community seed node are calculated, including: based on the connection relationship between each node, the positive behavior score and the negative behavior score of each seed node are determined; based on the number of edges of each node in each community and the number of nodes in each community, the community density of each community is calculated; based on the number of nodes in each community and the connection relationship between each community, the connection strength between two adjacent communities in each community is determined.

[0008] In a possible implementation, the activation probability of each seed node is calculated based on the influence factor of each community seed node, including: determining the intra-community activation probability of various sub-nodes in each community based on the positive behavior score and the negative behavior score of each seed node and the community density of each community; determining the inter-community activation probability of various sub-nodes between two adjacent communities based on the positive behavior score and the negative behavior score of each seed node and the connection strength between two adjacent communities in each community;

[0009] In a possible implementation, the seed nodes of each community are determined based on the influence scores of the nodes in each community, including: sorting the nodes in each community in descending order based on the influence scores to obtain a node sequence of each community; and determining the first K nodes in the node sequence of each community as the seed nodes of each community.

[0010] In a possible implementation, screening bridge nodes of multiple communities includes: determining the number of connections between each node in each community and other communities based on the connection relationship between each node in the multiple communities; determining the positive out-degree of each node in each community based on the number of connections between each node in each community and other communities, and the positive and negative relationships between each node and other communities, where the positive out-degree is used to characterize the number of connections with positive polarity relationships to other communities; determining the bridge node score of each node based on the number of connections between each node in each community and other communities, and the positive out-degree of each node; and determining the node with the largest bridge node score in each community as the bridge node.

[0011] In a possible implementation, after determining the net positive influence of various child nodes in each community based on the activation probability of each seed node and the information propagation model, it also includes: determining the nodes to be propagated based on the net positive influence of various child nodes; transmitting the information to be propagated to the nodes to be propagated, so as to realize information propagation control in the symbolic network.

[0012] In a possible implementation, based on the connection relationship between each node, a community is divided to obtain multiple communities, including: taking each node in the symbolic network as a separate community, dividing to obtain multiple initial communities; calculating the initial modularity of the multiple initial communities after division; fusing adjacent communities in the multiple initial communities, and calculating the modularity after fusion, and iteratively aggregating based on the modularity after fusion to obtain multiple communities.

[0013] In a possible implementation, adjacent communities in multiple initial communities are merged, and the modularity after fusion is calculated. Based on the modularity after fusion, iterative aggregation is performed to obtain multiple communities, including: step 1, merging the target node into the adjacent community and calculating the modularity after fusion; the target node is any node among the multiple nodes; step 2, calculating the difference between the modularity after fusion and the initial modularity; step 3, if the difference is positive, determining to merge the target node into the adjacent community to form multiple fused communities; step 4, if the difference is negative, moving the target node out of the adjacent community and replacing the target node, repeating steps 1 to 4 until multiple nodes are merged; step 5, taking multiple fused communities as multiple initial communities, repeating steps 1 to 5, and calculating the modularity of multiple fused communities until the modularity remains unchanged. Step 6, determining the multiple fused communities obtained by iteration as multiple communities.

[0014] In a second aspect, an embodiment of the present invention provides a propagation control device for a symbolic network based on a community structure, the propagation control device comprising: a communication module for obtaining a connection relationship between each node in the symbolic network and a polarity relationship between each node; a processing module for performing community division based on the connection relationship between each node to obtain multiple communities; screening bridge nodes of multiple communities, and calculating the influence score of each node in each community based on the bridge node of each community and the polarity relationship between each node; determining the seed node of each community based on the influence score of each node in each community; calculating the influence factor of each community seed node based on the connection relationship between each node in each community; calculating the activation probability of each seed node based on the influence factor of each community seed node, the activation probability including the activation probability within the community and the activation probability between communities; determining the net positive influence of various sub-nodes in each community based on the activation probability of each seed node and an information propagation model, the activation probability being a model parameter of the information propagation model.

[0015] In a third aspect, an embodiment of the present invention provides an electronic device, comprising a memory and a processor, the memory storing a computer program, the processor being used to call and run the computer program stored in the memory to perform the steps of the method described in the first aspect and any possible implementation method of the first aspect.

[0016] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and wherein when the computer program is executed by a processor, the steps of the method described in the first aspect and any possible implementation method of the first aspect are implemented.

