A social network key person hiding method and device based on target k-node decomposition

Abstract

The application relates to a social network key person hiding method and device based on target k node decomposition, and the method comprises the following steps: S1: assuming a non-directional and non-weighted network G, a k core decomposition algorithm is used to assign a core value C to each node in the network (v,G) , and a node with the highest core value in the social network is defined as a key person V k . S2: for each node, level weight information M is obtained, and a backtracking tree G BT is constructed for a target k node to obtain a candidate edge set H. S3: according to the candidate edge set H, an edge e with the most affected nodes is selected to be pruned and deleted. S4: step S3 is repeated until all nodes in the target k node are removed from the k core subgraph, that is, the effect of key person hiding is achieved. The application improves the onion decomposition algorithm and the backtracking tree algorithm, constructs a backtracking tree to obtain the least edge set in the graph which needs to be deleted, makes the specified target node separate from the k core subgraph, and thus realizes the hiding of the key person information in the social network.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data mining and social networks, and particularly relates to a social network key figure concealment method and device based on target k node decomposition. BACKGROUND

[0002] With the development of Internet technology, the network has become an indispensable part of our daily life, and it plays an important role in various complex systems such as social networks, power networks and transportation networks. Key nodes refer to nodes with important influence and connectivity in various networks. These nodes play a core role in the network, so the identification of key nodes is important for the overall function and stability of the network. For example, social networks as an important communication and information exchange platform in modern society, with the continuous growth of the number and scale of social network users, the connection relationship between nodes in the network is becoming increasingly complex. In social networks, key nodes are usually individuals or groups with extensive social relationships and high influence, who can accelerate information diffusion, guide public opinion, and influence the behavior decisions of other users through their demonstration effect. Studies have shown that attacks on key nodes can have serious consequences. How to analyze and understand the important nodes and their mutual relationships in these complex network structures has become an important research topic.

[0003] Due to its simplicity and efficiency, k-core as a key indicator to capture the participation of network structure has attracted widespread attention. The k-core is a maximal induced subgraph in which each node has a degree of at least k. The presence of nodes in the k-core depends on their adjacent relationships in the subgraph. Therefore, removing a portion of nodes or edges from the k-core can cause the separation of other nodes in the k-core. For example, some researchers found that removing a small number of edges can severely damage the structure of the k-core. Some scholars focus on the k-core minimization problem, use the concept of Shapley value in cooperative game theory and propose an efficient algorithm to handle the problem. Although the above methods are effective, they all attack the k-core from a global perspective and pay little attention to the disturbance of the distribution of individual nodes in the k-core. For example, the removal of key nodes in the k-core can significantly damage the network structure. However, the disturbance of removing nodes is not feasible in practice and can cause serious consequences. Therefore, in order to achieve concealment and feasibility, by trying to identify the critical edge, the removal of the edge will trigger the separation of a given target node from the k-core, so that the key node cannot be identified, achieving the effect of concealment.

[0004] Chinese patent CN 114301674 A proposes a method for hiding key figures in social networks based on a degree-preserving attack. This method protects the privacy of social users by reconnecting some links in the original network without changing the node's degree, thereby changing the node's core value across the entire network. However, this method attacks the k-core from a global perspective and does not address the vulnerability of local nodes within the k-core. Summary of the Invention

[0005] The present invention aims to overcome the above-mentioned shortcomings of the prior art and provides a method and device for hiding key figures in a social network based on target k-node decomposition.

[0006] The goal of the present invention is to selectively screen candidate edges for a given network and a target k-node set, so that all target nodes do not exist in the original k-core by removing a minimum set of edges. This problem is defined as the target k-node collapse problem. The method according to the present invention can integrate target node information into the candidate edge screening process, selectively screen candidate edges for the target node, and obtain the minimum budget for removing edges to crash the target k-node, thereby reducing computational complexity. In social networks, the study of the target k-node collapse problem helps to understand node vulnerabilities in the k-core model, allowing maintenance personnel to protect key relationships between users. In addition, it also helps to reduce the importance of confidential users, thereby reducing the attention of potential attackers to them.

[0007] The present invention improves the onion decomposition algorithm and the backtracking tree algorithm to construct a backtracking tree to obtain the minimum set of edges that need to be deleted in the graph, so that the specified target node is separated from the k-core, thereby realizing the concealment of key person information in the network.

