Social network structure anonymization method and device based on kernel value preserving algorithm

Abstract

The invention discloses a social network structure anonymization method and device based on a kernel value preserving algorithm. The method comprises the following steps: step 1, acquiring a social network reading and kernel value; in order to obtain the core value information of the social network nodes, data abstracted from the user relationship in the social network needs to be obtained, and network data with the relationship between the nodes and the user sides is constructed. Respectively acquiring a k-core value and a k-round value of the node according to a k-core decomposition algorithm and a hierarchical onion decomposition algorithm; step 2, obtaining an edge set Gc of a core to be removed; and screening out a connected edge set Gc of a core to be swapped out through a sub-graph composed of nodes with the core value being a k-max value. In the k-max subgraph, as long as the degree value of each node is greater than or equal to k-max, the maintenance of the core value can be realized; step 3, acquiring a to-be-exchanged peripheral edge set Gp; a to-be-exchanged peripheral edge set Gp is screened out from a peripheral sub-graph composed of nodes with the kernel values smaller than the k-max value, in the peripheral sub-graph, as long as the k-round values of all the nodes at the two ends of one edge are equal, neighbors can be exchanged with the edges of the core, and the kernel values of the core are maintained; and step 4, obtaining a social network anonymization result G ', and exchanging neighbors Gp between the edges Gc in the edge set of the core to be removed and the edges in the peripheral edge set to be exchanged, which are obtained in the step 2 and the step 3, representing an operable connection edge set, and keeping the core value of the network at the same time.

Description

Technical Field

[0001] The present invention belongs to the technical field of network security, and in particular relates to a method and device for anonymizing a social network structure based on a core value preservation algorithm. Background Art

[0002] With the advancement of technology, social networking platforms have become an indispensable part of people's daily lives. People share information, establish connections, and interact through these platforms. However, this highly interconnected social environment also brings serious privacy protection issues. In social networks, users' personal information and social relationship data are collected in large quantities, including but not limited to users' profiles, social behaviors, interaction records, etc. These data not only provide users with personalized service experience, but also provide a rich data source for social network analysis. However, the leakage and abuse of these data may lead to serious consequences, such as identity theft, fraudulent activities, malicious attacks, etc., posing a threat to users' privacy security.

[0003] In response to the above problems, the structure hiding technology of social relationship networks has been developed. This technology aims to hide or blur the nodes (users) and edges (relationships) in social networks so that third parties cannot easily obtain users' personal information and social relationships. Based on knowledge from multiple fields such as graph theory, data mining, and machine learning, various algorithms and technical means are needed to hide user relationships. These methods maintain the basic functions and structure of social networks as much as possible while protecting user privacy, thereby avoiding the risk of user privacy leakage and improving the security of social networks.

[0004] In addition, the continuous advancement of social network analysis technology enables researchers or third parties to reveal sensitive information such as users' social relationships and behavior patterns through in-depth analysis of user data. Although these analysis techniques help to understand the structure and behavior of social networks, they may also lead to the leakage of user privacy. For example, by analyzing the user's social relationship network, sensitive information such as the user's occupation and health status can be inferred, so as to accurately profile the user and conduct targeted marketing. In order to solve the above-mentioned problem of anonymization of social network structure, a kernel value preservation algorithm is proposed to retain the original characteristics of the network. 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 anonymizing a social network structure based on a kernel value preservation algorithm.

[0006] The first aspect of the present invention relates to a method for anonymizing a social network structure based on a kernel value preservation algorithm, comprising the following steps:

[0007] Step 1: Get the social network reading and core value; in order to obtain the core value information of the social network node, it is necessary to obtain the data of abstract user relationships in the social network and construct network data with nodes as users and edges as relationships. According to the k-core decomposition algorithm and the class onion decomposition algorithm, the k-core value and k-round value of the node are obtained respectively.

[0008] Step 2: Get the edge set Gc to be removed from the core; filter out the edge set Gc to be replaced from the core by filtering the subgraph composed of nodes with core values ​​equal to k-max. In the k-max subgraph, as long as the degree value of each node is greater than or equal to k-max, the core value can be maintained.