[0017] The present invention provides a method, device and equipment for controlling the spread of a symbolic network based on a community structure. The present invention divides the symbolic network to obtain multiple communities, considers the community structure and the polarity relationship of each node, calculates the influence score of each node, and then determines the seed node of each community. The influence factor of each community is calculated based on the connection relationship of each node, and then the activation probability of each seed node is determined. In this way, the information propagation model is improved by considering the community structure and the activation probability of the polarity relationship of each node, and the net positive influence of the seed node is accurately evaluated, thereby improving the accuracy of the node influence in the symbolic network. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0019] Figure 1 It is a flowchart of a method for controlling propagation of a symbolic network based on a community structure provided by an embodiment of the present invention;

[0020] Figure 2 It is a structural schematic diagram of a communication control device of a symbolic network based on a community structure provided by an embodiment of the present invention;

[0021] Figure 3 It is a structural schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0022] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0023] In the description of the present invention, unless otherwise specified, " / " means "or", for example, A / B can mean A or B. "And / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, "at least one" and "plurality" refer to two or more. The words "first", "second", etc. do not limit the quantity and execution order, and the words "first", "second", etc. do not limit them to be different.

[0024] In the embodiments of the present application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a concrete way for easy understanding.

[0025] In addition, the terms "including" and "having" and any variations thereof mentioned in the description of this application are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or modules is not limited to the listed steps or modules, but may optionally include other steps or modules that are not listed, or may optionally include other steps or modules that are inherent to these processes, methods, products or devices.

[0026] In order to make the purpose, technical solutions and advantages of the present invention more clear, the following will be described through specific embodiments in conjunction with the accompanying drawings of the present invention.

[0027] As described in the background technology, existing influence maximization algorithms are usually based on independent cascade models or linear threshold models, ignoring the node polarity relationship in the symbolic network. In symbolic networks, positive and negative relationships are crucial to the dissemination results, and directly using traditional models cannot measure the impact of negative relationships on the dissemination of influence. In addition, since users in real social networks have their own social circles. A social circle refers to the community structure in a social network, that is, a group of users have strong interactions with each other, forming a relatively close sub-network. Users in each social circle usually have similar interests, backgrounds or relationships, which leads to the rapid spread of information within these social circles. However, the connections between social circles are relatively weak, which means that information usually encounters obstacles when it spreads across communities. Failure to consider these factors will also lead to inaccurate assessments of influence.

[0028] Net positive influence definition: For each node v∈V in a signed network, there is a positive influence function and a negative impact function When node v is activated, its outward positive influence is given by the positive influence function Determine, the negative influence of outward propagation is determined by the negative influence function The net positive influence of outward propagation is determined by f(v) = f + (v)-f - (v) Decision.

[0029] The net positive influence maximization problem is defined as follows: Given a signed network G(V,E,A), where V represents the node set, corresponding to the user group in the signed network, E represents the set of edges between nodes, corresponding to the relationship between users in the signed network, and A is an adjacency matrix, in which each element represents the symbolic label of each edge, and the edge value can take two cases: 1 and -1, corresponding to positive polarity relationship and negative polarity relationship respectively. Given an information propagation model and a non-negative integer K, the net positive influence maximization problem is to find a set S containing K seed nodes that maximizes the net positive influence, that is,

[0030]

[0031] This paper proposes a bridge node and polarity-based method to solve the problem of maximizing net positive influence in a signed network. This method takes into account the communities in which social network users are located and the relationships between users, and aims to maximize the spread of positive information and minimize the spread of negative information in social networks by accurately selecting seed users. This paper improves the propagation model and combines multiple factors such as the polarity relationship of nodes, community structure, and community density to effectively improve the time efficiency of influence evaluation.

[0032] In order to explore the hidden structures and relationships in the network, the present invention divides the network through a community detection algorithm to extract the community structure in the network. Then, by analyzing the polarity of the node and its shortest path to the community bridge node, a new influence evaluation index is proposed, and the top K nodes are selected as the optimal seed set. Then, the traditional independent cascade propagation model is improved by combining the community affiliation, polarity relationship and community density of the node to better simulate the information propagation in the symbolic network. Finally, the improved independent cascade model is used to measure the propagation range of the net positive influence.

[0033] like Figure 1 As shown, the embodiment of the present invention provides a method for controlling the spread of a symbolic network based on a community structure. The method for controlling the spread includes steps S101-S105.

[0034] S101: Acquire the connection relationship between nodes in the symbolic network and the polarity relationship between nodes.