[0008] A first aspect of the present invention relates to a method for hiding key figures in a social network based on target k-node decomposition, comprising the following steps:

[0009] S1: Assume an undirected and unweighted network G and use the k-core decomposition algorithm to assign a core value C to each node in the network. (v,G) , define the node with the highest core value in the social network as the key person V k ;

[0010] S2: For each node, obtain its hierarchical weight information M and construct a backtracking tree G for the target k node BT Get the candidate edge set H;

[0011] S3: Select the edge e with the most affected nodes according to the candidate edge set H for pruning and deletion;

[0012] S4: Repeat step S3 until all nodes in the target k node are removed from the k core, that is, the key figure hiding effect is achieved.

[0013] Specifically, in step S1: assuming an undirected and unweighted social network G(V,E), V is the set of nodes in the network, E is the set of links in the network, and a kernel value C is assigned to each node in the network using the k-kernel decomposition algorithm. (v,G) The core value of a node is equivalent to the importance of the node. The higher the core value of a node, the higher the importance of the node. The node with the highest core value in the social network is defined as the key person V k .

[0014] The step S2 specifically includes the following steps:

[0015] S2.1: For each node in the network, find its neighbor nodes N (v,G) and node degree d (v,G) ;

[0016] S2.2: According to theorem, if and only if When removing the edge (u,v) will cause the k nodes to collapse and theorem k=C (u,G) ≤C (v,G) , after removing (u, v), only the nodes with k cores will crash, and the number of cores will be reduced by 1 at most, and the initial candidate edge set P = {(u, v)|(u, v)∈E,min(C (u,G) ,C (v,G) )=k};

[0017] S2.3: Use the improved onion decomposition algorithm to divide the k nodes of graph G into different layers, so as to further simplify the candidate edge set; first calculate the node degree value d of all nodes (v,G) , and sort the node degrees to get the degree order, peel off the nodes from small to large according to the degree order, and divide the nodes into different layers according to how many peelings are required to reach a given node; that is, if the degree of the node of the k-shell in the k-core is exactly k, it belongs to the first local layer; if after removing the first layer, the node with the maximum degree of k belongs to the second local layer, and so on to get the number of layers M for each node;

[0018] S2.4: Based on the number of layers M, prioritize neighbors with higher collapse orders, and use the breadth-first search algorithm to iteratively backtrack neighbors that maintain increasing collapse order relative to the current node; first use the improved onion decomposition algorithm to calculate the layer value of each node and initialize the queue to the target node set; then initialize an empty directed graph G BT , which contains a node set T and an empty edge set; then traverse each node u in the queue, find all nodes v adjacent to u and with a core number of k, and store these nodes in the set N′, then find all nodes v in the set with a layer number less than node u, and store these nodes in the set N * , update the queue, remove u from the queue, and replace all* In but not in the backtracking tree G BT The nodes in G are added to the queue and updated BT , all in N * Add the node v and edge (u,v) in G BT Finally, we get the simplified candidate edge set H = E BT ∪{(u,v)|u∈T,v∈N(u,G k )\V BT )};

[0019] Specifically, step S3 includes: traversing each edge e in the candidate edge set H, and pruning edge e according to the pruning rule; first, removing edge e from the backtracking tree to obtain a new subgraph; at the same time, initializing the pruned follower set to the intersection of the follower set of edge e and the target node set T; then finding all the node sets Z in the target node set T that exclude zero-indegree nodes, removing node set Z from them, and adding node set Z to the pruned follower set to obtain the pruned follower set; selecting the edge e with the most followers after pruning. * , e * Add to the set of removed edges, and finally get the set of removed edges.

[0020] A second aspect of the present invention relates to a device for hiding key figures in a social network based on target k-node decomposition, comprising a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, they are used to implement a method for hiding key figures in a social network based on target k-node decomposition of the present invention.

[0021] A third aspect of the present invention relates to a computer-readable storage medium having a program stored thereon. When the program is executed by a processor, the method for hiding key figures in a social network based on target k-node decomposition of the present invention is implemented.

[0022] The innovation of this invention lies in proposing a method for hiding key figures in social networks based on target k-node decomposition. This method can remove the minimum set of edges in the graph, separating the specified target node from the k-core, thereby effectively avoiding the risk of user privacy leakage. The main innovations are as follows:

[0023] (1) An improved onion decomposition algorithm is proposed, and the judgment condition of node stripping is modified, so that it can perform a more detailed hierarchical division of k-core nodes. The present invention proposes an improved onion decomposition algorithm and applies it to the field of anonymization of key figures in social networks. Compared with the onion decomposition algorithm, the improved onion decomposition algorithm modifies the judgment condition of node stripping, so that it can perform a more detailed hierarchical division of k-core nodes, until a hierarchical weight information M is returned for all k nodes, thereby providing a basis for subsequent edge set screening and reduction. This method performs a more detailed hierarchical division of k-core nodes and can more effectively deal with the k-node stratification problem in social networks.