[0009] Step 3 obtains the peripheral edge set Gp to be exchanged; the peripheral edge set Gp to be exchanged is screened out by the peripheral subgraph composed of nodes whose core values ​​are less than the k-max value. In the peripheral subgraph, as long as the k-round value of each node at both ends of an edge is equal, the neighbors can be exchanged with the core edge to maintain the core core value.

[0010] Step 4 obtains the social network anonymization result G'; the edges Gc in the edge set to be removed from the core obtained in steps 2 and 3 are exchanged with the neighbors Gp in the edge set to be exchanged, representing an operable edge set while maintaining the core value of the network.

[0011] The steps of obtaining the social network reading and verification value in step 1 are as follows:

[0012] Step 1-1: In the initial state, input the relationship data of the social network to construct a social network G = (V, E) with nodes as users and edges as relationships, where V represents the set of nodes and E represents the set of edges;

[0013] Step 1-2: Use the k-core decomposition algorithm to obtain the k-core value of each node in the network, and record the maximum value as k-max;

[0014] Step 1-3: Use the hierarchical onion decomposition algorithm to obtain the k-round value of each node in the network.

[0015] The steps of step 2 to obtain the edge set Gc to be removed from the core are as follows:

[0016] Step 2-1: Initial state, input social network G, core value k-core and maximum value k-max of k-core;

[0017] Step 2-2: Filter the nodes whose k-core value is k-max, and the edges between the nodes, to obtain the core subgraph Gt that they jointly constitute;

[0018] Step 2-3: Traverse each edge in the core subgraph Gt, and record the edges whose degree values ​​of the nodes at both ends of the edge in the core subgraph are greater than k-max as Gc;

[0019] Step 2-4: Output the final Gc result.

[0020] Step 2-5: Repeat 2-2 until all data in T are run;

[0021] Wherein, step 3 specifically includes:

[0022] Step 3-1: Initial state, input social network G, core value k-core, k-round value and maximum value k-max of k-core;

[0023] Step 3-2: Take the nodes with k-core values ​​equal to k and the edges between the nodes to form a k-shell subgraph. The initial value of k is set to 1.

[0024] Step 3-3: For all the edges in the k-shell subgraph, determine the k-round values ​​of the nodes on both sides of the edge. If the k-round values ​​of the nodes on both sides are equal, record them in the peripheral edge set Gp to be exchanged.

[0025] Step 3-4: After the value of k increases by 1, repeat steps 3-2 and 3-3 until the value of k is equal to k-max-1;

[0026] Step 3-5: Output the final Gp result.

[0027] The steps of obtaining the social network anonymization result G' in step 4 are as follows:

[0028] Step 4-1: In the initial state, input the social network G, the core value k-core and the maximum value k-max of k-core, and the edge Gc in the edge set to be removed from the core and the edge in the edge set to be exchanged with the neighbor Gp;

[0029] Step 4-2: Select an edge Ec in Gc as the core edge to be exchanged;

[0030] Step 4-3: For the edges belonging to Gp in the k-shell subgraph where k is equal to k-max-1, determine whether they have a neighbor relationship with the nodes in Ec. If not, the exchange is successful. Otherwise, reduce the k value and then determine the Gp value that meets the conditions until Gp is ​​traversed;

[0031] Step 4-4: Repeat steps 4-2 and 4-3 until all exchangeable edges in Gc have been exchanged, and obtain the social network anonymization result G';

[0032] Step 4-4: Verify whether the core values ​​k-core and k-round of the nodes in G' are the same as those of the original social network G, and verify that anonymization is successfully completed;

[0033] Step 4-5: Output the final G' result. In G, A and B have a neighbor relationship, but after anonymization, the edge between A and B no longer exists, and the edge between C and D no longer exists; while protecting the privacy of the ABCD relationship, the degree value and kernel value have not changed.

[0034] A second aspect of the present invention relates to a social network structure anonymization device based on a core value preservation algorithm, 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 social network structure anonymization method based on a core value preservation algorithm of the present invention.

[0035] A third aspect of the present invention relates to a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements a social network structure anonymization method based on a core value preservation algorithm of the present invention.