[0035] S102: Based on the connection relationship between each node, use to divide the community to obtain multiple communities.

[0036] It should be noted that when performing community division in a symbolic network, the key goal is to divide the nodes in the network into several closely connected small groups, namely communities, based on their connection relationships. The core principle of community division is to maximize the connection density between nodes within the community while minimizing the connection between communities. Specifically, the principle of community division can be summarized as follows: in the same community, the connection between nodes should be as close as possible so that information can be efficiently transmitted within the community; while the node connections between different communities should be relatively small, thereby limiting the scope and speed of information transmission across communities.

[0037] An important division principle is to maximize the modularity of the community. Modularity is one of the criteria for measuring the quality of community division, indicating whether the node connections within the community are tight enough and whether the connections between communities are sparse enough. By maximizing modularity, it is ensured that after the community division, the nodes within the community maintain a high degree of interconnectivity, while the node connections between communities are relatively few.

[0038] Exemplarily, the embodiments of the present invention can perform community detection, and divide the network into several communities through the community detection algorithm to discover hidden structures and relationships. The Louvain algorithm is used to divide the symbolic network into multiple communities. The purpose of community division is to better understand the role of nodes in the local network and the impact of community structure on information dissemination when information is propagated.

[0039] It should be noted that the community characteristics after the community is divided are that each community shows high internal connectivity, that is, the nodes in the community have relatively close connections with each other. Such a community has a high internal connection density, which can ensure that information, especially positive information, can spread quickly within the community. The multi-path propagation effect of information is particularly obvious in high-density communities, and the positive relationship between nodes increases the spread of positive influence.

[0040] The connections between different communities are relatively sparse, and this low external connectivity means that cross-community information dissemination is more difficult. The key nodes for cross-community dissemination are so-called "bridge nodes". These bridge nodes are usually located at the boundaries of communities and are responsible for connecting a small number of nodes in different communities. Since cross-community information dissemination mainly depends on these bridge nodes, controlling the positive and negative polarity relationship of bridge nodes is crucial to the control of positive and negative information in the information dissemination process. In the process of positive information dissemination, bridge nodes can increase the spread of positive influence. In the dissemination of negative information, the role of bridge nodes is equally important. By reasonably selecting seed nodes, the spread of negative information can be effectively reduced.

[0041] As a possible implementation manner, step S102 may be specifically implemented as S1021-S1023.

[0042] S1021, taking each node in the symbolic network as a separate community, dividing it to obtain multiple initial communities;

[0043] S1022, calculating the initial modularity of multiple initial communities after division;

[0044] S1023, adjacent communities in the multiple initial communities are merged, and the modularity after merging is calculated. Based on the modularity after merging, iterative aggregation is performed to obtain multiple communities.

[0045] Exemplarily, step S1023 may be specifically implemented as steps one to six.

[0046] Step 1: merge the target node into the adjacent community and calculate the modularity after fusion; the target node is any node among multiple nodes;

[0047] Step 2: Calculate the difference between the fused modularity and the initial modularity;

[0048] Step 3: If the difference is positive, it is determined that the target node is merged into the adjacent community to form multiple fusion communities;

[0049] Step 4: If the difference is negative, move the target node out of the adjacent community and replace the target node, repeating steps 1 to 4 until multiple nodes are merged;

[0050] Step 5: Using multiple fusion communities as multiple initial communities, repeating steps 1 to 5, and calculating the modularity of multiple fusion communities until the modularity remains unchanged;

[0051] Step six, determining the multiple fusion communities obtained through iteration as multiple communities.

[0052] S103: Filter bridge nodes of multiple communities, and calculate the influence score of each node in each community based on the bridge nodes of each community and the polarity relationship between each node.

[0053] In the embodiment of the present application, the bridge node is a node that connects the most communities and has the most positive relationships with nodes in the community.

[0054] As a possible implementation manner, the embodiment of the present invention can screen bridge nodes of multiple communities through steps S1031-S1034.

[0055] S1031. Determine the number of connections between each node in each community and other communities based on the connection relationship between each node in multiple communities.

[0056] For example, the first condition for a bridge node is that it needs to be connected to nodes in multiple different communities. i , calculate the number of different communities it is connected to, the calculation formula is as follows.

[0057] C(v i )=|{C(v j )|v j ∈Successors(v i )}|;

[0058] Among them, C(v j ) represents node v j The community to which you belong. i ) is the node v j The community you belong to. i ) is the node v i The set of out-degree neighbor nodes.