[0024] (2) A backtracking tree algorithm is proposed to screen the candidate edges of the target node, which greatly reduces the number of candidate edges. The present invention screens the candidate edges of the target node based on the node level weight obtained by the improved onion decomposition algorithm, and a backtracking tree algorithm is proposed for this purpose. The algorithm takes the target node as the root node, uses breadth-first search to traverse all neighbor nodes whose level weight is less than the current root node, and adds the traversed nodes to the backtracking tree G in order. bt In, G bt All the edges included in are the candidate edges considered to crash the target node. This method greatly reduces the number of candidate edges and significantly reduces the computational complexity.

[0025] (3) A target k-node attack algorithm based on the improved onion algorithm is proposed to obtain a reduced set of candidate links, and the link with the greatest influence among the candidate links is iteratively deleted until all target nodes collapse. After reducing the candidate links, the present invention designs a k-node attack algorithm based on the improved onion algorithm to achieve the collapse of the target k nodes. The algorithm iteratively deletes the link with the greatest influence among the candidate links until all target nodes collapse, thereby achieving the anonymization of key figures in the social network. This strategy significantly reduces the computational complexity by simplifying the candidate links, ensuring the efficiency of the anonymization process.

[0026] Therefore, the present invention proposes a method for hiding key figures in social networks based on target k-node decomposition, which has significant innovation and practicality in protecting user privacy, maintaining social network functions and improving security.

[0027] The present invention improves the onion decomposition algorithm and the backtracking tree algorithm to construct a backtracking tree to obtain the minimum set of edges that need to be deleted from the graph, so that the specified target node is separated from the k-core subgraph, thereby realizing the concealment of key person information in the social network.

[0028] The advantages and effectiveness of this invention over existing technologies are as follows: The proposed method significantly reduces the number of candidate edges that need to be deleted to achieve anonymity, significantly reducing computation time and improving efficiency. By deleting edges and changing the k-core values ​​of key nodes, the nodes of key individuals in a social network can be effectively hidden without being discovered by hackers, thereby improving network security. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. 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 any creative work.

[0030] Figure 1 is a flow chart of the method of the present invention.

[0031] Figure 2 This is a flow chart of the onion decomposition algorithm improved in step S2 of the present invention.

[0032] Figure 3 The present invention is a schematic diagram generated after selecting some nodes in the network to perform an improved onion decomposition algorithm and construct a backtracking tree. Specific implementation methods

[0033] Various exemplary embodiments of the present invention will now be described in detail. This detailed description should not be considered as limiting the present invention, but rather as a more detailed description of certain aspects, features, and embodiments of the present invention.

[0034] It should be understood that the terms described herein are intended only to describe particular embodiments and are not intended to limit the present invention. In addition, for numerical ranges herein, it should be understood that each intermediate value between the upper and lower limits of the range is also specifically disclosed. Each smaller range between any intermediate value within a stated value or stated range and any other stated value or intermediate value within the stated range is also encompassed by the present invention. The upper and lower limits of these smaller ranges may be independently included or excluded within the scope.

[0035] Unless otherwise indicated, all technical and scientific terms used herein have the same meanings as commonly understood by those skilled in the art. Although only preferred methods and materials are described herein, any methods and materials similar or equivalent to those described herein may also be used in the practice or testing of the present invention. All documents mentioned in this specification are incorporated by reference to disclose and describe the methods and / or materials related to the documents. In the event of any conflict with any incorporated document, the contents of this specification shall prevail.

[0036] Many modifications and variations to the illustrative embodiments described herein will be apparent to those skilled in the art from consideration of this specification and practice of the subject technology. Additional embodiments of the present subject technology will be apparent to those skilled in the art from consideration of the specification and practice of the subject technology. The specification and examples given herein are exemplary only. It is to be understood that the specification and examples are illustrative of the subject technology and are mentioned merely to provide a general understanding of the subject technology. Accordingly, the subject technology is not limited to the specific details listed herein.

[0037] As used herein, the terms "comprise", "comprising", "include", "including", "have", "having" or the like are open-ended and do not exclude additional, unrecited elements or method steps.

[0038] The "parts" described in the present application are measured by mass unless otherwise specified.