[0036] Efficiency analysis: The core value preservation algorithm can also be implemented by exhaustive method. The social network G has m edges. The exhaustive method needs to execute (m-1) 2 The edge Gc in the edge set to be removed from the core of the present invention exchanges its neighbor Gp with the edge in the peripheral edge set to be exchanged, assuming that they have p(m) and q(m) edges respectively, where p and q are less than 1 and p+q≤1. The edge exchange operation on Gc and Gp requires the execution of pq(m) 2 times, much less than (m-1) 2 The social network structure anonymization method based on the core value preservation algorithm improves the efficiency of the core value preservation algorithm by screening Gc and Gp.

[0037] The key points of the present invention are:

[0038] 1. Comprehensive network structure protection: Different from the existing technology that mainly focuses on node characteristics, the present invention will focus on the protection of the overall network structure to ensure that important structural information of the network is retained during the anonymization process.

[0039] 2. Anonymization method with core value preservation: This paper proposes a new anonymization method that can maintain the core value characteristics of nodes in the network while changing the network structure. This means that the results of scientific calculations and analysis performed on the anonymized network will be closer to the original social network, ensuring the accuracy and reliability of the research.

[0040] 3. Wide versatility: The technical solution of the present invention is applicable to all types of social network structures, and can achieve effective anonymization protection regardless of the size of the network and the complexity of the structure.

[0041] This paper proposes a social network structure anonymization method based on a core value preservation algorithm, which can protect user privacy while maintaining the basic functions and structure of the social network to the maximum extent, effectively avoiding the risk of user privacy leakage and improving the security of the social network. The main innovations are as follows:

[0042] 1. Application of the core value preservation algorithm. This invention applies the core value preservation algorithm to the field of social network structure anonymization for the first time. By introducing the k-core decomposition and onion decomposition algorithms, the core value of each node in the social network can be accurately calculated, thereby providing a basis for subsequent edge set screening and exchange. This method not only ensures the rationality of the network structure after anonymization, but also ensures that the core characteristics of the social network are preserved.

[0043] 2. Accurate determination of the core edge set Gc to be removed and the peripheral edge set Gp to be exchanged. The present invention determines the core edge set Gc to be removed and the peripheral edge set Gp to be exchanged through precise calculation and analysis. This method ensures that the core value characteristics of the network can be maintained to the maximum extent when edge exchange is performed, while avoiding unnecessary edge exchange and improving the efficiency of the algorithm.

[0044] 3. Optimization of edge exchange strategy. After determining the core edge set Gc to be removed and the peripheral edge set Gp to be exchanged, the present invention adopts an optimized edge exchange strategy. By traversing the edges in Gc and exchanging them with the edges in Gp, and verifying the core value after the exchange, the anonymization of the social network structure is achieved. This strategy not only ensures the rationality of the network structure after anonymization, but also ensures the efficiency of the anonymization process.

[0045] Therefore, the present invention has significant innovation and practicality in protecting user privacy, maintaining social network functions and improving security. In order to solve the problem of leaking user privacy when social platforms disclose user relationship network data sets, existing methods often achieve this through random interactive edges. Although the anonymization of the network structure can be achieved, it also changes the characteristics of the network structure, affecting the results of subsequent data analysis and scientific calculations. The method proposed in the present invention retains the core value of social network data, reduces the impact on data analysis and scientific calculation results while realizing the anonymization of user relationships, and can be used for the anonymous disclosure of social network data.

[0046] In order to solve the problem of user privacy leakage when social platforms disclose user relationship network data sets, existing methods often achieve this through random interactive edges. Although the network structure can be anonymized, it also changes the characteristics of the network structure, affecting the results of subsequent data analysis and scientific calculations. The method proposed in the present invention retains the core value of social network data, realizes the anonymization of user relationships, and reduces the impact on data analysis and scientific calculation results, which can be used for the anonymous disclosure of social network data.

[0047] The advantages of the present invention are:

[0048] 1. Application of the core value preservation algorithm. This invention applies the core value preservation algorithm to the field of social network structure anonymization for the first time. By introducing the k-core decomposition and onion decomposition algorithms, the core value of each node in the social network can be accurately calculated, thereby providing a basis for subsequent edge set screening and exchange. This method not only ensures the rationality of the network structure after anonymization, but also ensures that the core characteristics of the social network are preserved.