[0059] S1032: Determine the positive out-degree of each node in each community based on the number of connections between each node in each community and other communities, and the positive and negative relationships between each node and other communities.

[0060] In some embodiments, positive outdegree is used to characterize the number of connections that have positive polarity relationships to other communities.

[0061] For example, for each node v i , calculate the positive out-degree with positive polarity relationship, the calculation formula is as follows.

[0062]

[0063] Among them, I(A i,j =1) is the indicator function. i to v j When the edge is a positive relationship, the value is 1, otherwise it is 0. i ) is the node v i The node set pointed to, For node v i The positive outgoing value, Av i ,v j For node v i and node v j The value corresponding to the connection relationship between them.

[0064] S1033: Determine the bridge node score of each node based on the number of connections between each node in each community and other communities and the positive out-degree of each node.

[0065] For example, each node vi The score of is the product of its positive outdegree and the number of different communities it connects to. The calculation formula is as follows.

[0066]

[0067] in, For node v i The bridge node fraction, For node v i The number of different communities connected, For node v i The positive outgoing degree of .

[0068] S1034. Determine the node with the largest bridge node score in each community as a bridge node.

[0069] For example, within each community, select the i The node with the highest value serves as the bridge node of the community.

[0070] It should be noted that the embodiment of the present invention can propose a new influence evaluation index by analyzing the polarity relationship of the node and its shortest path to the community bridge node, and select the top K nodes as the optimal seed set. First, in each community, a node that is connected to more different communities and has more positive relationships with the connected communities is selected as a bridge node. Then, the shortest path from each node to the community bridge node is calculated within each community to measure the role of the node in information dissemination. At the same time, the three-degree separation theory of social network information dissemination is taken into account. Finally, the formula for calculating the influence score of the node within the community is as follows.

[0071]

[0072] Among them, BPS is the node influence score, is the number of positive ties between node u and its neighbors, is the number of negative ties between node u and its neighbors, v is the bridge node of the community, s u,v is the shortest path length from node u to bridge node v.

[0073] S104. Determine the seed node of each community based on the influence score of each node in each community.

[0074] As a possible implementation manner, step S104 may be specifically implemented as steps S1041 - S1042 .

[0075] S1041. Based on the influence scores, sort the nodes in each community in descending order to obtain a node sequence for each community.

[0076] S1042: Determine the first K nodes in the node sequence of each community as the seed nodes of each community.

[0077] S105. Calculate the influence factor of each community seed node based on the connection relationship between each node in each community.

[0078] In some embodiments, the influence factors include positive behavior scores, negative behavior scores, community density, and inter-community connection strength.

[0079] As a possible implementation manner, step S105 may be specifically implemented as steps S1051 - S1053 .

[0080] S1051. Determine the positive behavior score and negative behavior score of each seed node based on the connection relationship of each node.

[0081] In some embodiments, the positive behavior score is used to evaluate the degree of positive relationship between evaluation nodes. When the relationship between two nodes is positive, the indicator function takes a value of 1, otherwise it takes a value of 0.

[0082] Exemplarily, the embodiment of the present invention can determine the positive behavior score by the following formula.

[0083]

[0084] Among them, PB i,j is the positive behavior score between node i and node j, I(A i,m =1) is an indicator function, which takes the value 1 when the edge relationship between node i and node m is positive, otherwise it takes the value 0. n,j =1) is an indicator function, which takes the value of 1 when the edge relationship between node n and node j is positive, otherwise it takes the value of 0. m is the out-degree neighbor of node i, n is the in-degree neighbor of node j, out_neighbors(i) is the out-degree neighbor set of node i, and in_neighbors(j) is the in-degree neighbor set of node j.

[0085] In a symbolic social network, take the number of positive behavior scores between node A and its neighbor node B as an example. When node A has multiple neighbor nodes that receive its positive behavior, its influence will be dispersed among these nodes, which results in a relative decrease in the influence received by a single neighbor node B from node A. In addition, if node B also receives positive behaviors from multiple of its neighbor nodes, the relative influence of node A on B will be further reduced. Therefore, the combination of these two dimensions jointly determines the potential for the spread of positive behaviors between nodes.

[0086] In some embodiments, the negative behavior score is used to evaluate the degree of negative relationship between network nodes. When the relationship between nodes is negative, the indicator function takes a value of 1, otherwise it takes a value of 0.