[0039] Example 1

[0040] The present embodiment provides a social network key figure concealment method based on target k-node decomposition. With the wide application of social networks, user privacy protection is increasingly valued. A certain social network platform has millions of registered users, covering many well-known public figures, corporate executives, and scientific researchers. The social dynamics and personal information of these key figures are widely concerned, and the risk of privacy leakage is increasing. Using the concealment method proposed in the present application, the key figures can be protected, and the impact of deliberate attacks on key figures can be avoided. As shown in the figure, it specifically includes the following implementation steps:

[0041] Step S1: define the social network as an undirected and unweighted network G(V, E), V is the set of users in the network, and E is the set of user relationships in the network; the small social network consists of seven users, v1-v7; first, use the k-core decomposition algorithm to assign a core value C (v,G) to each user in the network, and the core values of the users are: Define the user with the highest core value in the social network as the key figure V k , at this time, user v7 is the key figure to be concealed.

[0042] Step S2: for each user, obtain its hierarchical weight information M and construct a backtracking tree G BT for the target user to obtain a candidate relationship edge set H; including the following steps:

[0043] S2.1: for each user in the network, find its neighbor user N (v,G) and node degree value d (v,G) , the node degree values are:

[0044] S2.2: according to the theorem, if and only if When edge (u, v) is removed, k-node collapses and the theorem holds for k = C (u,G) ≤ C (v,G) When edge (u, v) is removed, only the k-node with core number k collapses, and the core number of the node is reduced by at most 1. The initial candidate relationship edge set P = {(u, v) | (u, v) E, min (C (u,G) , C (v,G) ) = k}, for which the initial candidate relationship edge set P = {(v1, v2), (v2, v3), (v3, v4), (v4, v5), (v5, v6), (v2, v7), (v5, v7)}.

[0045] S2.3: Divide the users of network G into different layers using the improved onion decomposition algorithm, so as to further reduce the candidate relationship edge set; first, calculate the node degree value d (v,G) of all users, sort the node degree value to obtain the degree sequence, and peel off the user nodes from small to large according to the degree sequence, and divide the user nodes into different layers according to how many times it takes to reach a given user node; that is, if the degree of the k-shell node in the k-core is exactly k, it belongs to the first local layer; if the first layer is removed, the node with the maximum degree of k belongs to the second local layer, and so on to obtain the layer number M of each user node, the layer numbers are: M1 = 1, M3 = 1, M4 = 1, M6 = 1, M2 = 2, M5 = 2, M7 = 3.

[0046] S2.4: According to the layer number M, prefer to consider the neighbor with higher collapse order, take the target user node as the root node, and perform breadth-first search to visit all neighbor nodes with layer number less than the current node; first, calculate the layer number value of each node using the improved onion decomposition algorithm, and initialize the queue as the target node set; then initialize an empty directed graph G BT , which contains node set T and empty edge set; then traverse each node u in the queue, find all nodes v adjacent to u and with core number k, and store these nodes in set N', then find all nodes v in set N * and with layer number less than node u, and store these nodes in set N * , update the queue, remove u from the queue, and add all nodes in N BT but not in the backtracking tree G BT to the queue, update G * , add all nodes v in N BT and edge (u, v) to G BT ; finally, the reduced candidate relationship edge set H = E k ∪ {(u, v) | u T, v N (u, G BT )\ V}, for this network, the reduced candidate relationship edge set H = {(v1, v2), (v2, v3), (v2, v7), (v4, v5), (v5, v6), (v5, v7)}.

[0047] Step S3: traverse each edge e in the candidate relationship edge set H, and prune the edge e according to the pruning rule. First, remove the edge e from the backtracking tree to obtain a new subgraph network; at the same time, initialize the pruned follower set as the intersection of the follower set of the edge e and the target node set T; then find all the node set Z in the target node set T excluding the zero-in-degree node, remove the node set Z from the node set Z, and add the node set Z to the pruned follower set to obtain the pruned follower set; select the edge e with the most pruned followers * , add e * to the removed edge set, and finally obtain the removed edge set: {(v2, v7), (v5, v7)}.

[0048] Step S4: repeat step S3 until all nodes in the target node are removed from the k-core; for this small social network, the first iteration deletes the edge between nodes v2 and v7, thereby obtaining an updated subgraph; the second iteration deletes the edge between nodes v5 and v7, thereby achieving the anonymity of the core position of node v7, that is, the algorithm achieves the effect of hiding the key person in the social network.

[0049] By implementing the social network key person hiding method based on target k node decomposition provided by the present application, the privacy of the key person is successfully protected, and the risk of user privacy leakage is reduced. At the same time, the method also provides a privacy protection scheme for other similar platforms.