[0049] 2. Accurately determine the core edge set Gc to be removed and the peripheral edge set Gp to be exchanged. The present invention determines the edge set Gc to be removed from the core and the edge set Gp to be exchanged in the periphery through precise calculation and analysis. The core value preservation algorithm can also be implemented by an exhaustive method. The social network G has m edges, and the exhaustive method needs to be executed to implement the core value preservation algorithm. The edge Gc in the edge set to be removed from the core in this method exchanges neighbors Gp with the edge in the peripheral edge set to be exchanged. It is assumed that they have p(m) and q(m) edges respectively, where p and q are less than 1 and p+q≤1. The edge exchange operation on Gc and Gp needs to be executed times, which is much less than. The social network structure anonymization method based on the core value preservation algorithm improves the efficiency of the core value preservation algorithm by screening Gc and Gp.

[0050] 3. Optimize the edge exchange strategy. After determining the core edge set Gc to be removed and the peripheral edge set Gp to be exchanged, the present invention adopts an optimized edge exchange strategy. By traversing the edges in Gc and exchanging them with the edges in Gp, and verifying the core value after the exchange, the anonymization of the social network structure is achieved. This strategy not only ensures the rationality of the network structure after anonymization, but also ensures the efficiency of the anonymization process. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 is a flow chart of the method of the present invention.

[0052] Figure 2 It is a schematic diagram of obtaining the core value of a social network node of the present invention.

[0053] Figure 3 Schematic diagram of the present invention for exchanging edge neighbors. DETAILED DESCRIPTION

[0054] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0055] Example 1

[0056] This embodiment relates to a social network structure anonymization method based on a kernel value preservation algorithm.

[0057] Based on the purpose of anonymizing the social network structure, the process of the method of the present invention is as follows: Figure 1 As shown, there are three main steps: 1. According to the results of k-core decomposition and onion decomposition, find the edge set Gc to be removed from the core. 2. According to the results of k-core decomposition and onion decomposition, find the peripheral edge set Gp to be exchanged. 3. Traverse the edges Gc in the edge set to be removed from the core and exchange neighbors Gp with the edges in the peripheral edge set to be exchanged, and obtain the structural anonymization result after verifying the core value.

[0058] To achieve the above objectives, the present invention proposes a social network structure anonymization method based on a kernel value preservation algorithm, and the specific steps are as follows:

[0059] Step 1: Get social network read and core values.

[0060] In order to obtain the core value information of social network nodes, it is necessary to obtain the data of abstract user relationships in the social network and construct network data with nodes as users and edges as relationships. The k-core value and k-round value of the node are obtained according to the k-core decomposition algorithm and the class onion decomposition algorithm respectively.

[0061] like Figure 2 As shown, the steps for obtaining social network reading and core value are as follows:

[0062] Step 1-1: In the initial state, input the relationship data of the social network to construct a social network G = (V, E) with nodes as users and edges as relationships, where V represents the set of nodes and E represents the set of edges;

[0063] Step 1-2: Use the k-core decomposition algorithm to obtain the k-core value of each node in the network, and record the maximum value as k-max;

[0064] Step 1-3: Use the hierarchical onion decomposition algorithm to obtain the k-round value of each node in the network.

[0065] Step 2: Get the edge set Gc to be removed from the core.

[0066] The present invention selects the edge set Gc of the core to be replaced by the subgraph composed of nodes whose core value is k-max value. In the k-max subgraph, as long as the degree value of each node is greater than or equal to k-max, the core value can be maintained.

[0067] The steps to obtain the edge set Gc to be removed from the core are as follows:

[0068] Step 2-1: Initial state, input social network G, core value k-core and maximum value k-max of k-core;

[0069] Step 2-2: Filter the nodes whose k-core value is k-max, and the edges between the nodes, to obtain the core subgraph Gt that they jointly constitute;

[0070] Step 2-3: Traverse each edge in the core subgraph Gt, and record the edges whose degree values ​​of the nodes at both ends of the edge in the core subgraph are greater than k-max as Gc;

[0071] Step 2-4: Output the final Gc result.