[0087] Exemplarily, the embodiment of the present invention may determine the negative behavior score by the following formula.

[0088]

[0089] Among them, NB i,j is the negative behavior score between node i and node j, I(A i,m =-1) is an indicator function, which takes the value 1 when the edge relationship between node i and node m is negative, otherwise it takes the value 0. n,j =-1) is an indicator function, which takes the value of 1 when the edge relationship between node i and node m is negative, otherwise it is 0.

[0090] S1052: Calculate the community density of each community based on the number of edges of each node in each community and the number of nodes in each community.

[0091] In some embodiments, community density is an indicator used to measure the degree of interconnection between nodes in a specific community, and reveals the connectivity of the community structure by measuring the ratio of the actual number of connections to the possible number of connections in the community.

[0092] Exemplarily, the calculation formula for the community density of the community where node i is located is as follows.

[0093]

[0094] Among them, CD i is the community density of the community where node i is located, is the number of edges in the community where node i is located, comm i is the number of nodes in the community where node i is located.

[0095] In a high-density community, each node is connected to most other nodes in the community, and information can spread quickly through multiple paths. Because the paths are shorter and each node can affect multiple other nodes after activation, the speed of information propagation in a high-density community is significantly faster.

[0096] S1053: Based on the number of nodes in each community and the connection relationship between each community, determine the connection strength between two adjacent communities in each community.

[0097] In some embodiments, connection strength is an indicator that measures the ease with which information can be transmitted between different communities, by analyzing the number of direct connections between communities to assess the ease with which information can flow from one community to another.

[0098] Exemplarily, the calculation formula for the connection strength between two communities is as follows.

[0099]

[0100] Among them, CS A,B is the connection strength between community A and community B, e i,j represents the edge between node i and node j, A is community A, B is community B, E is the edge set between network nodes, |comm A | is the number of nodes in community A, |comm B | is the number of nodes in community B.

[0101] Strong connections between communities mean that there are more nodes connected between the two communities, which increases the chances of inter-community communication. Even if a node fails to successfully spread information, other nodes connected to external communities may still successfully activate neighbors, thereby increasing the success rate of cross-community communication.

[0102] S106. Based on the influence factor of each community seed node, the activation probability of each seed node is calculated.

[0103] In the embodiment of the present application, the activation probability includes the intra-community activation probability and the inter-community activation probability.

[0104] As a possible implementation manner, step S106 may be specifically implemented as steps S1061 - S1062 .

[0105] S1061. Based on the positive behavior score and the negative behavior score of each seed node and the community density of each community, determine the intra-community activation probability of various sub-nodes of each community.

[0106] In some embodiments, the probability of activation within a community reflects the possibility of information propagation between nodes in the same community, taking into account the positive and negative relationships between nodes and the density of the community. Higher positive behaviors and community density increase the possibility of propagation.

[0107] Exemplarily, the embodiment of the present invention can calculate the activation probability within the community by the following formula.

[0108]

[0109] Among them, P i.j is the activation probability between node i and node j in the community, CD i is the community density of the community where node i is located, PB i,j is the positive behavior score between node i and node j, NB i,j is the negative behavior score between node i and node j, Ai,j is the value of the connection relationship between node i and node j.

[0110] In a highly connected community, information is propagated through multiple paths, allowing it to quickly cover most nodes in the community. In addition, the positive and negative (negative) behavior measurements between nodes can potentially affect the propagation path lengths of positive and negative influences.

[0111] S1062: Based on the positive behavior score and the negative behavior score of each seed node and the connection strength between two adjacent communities in each community, determine the inter-community activation probability of various sub-nodes between two adjacent communities.

[0112] In some embodiments, the inter-community activation probability is used to evaluate the potential for information dissemination between nodes across communities. The model considers both the strength of connections between communities and the positive and negative relationships between nodes. Stronger connections and positive behaviors enhance the efficiency of cross-community dissemination.

[0113] Exemplarily, the embodiment of the present invention can calculate the inter-community activation probability by the following formula.

[0114]

[0115] Among them, P i.j is the activation probability between nodes i and j in the community, CS A,B is the connection strength between community A and community B, PB i,j is the positive behavior score between node i and node j, NB i,j is the negative behavior score between node i and node j, A i,j is the value of the connection relationship between node i and node j.

