Method for promoting importance level of social network characters based on edge growth

Abstract

Method for promoting importance level of social network characters based on edge growth, including: Step S1, assuming an undirected and unweighted network G, using the K-core decomposition algorithm to assign core values to each node in the network G, and selecting the subgraph network G' with the highest core value as the subgraph whose core value is expected to be promoted; S2, pruning the subgraph G' according to the degree values of the nodes in G', removing the nodes with small degree values, and only retaining the largest connected subgraph G'', calculating the priorities of each node in the subgraph G'' according to the path length and degree value between the target nodes, and generating a candidate node list; S3, generating the links to be connected within the candidate node list, sorting the links according to the priorities of the nodes by the calculation formula, and sequentially adding the links with the highest priority to the original network; The present invention uses the method of edge growth to generate a list of links to be connected, reduces the complexity of link search, and cooperates with the greedy algorithm to quickly and efficiently increase the core value of the expected nodes, and effectively promotes the importance level of the leader characters in the social network.

Description

Technical Field

[0001] The present invention relates to data mining and network node processing technologies, and particularly to a method for promoting the importance level of social network characters based on edge growth. Background Art

[0002] Many researchers have proposed many methods to identify key nodes in a network or measure the influence of users in an online social network. The K-core decomposition algorithm is widely used in bioinformatics, network visualization, and importance measurement of complex networks. The K-core structure of a network can be obtained using the K-core decomposition algorithm. Generally speaking, the K-core structure of a network is the largest subgraph in the network, where each node has at least k links connecting to other nodes in the subgraph. Therefore, the degree of all vertices in it is at least k. In the research of the K-core decomposition algorithm, people have begun to notice the criticality that appears during the K-core decomposition process, that is, there are only two results after the decomposition process. One is that the network disappears and no nodes can survive during the pruning process; the other is that no nodes can be further removed, so that the network finally has a K-core subgraph with a fixed size. On the contrary, the method of using edge growth can utilize the criticality of the K-core decomposition algorithm to promote the importance level of nodes.

[0003] A social network is a social structure composed of a group of social participants, a group of binary relationships, and other social interactions among the participants. The growth of social networks has greatly enhanced people's awareness of interconnectedness in social life. Social networks and their analysis are an inherently interdisciplinary academic field that emerges from social psychology, sociology, statistics, and graph theory. Many important analysis problems are now viewed from the perspective of networks. For example: how events and attributes spread and are distributed among humans, why gossip spreads faster than reality, and how traffic congestion occurs and is alleviated. Scholars believe that key figures in social networks have an important influence on the propagation process and scope.

[0004] Related scholars studied 29 corruption cases exposed on Sina Weibo, found the influence degree of opinion leaders among various users, and found that news agencies and online media are the main opinion leaders and can dominate public opinion to a certain extent. Using social networks and econometrics for modeling, comparing and analyzing the relationship between opinion leaders and ordinary nodes, it is found that the authority of opinion leaders can enhance the group influence. Through questionnaire surveys, based on the perspectives of individual motivation and social capital, the factors influencing Weibo opinion leaders are studied, and the data results show that social interaction relationships, trust, and recognition all have significant effects on Weibo opinion leaders.

[0005] Opinion leaders in social networks play an important role in guiding public opinion in social networks. Social media such as Weibo, WeChat, and short videos can spread and report on major emergencies, uncover the truth, and convey public sentiment. In the process of a large number of social media information interactions and transmissions, the relevant figures in the spread of public opinion events play a guiding and promoting role in these events. In cognitive warfare, quickly seizing the high ground of public opinion is crucial for safeguarding our country's culture in related fields including the virtual space, guiding the healthy development of relevant groups, and building a clear and orderly online space in our country.

[0006] With the continuous expansion of the scale of e-commerce, the number and variety of goods have grown rapidly, and customers need to spend a lot of time to find the products they want to buy. This process of browsing a large amount of irrelevant information and products will undoubtedly cause continuous loss of consumers overwhelmed by the problem of information overload. In user-based recommendation systems, recommendations from friends are very important for increasing users' trust in the recommendation results. Key figures in social networks can also play a huge role in network marketing. For example, current live streaming with goods is a form of viral marketing that relies on spontaneous word-of-mouth publicity from users to achieve a rapid snowballing spread effect. According to the current real situation of social media information dissemination, quickly enhancing the role and influence of key figures in social networks is an important entry point for solving and realizing the above problems.