[0072] Step 2-5: Repeat 2-2 until all data in T are run;

[0073] Step 3: Get the peripheral edge set Gp to be exchanged.

[0074] The present invention selects the peripheral edge set Gp to be exchanged by filtering the peripheral subgraph composed of nodes whose core values ​​are less than the k-max value. In the peripheral subgraph, as long as the k-round value of each node at both ends of an edge is equal, the neighbors can be exchanged with the core edge to maintain the core core value.

[0075] Step 3-1: Initial state, input social network G, core value k-core, k-round value and maximum value k-max of k-core;

[0076] Step 3-2: Take the nodes with k-core value equal to k and the edges between the nodes to form a k-shell subgraph. The initial value of k is set to 1.

[0077] Step 3-3: For all the edges in the k-shell subgraph, determine the k-round values ​​of the nodes on both sides of the edge. If the k-round values ​​of the nodes on both sides are equal, record them in the peripheral edge set Gp to be exchanged.

[0078] Step 3-4: After the value of k increases by 1, repeat steps 3-2 and 3-3 until the value of k is equal to k-max-1

[0079] Step 3-5: Output the final Gp result.

[0080] Step 4: Obtain the social network anonymization result G'.

[0081] The edges Gc in the edge set to be removed from the core obtained in step 2 and step 3 exchange neighbors Gp with the edges in the peripheral edge set to be exchanged, representing an operable edge set while maintaining the core value of the network.

[0082] like Figure 3 As shown, the steps of social network anonymization result G' are as follows:

[0083] Step 4-1: In the initial state, input the social network G, the core value k-core and the maximum value k-max of k-core, and the edge Gc in the edge set to be removed from the core and the edge in the edge set to be exchanged with the neighbor Gp;

[0084] Step 4-2: Select an edge Ec in Gc as the core edge to be exchanged;

[0085] Step 4-3: For the edges belonging to Gp in the k-shell subgraph where k is equal to k-max-1, determine whether they have a neighbor relationship with the nodes in Ec. If not, the exchange is successful. Otherwise, reduce the k value and then determine the Gp value that meets the conditions until Gp is ​​traversed;

[0086] Step 4-4: Repeat steps 4-2 and 4-3 until all exchangeable edges in Gc have been exchanged, and obtain the social network anonymization result G';

[0087] Step 4-4: Verify whether the core values ​​k-core and k-round of the nodes in G' are the same as those of the original social network G, and verify that anonymization is successfully completed;

[0088] Step 4-5: Output the final G' result. In G, A and B have a neighbor relationship, but after anonymization, the edge between A and B no longer exists, and the edge between C and D no longer exists. While protecting the privacy of the ABCD relationship, the degree value and kernel value have not changed.

[0089] Example 2

[0090] The present embodiment relates to a social network structure anonymization device based on a core value preservation algorithm, including 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 social network structure anonymization method based on a core value preservation algorithm of embodiment 1.

[0091] Example 3

[0092] This embodiment relates to a computer-readable storage medium on which a program is stored. When the program is executed by a processor, a social network structure anonymization method based on a core value preservation algorithm of embodiment 1 is implemented.

[0093] The contents described in the embodiments of this specification are merely an enumeration of the implementation forms of the inventive concept. The protection scope of the present invention should not be regarded as limited to the specific forms described in the embodiments. The protection scope of the present invention also extends to equivalent technical means that can be conceived by those skilled in the art based on the inventive concept.

Claims

1. A method for anonymizing social network structure based on a kernel value preservation algorithm, comprising the following steps: Step 1: Obtain social network reading and core value; in order to obtain the core value information of social network nodes, it is necessary to obtain the data of abstracted user relationships in the social network and construct network data with nodes as users and edges as relationships; obtain the k-core value and k-round value of the node according to the k-core decomposition algorithm and the class onion decomposition algorithm respectively; Step 2: Obtain the edge set Gc to be removed from the core; filter out the edge set Gc to be replaced from the core by the subgraph composed of nodes whose core value is k-max value; in the k-max subgraph, as long as the degree value of each node is greater than or equal to k-max, the core value can be maintained; Step 3: Obtain the peripheral edge set Gp to be exchanged; filter out the peripheral edge set Gp to be exchanged by the peripheral subgraph composed of nodes whose core values ​​are less than the k-max value. In the peripheral subgraph, as long as the k-round value of each node at both ends of an edge is equal, the neighbors can be exchanged with the core edge to maintain the core core value. Step 4 obtains the social network anonymization result G'; the edges Gc in the edge set to be removed from the core obtained in steps 2 and 3 are exchanged with the neighbors Gp in the edge set to be exchanged, representing an operable edge set while maintaining the core value of the network.