[0116] When the strength of inter-community connections is high, the chances of inter-community communication are increased. Even if a node fails to successfully propagate information, other nodes connected to external communities may still successfully activate neighbors, thereby increasing the success rate of cross-community communication. When the positive (negative) behavior measurement value of cross-border nodes is high, the probability of successful information propagation will also increase.

[0117] S107. Based on the activation probability of each seed node and the information propagation model, determine the net positive influence of various child nodes in each community, and the activation probability is a model parameter of the information propagation model.

[0118] In some embodiments, the influence score of each node, that is, the net positive influence of the user. Using a real social network as an experimental data set, the seed user is brought into the influence propagation model for propagation, thereby obtaining the positive influence of the seed user v that is propagated outward and negative influence The positive influence is calculated by the number of other users activated by the seed user to the positive activation state, and the negative influence is calculated by the number of nodes activated by the seed user to the negative activation state, thereby achieving the net positive influence measurement of the seed user f(v) = f + (v)-f - (v).

[0119] The embodiment of the present invention introduces a community structure and makes innovative improvements to the traditional activation probability design. This innovative design is based on two motivations: improving the propagation efficiency and controlling the boundary of information diffusion.

[0120] The information propagation model increases the number of nodes in the actively activated state by efficiently propagating positive information within the community. In the community, the connection density between nodes is high, the propagation path is short, and information can be propagated in parallel through multiple paths, allowing more nodes to quickly enter the actively activated state. This design maximizes the propagation effect of positive information, thereby increasing the number of positively activated nodes and improving the net positive influence.

[0121] In cross-community information dissemination, when the connection strength between communities is high, the chance of inter-community dissemination is increased. Even if a node fails to successfully disseminate information, other nodes connected to external communities may still successfully activate neighbors, thereby increasing the success rate of cross-community dissemination. On the contrary, when the connection between communities is weak, information diffusion may be restricted within the community boundary, forming a natural dissemination resistance and reducing the scope of information diffusion in the entire network. Through this boundary control mechanism, the model not only effectively improves the efficiency of information dissemination within the community, but also limits unnecessary information diffusion, ensures the controllability of dissemination, and further enhances the dissemination coverage of the entire network.

[0122] The embodiment of the present invention extends the IC model and proposes an independent cascade model that combines polarity and community topology. In this model, in order to more accurately simulate the information propagation mechanism in symbolic social networks, not only the polarity relationship between nodes is considered, but also the important influence of community structure on information propagation is particularly highlighted.

[0123] The information propagation process of this model is as follows: Information propagation starts by setting the nodes in the seed node set S to an active activation state, and the other nodes in the network are set to an inactive state. At the initial moment of propagation, each node in the seed node set S has only one chance to try to activate its inactive neighbor node with probability P. The result of the activation attempt is not only affected by the state of the activated node itself, but also closely related to the polarity relationship between its connected nodes. Specifically, when a node u is activated by its neighbor node v at time t-1, it will decide whether to transform itself into an active or passive state according to the following formula at time t, and try to activate its inactive neighbor node. Thereafter, the above process is continuously iterated until no new nodes are activated. For the newly activated node u, the calculation formula of its state State(u) is as follows.

[0124] State(u)=State(v)×A v,u ;

[0125] Among them, A v,u Represents edge e v,u The polarity relationship between nodes is 1 and -1, respectively, indicating a positive relationship and a negative relationship between nodes. State(u) is the state of node u, and State(v) is the state of node v.

[0126] When estimating the information activation probability P between nodes, the model comprehensively considers the positive and negative polarity behaviors between nodes, the community density of the community to which the node belongs, and the connection strength between different communities.

[0127] The present invention provides a propagation control method for a symbolic network based on a community structure. The symbolic network is divided to obtain a plurality of communities. The community structure and the polarity relationship of each node are considered, and the influence score of each node is calculated, and then the seed node of each community is determined. The influence factor of each community is calculated based on the connection relationship of each node, and then the activation probability of each seed node is determined. Thus, the information propagation model is improved by considering the community structure and the activation probability of the polarity relationship of each node, and the net positive influence of the seed node is accurately evaluated, thereby improving the accuracy of the node influence in the symbolic network.

[0128] Optionally, the method for controlling propagation of a symbolic network based on a community structure provided by an embodiment of the present invention further includes steps S201 - S202 after step S107 .

[0129] S201. Determine the nodes to be propagated based on the net positive influence of various child nodes.

[0130] S202: Transmit the information to be propagated to the node to be propagated, so as to realize information propagation control in the symbolic network.