[0050] In summary, the present application uses a social network key person hiding method based on target k node decomposition, which selectively filters candidate edges for target nodes, overcomes the problem of low calculation speed of previous key person information hiding algorithms on large-scale networks, can successfully hide key person nodes in social networks, effectively maintains the security of the network, and meets the actual use requirements.

[0051] Embodiment 2

[0052] The present embodiment relates to a social network key person hiding device based on target k node decomposition, comprising a memory and one or more processors, the memory storing executable code, and the one or more processors executing the executable code to implement the social network key person hiding method based on target k node decomposition of embodiment 1.

[0053] Embodiment 3

[0054] The embodiment relates to a computer readable storage medium, which stores a program, and the program is executed by a processor to realize a social network key person concealment method based on target k node decomposition according to the embodiment 1.

[0055] The above-described embodiments are only used to describe the preferred modes of the present application, and are not used to limit the scope of the present application. Without departing from the design spirit of the present application, various modifications and improvements of the technical solutions of the present application made by those skilled in the art shall fall within the protection scope of the present application.

Claims

1. A method for hiding key figures in social networks based on target k-node decomposition, characterized in that: The steps include: S1: Assign core values ​​to nodes in the social network and define key figures; S2: Obtain node-level weight information and construct a backtracking tree for the target node to obtain a candidate edge set; specifically, it includes: S2.1: For each node in the network, find its neighbor nodes and node degree ; S2.2: According to theorem, if and only if When removing the edge will cause k nodes to crash and theorem to assume , remove After that, only the number of cores is The node will crash, and its core number will be reduced by 1 at most, and the initial candidate edge set will be obtained. ; S2.3: Using the improved onion decomposition algorithm to The k nodes are divided into different layers, so that the candidate edge set can be further simplified; first calculate the node degree values ​​of all nodes , and sort the node degrees to get the degree order, peel off the nodes from small to large according to the degree order, and divide the nodes into different layers according to how many times it takes to reach a given node; that is, if the degree of the node of the k-shell in the k-core is exactly , it belongs to the first local layer; if the first layer is removed, the node degree is at most The nodes belong to the second local layer, and so on to get the number of layers each node belongs to. ; S2.4: According to the number of layers Prioritize neighbors with higher collapse orders, and use the breadth-first search algorithm to iteratively backtrack neighbors that maintain increasing collapse order relative to the current node; first use the improved onion decomposition algorithm to calculate the layer value of each node and initialize the queue to the target node set; then initialize an empty directed graph , which contains a collection of nodes and an empty edge set; then traverse each node in the queue , find all adjacent and the number of cores is Node , and store these nodes in a collection , then find all nodes in the set whose layers are less than Node , and store these nodes in a collection , update the queue, Remove from the queue and all In but not in the backtracking tree The nodes in the queue are added and updated , all in Nodes in and the edge Add to Finally, we get the simplified candidate edge set ; S3: Select the most affected edge pruning and deletion based on the candidate edge set; S4: Repeat step S3 until all nodes in the target node are removed from the k-core subgraph, achieving the effect of hiding key figures.

2. The method for hiding key figures in a social network based on target k-node decomposition according to claim 1, characterized in that: The step S1 of assigning core values ​​to nodes in the social network and defining key figures specifically includes the following steps: S1.1: First, use the k-core decomposition algorithm to assign a core value to each node in the network ,The core value of a node is equivalent to the importance of the node. The higher the core value of the node, the higher the importance of the node; S1.2: Define the node with the highest core value in the social network as the key person .

3. The method for hiding key figures in a social network based on target k-node decomposition according to claim 1, characterized in that: Step S3 of selecting the most affected edge to be pruned and deleted based on the candidate edge set specifically includes the following steps: S3.1: Traverse the candidate edge set Each edge in , according to the trimming rule for the edge Perform pruning; S3.2: Removing edges from a backtracking tree , get a new subgraph, and initialize the pruned follower set as edge The follower set and target node set of The intersection of S3.3: Find all target node sets The node set excluding zero-indegree nodes , remove the node set from it , the node set Add the pruned follower set to obtain the pruned follower set; S3.4: Select the edge with the most followers after pruning ,Will Add to the set of removed edges, and finally get the set of removed edges.

4. A device for hiding key figures in social networks based on target k-node decomposition, characterized in that: The method comprises a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, they are used to implement a method for hiding key figures in a social network based on target k-node decomposition according to any one of claims 1 to 3.

5. A computer-readable storage medium, characterized in that A program is stored thereon, and when the program is executed by a processor, the method for hiding key figures in a social network based on target k-node decomposition according to any one of claims 1 to 3 is implemented.

Images