2. The social network structure anonymization method based on the core value preservation algorithm according to claim 1, characterized in that: The steps of social network reading and verification value acquisition described in step 1 are as follows: Step 1-1: In the initial state, input the relationship data of the social network to construct a social network G = (V, E) with nodes as users and edges as relationships, where V represents the set of nodes and E represents the set of edges; Step 1-2: Use the k-core decomposition algorithm to obtain the k-core value of each node in the network, and record the maximum value as k-max; Step 1-3: Use the hierarchical onion decomposition algorithm to obtain the k-round value of each node in the network.

3. The social network structure anonymization method based on the core value preservation algorithm according to claim 1, characterized in that: Step 2: The steps to obtain the edge set Gc to be removed from the core are as follows: Step 2-1: Initial state, input social network G, core value k-core and maximum value k-max of k-core; Step 2-2: Filter the nodes whose k-core value is k-max, and the edges between the nodes, to obtain the core subgraph Gt that they jointly constitute; Step 2-3: Traverse each edge in the core subgraph Gt, and record the edges whose degree values ​​of the nodes at both ends of the edge in the core subgraph are greater than k-max as Gc; Step 2-4: Output the final Gc result; Step 2-5: Repeat step 2-2 until all data in T are processed.

4. The method for anonymizing a social network structure based on a core value preservation algorithm according to claim 1, characterized in that: Step 3 specifically includes: Step 3-1: Initial state, input social network G, core value k-core, k-round value and maximum value k-max of k-core; Step 3-2: Take the nodes with k-core values ​​equal to k and the edges between the nodes to form a k-shell subgraph. The initial value of k is set to 1. Step 3-3: For all the edges in the k-shell subgraph, determine the k-round values ​​of the nodes on both sides of the edge. If the k-round values ​​of the nodes on both sides are equal, record them in the peripheral edge set Gp to be exchanged. Step 3-4: After the value of k increases by 1, repeat steps 3-2 and 3-3 until the value of k is equal to k-max-1; Step 3-5: Output the final Gp result.

5. The method for anonymizing a social network structure based on a core value preservation algorithm according to claim 1, characterized in that: Step 4: The steps to obtain the social network anonymization result G' are as follows: Step 4-1: In the initial state, input the social network G, the core value k-core and the maximum value k-max of k-core, and the edge Gc in the edge set to be removed from the core and the edge in the edge set to be exchanged with the neighbor Gp; Step 4-2: Select an edge Ec in Gc as the core edge to be exchanged; Step 4-3: For the edges belonging to Gp in the k-shell subgraph where k is equal to k-max-1, determine whether they have a neighbor relationship with the nodes in Ec. If not, the exchange is successful. Otherwise, reduce the k value and then determine the Gp value that meets the conditions until Gp is ​​traversed; Step 4-4: Repeat steps 4-2 and 4-3 until all exchangeable edges in Gc have been exchanged, and obtain the social network anonymization result G'; Step 4-4: Verify whether the core values ​​k-core and k-round of the nodes in G' are the same as those of the original social network G, and verify that anonymization is successfully completed; Step 4-5: Output the final G' result; A and B have a neighbor relationship in G, but after anonymization, the edge between A and B no longer exists, and the edge between C and D no longer exists; while protecting the privacy of the ABCD relationship, the degree value and kernel value have not changed.

6. A social network structure anonymization device based on a kernel value preservation algorithm, characterized in that: It 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, it is used to implement a social network structure anonymization method based on a core value preservation algorithm as described in any one of claims 1-5.

7. A computer-readable storage medium, characterized in that: A program is stored thereon, and when the program is executed by a processor, a social network structure anonymization method based on a core value preservation algorithm as described in any one of claims 1 to 5 is implemented.

Images