[0131] The embodiment of the present invention proposes a method based on bridge nodes and polarity: The present invention adopts a node influence measurement method based on communities, which reduces the overall computational complexity. By analyzing the polarity relationship of nodes and their distance from bridge nodes, a node influence evaluation method is proposed, which can effectively select a seed node set that can maximize the net positive influence.

[0132] The embodiment of the present invention is based on the improvement of the propagation model of the symbolic network: the present invention proposes an improved propagation model based on the particularity of the positive and negative polarity relationship and the community structure characteristics in the symbolic network. The model designs a reasonable activation probability for the nodes within the community, realizing the rapid propagation of information within the community; at the same time, it takes into account the frequency of communication between the two communities to avoid excessive attenuation or loss of information when propagating across communities, thereby improving the success rate of cross-community propagation.

[0133] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0134] The following is an embodiment of the device of the present invention. For details not described in detail therein, reference may be made to the corresponding method embodiment described above.

[0135] Figure 2 The schematic diagram of the structure of a communication control device for a symbolic network based on a community structure provided by an embodiment of the present invention is shown. The communication control device 300 includes a communication module 301 and a processing module 302 .

[0136] The communication module 301 is used to obtain the connection relationship between each node in the symbolic network;

[0137] The processing module 302 is used to divide the community based on the connection relationship of each node to obtain multiple communities; based on the connection relationship of each node in the multiple communities, screen and obtain the bridge node of each community; the bridge node is the node that connects the most communities and has the most positive relationships with each node in the community; based on the connection relationship of each node in the multiple communities and the bridge node of each community, calculate the influence score of each node in each community; based on the influence score of each node in each community, determine the seed node of each community; based on the seed node of each community, perform information dissemination control.

[0138] Figure 3 Schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. Figure 3As shown, the electronic device 400 includes: a processor 401, a memory 402, and a computer program 403 stored in the memory 402 and executable on the processor 401. When the processor 401 executes the computer program 403, the steps in the above-mentioned method embodiments are implemented, for example Figure 1 Alternatively, when the processor 401 executes the computer program 403, the functions of each module / unit in the above-mentioned device embodiments are realized, for example, Figure 2 The functions of the communication module 301 and the processing module 302 are shown.

[0139] Exemplarily, the computer program 403 may be divided into one or more modules / units, which are stored in the memory 402 and executed by the processor 401 to implement the present invention. The one or more modules / units may be a series of computer program instruction segments that can implement specific functions, and the instruction segments are used to describe the execution process of the computer program 403 in the electronic device 400. For example, the computer program 403 may be divided into Figure 2 The communication module 301 and the processing module 302 are shown.

[0140] The processor 401 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0141] The memory 402 may be an internal storage unit of the electronic device 400, such as a hard disk or memory of the electronic device 400. The memory 402 may also be an external storage device of the electronic device 400, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device 400. Further, the memory 402 may also include both an internal storage unit of the electronic device 400 and an external storage device. The memory 402 is used to store the computer program and other programs and data required by the terminal. The memory 402 may also be used to temporarily store data that has been output or is to be output.

[0142] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such 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, and should all be included in the protection scope of the present invention.

Claims

1. A method for controlling the spread of a symbolic network based on a community structure, characterized in that: include: Obtain the connection relationship of each node in the symbolic network and the polarity relationship between each node; Based on the connection relationship between each node, the community is divided into multiple communities. Screening the bridge nodes of the multiple communities, and calculating the influence score of each node in each community based on the bridge nodes of each community and the polarity relationship between the nodes; Determining the seed node of each community based on the influence score of each node in each community; Based on the connection relationship between each node in each community, calculate the influence factor of each community seed node; Based on the influence factor of each community seed node, the activation probability of each seed node is calculated, and the activation probability includes the activation probability within the community and the activation probability between communities; Based on the activation probability of each seed node and the information propagation model, the net positive influence of various child nodes in each community is determined, and the activation probability is a model parameter of the information propagation model.

2. The method for controlling the propagation of a symbolic network based on a community structure according to claim 1, characterized in that: The influencing factors include positive behavior scores, negative behavior scores, community density, and inter-community connection strength; The calculation of the influence factor of each community seed node based on the connection relationship of each node in each community includes: Based on the connection relationship of each node, determine the positive behavior score and negative behavior score of each seed node; Calculate the community density of each community based on the number of edges connected to each node in each community and the number of nodes in each community; Based on the number of nodes in each community and the connection relationship between each community, the connection strength between two adjacent communities in each community is determined.

3. The method for controlling the propagation of a symbolic network based on a community structure according to claim 2, characterized in that: The activation probability of each seed node is calculated based on the influence factor of each community seed node, including: Based on the positive behavior score and negative behavior score of each seed node and the community density of each community, determine the intra-community activation probability of various child nodes in each community; Based on the positive behavior score and negative behavior score of each seed node and the connection strength between two adjacent communities in each community, the inter-community activation probability of various sub-nodes between two adjacent communities is determined.

4. The method for controlling the propagation of a symbolic network based on a community structure according to claim 1, characterized in that: The step of determining the seed node of each community based on the influence score of each node in each community includes: Based on the influence scores, sort the nodes in each community in descending order to obtain a node sequence for each community; The first K nodes in the node sequence of each community are determined as seed nodes of each community.

5. The method for controlling the propagation of a symbolic network based on a community structure according to claim 1, characterized in that: The screening of the bridge nodes of the plurality of communities includes: Determine the number of connections between each node in each community and other communities based on the connection relationship between each node in the multiple communities; Based on the number of connections between each node in each community and other communities, and the positive and negative relationships between each node and other communities, determine the positive out-degree of each node in each community, where the positive out-degree is used to characterize the number of connections with positive polarity relationships to other communities; Determine a bridge node score of each node based on the number of connections between each node in each community and other communities and the positive out-degree of each node; The node with the largest bridge node score in each community is determined as the bridge node.

6. The method for controlling the spread of a community structure-based symbolic network according to claim 1, characterized in that: After determining the net positive influence of various sub-nodes in each community based on the activation probability of each seed node and the information propagation model, the following is also included: Determine the nodes to be propagated based on the net positive influence of various child nodes; The information to be propagated is transmitted to the node to be propagated, so as to realize the information propagation control in the symbolic network.

7. The method for controlling the propagation of a symbolic network based on a community structure according to claim 1, characterized in that: Based on the connection relationship of each node, the community is divided using to obtain multiple communities, including: Taking each node in the symbolic network as a separate community, dividing the community to obtain a plurality of initial communities; Calculate the initial modularity of multiple initial communities after partitioning; Adjacent communities in the multiple initial communities are merged, and the modularity after merging is calculated. Based on the modularity after merging, iterative aggregation is performed to obtain the multiple communities.

8. The method for controlling the propagation of a symbolic network based on a community structure according to claim 7, characterized in that: The step of fusing adjacent communities in the multiple initial communities, calculating the modularity after fusion, and performing iterative aggregation based on the modularity after fusion to obtain the multiple communities includes: Step 1: merge the target node into the adjacent community and calculate the modularity after fusion; the target node is any node among the multiple nodes; Step 2, calculating the difference between the modularity after fusion and the initial modularity; Step 3: If the difference is positive, it is determined that the target node is merged into the adjacent community to form multiple fusion communities; Step 4: if the difference is negative, the target node is moved out of the adjacent community, and the target node is replaced, and steps 1 to 4 are repeated until the fusion of the multiple nodes is completed; Step 5, taking the multiple fusion communities as multiple initial communities, repeating steps 1 to 5, and calculating the modularity of the multiple fusion communities until the modularity remains unchanged; Step six: determine the multiple fusion communities obtained through iteration as the multiple communities.

9. A communication control device for a symbolic network based on a community structure, characterized in that: include: A communication module, used to obtain the connection relationship of each node in the symbolic network and the polarity relationship between each node; A processing module is used to divide communities based on the connection relationship between each node to obtain multiple communities; Screening the bridge nodes of the multiple communities, and calculating the influence score of each node in each community based on the bridge nodes of each community and the polarity relationship between the nodes; Determining the seed node of each community based on the influence score of each node in each community; Based on the connection relationship between each node in each community, calculate the influence factor of each community seed node; Based on the influence factor of each community seed node, the activation probability of each seed node is calculated, and the activation probability includes the activation probability within the community and the activation probability between communities; Based on the activation probability of each seed node and the information propagation model, the net positive influence of various child nodes in each community is determined, and the activation probability is a model parameter of the information propagation model.

10. An electronic device, characterized in that: The electronic device comprises a memory and a processor, the memory stores a computer program, and the processor is used to call and run the computer program stored in the memory to execute the steps of the method according to any one of claims 1 to 8.

Images