[0007] The invention patent with the publication number CN 104010003 A proposes a penetration method for quickly enhancing the importance of Sybil nodes. The main core is to increase the in-degree of Sybil nodes to quickly enhance the importance of Sybil nodes, making it easier for Sybil nodes to integrate into the original network and avoid detection of Sybil nodes by the Kademlia network. It can be seen that the index for measuring the importance of nodes in this invention is degree centrality. This invention only uses the topological property of node degree to enhance the importance of injected false nodes, does not use more topological properties of network nodes to enhance node importance, and its various calculation methods in the Internet of Things network are complex, with low time efficiency during operation, and its generalization and applicability will also be greatly reduced in other networks. Summary of the Invention

[0008] This invention aims to overcome the above-mentioned shortcomings of the prior art and provides a method for enhancing the importance level of social network figures based on edge growth.

[0009] Based on the classic network node importance ranking index, the K-core decomposition algorithm, this invention uses the edge growth method, that is, screening links that meet the restrictive conditions, adding some links to the original network, and through adversarial attacks, changing the core values of network nodes, increasing the core values of nodes, effectively enhancing the importance level of figures in social networks and making them key figures in the network.

[0010] The technical solution adopted by the present invention to achieve the above-mentioned invention purpose is as follows:

[0011] A method for promoting the importance level of social network characters based on edge growth includes the following steps:

[0012] S1: Assume an undirected unweighted network G. Using the K-core decomposition algorithm, obtain the core values of each node in the network G, and select the subgraph G' with the highest core value as the subgraph whose core value is expected to be promoted.

[0013] S2: Define the list of nodes with the highest degree values in the subgraph G' as top. Delete the subgraph G' according to the degree values of the nodes in G', remove the nodes with small degree values, and only retain the largest connected subgraph G" containing top. Calculate the priority of each node in the subgraph G" according to the path length and degree value between the target nodes, sort the nodes according to the priority, and generate a candidate node list candidate.

[0014] S3: Generate the links to be connected within the candidate node list candidate, sort the links according to the calculation formula of the node priority, generate a link list, and use the greedy algorithm to sequentially add the links with the highest priority to the original network.

[0015] Preferably, in the step S1: Assume an undirected unweighted social network G=(V, E), where V is the set of nodes in the network and E is the set of links in the network. Use the K-core decomposition algorithm to assign a core value to each node in the network, and select a series of nodes with the highest core value and their links to form the subgraph G'. The graph G'=(V', E') is a subgraph of another graph G=(V, E). Its vertex set V' is a subset of V, and its edge set E' is a subset of E. The subgraph relationship is denoted as

[0016] Preferably, in the step S2:

[0017] S2.1: Define one or several nodes with the highest degree values in the subgraph G' as top, and delete the subgraph G' layer by layer according to the degree values of the nodes in G', remove the nodes with small degree values until top only exists in the largest connected subgraph G";

[0018] S2.2: Calculate the node distance between any node i in the network and the top node. The node distance refers to the number of edges included in the shortest path between two nodes, denoted as X i ; Normalize the node distance. The normalization formula is as follows formula (1):

[0019]

[0020] Where, X minis the minimum value of all node distances, X max is the maximum value among all node distances;

[0021] S2.3: Calculate the degree of each node, denoted as Y i Normalize the degree of the node. The normalization formula is as follows (2):

[0022]

[0023] Among them, Y min is the minimum value among all degree values, Y max is the maximum value among all degree values;

[0024] S2.4: The priority of node i can be calculated according to formula (3):

[0025] ρ i =x i +y i (3)

[0026] ρ i The smaller it is, the higher the priority of node i. The nodes are sorted according to the priority and a candidate node list candidate is generated.

[0027] Preferably, in step S3: generating links to be connected in the candidate node list candidate, the links do not include existing links in the original network, and sorting the links according to the priority of the nodes. The priority is assigned in the following manner (4):

[0028]

[0029] Generate a link list and use the greedy algorithm to add the link with the highest priority to the original network. Note that each time a new link is added to the network, the link priority needs to be updated before the link can be added again, until the maximum core value of the original network G changes. At this point, the importance level of some nodes in the original network G has been effectively improved.

[0030] For real social networks, the node importance ranking algorithm provided by the K-core decomposition algorithm is very sensitive to perturbations in the network topology, making it unreliable for incomplete or noisy systems. This invention exploits this characteristic of social networks and uses small perturbations through edge growth to quickly and effectively improve the importance level of people in the social network.

[0031] The advantages of the present invention are as follows: Using the K-core decomposition algorithm as an index to measure the importance of nodes, and by means of edge growth method, based on the topological structure of the network, it can ensure that under minor perturbations, the core value of nodes is increased, effectively enhancing the importance level of characters in the social network, and calculating the priority ranking of links to improve the operation efficiency of the greedy algorithm. Brief Description of the Drawings

[0032] Figure 1 is the flowchart of the method of the present invention;

[0033] Figure 2 is an application case of the present invention on the dolphin dataset, and the node with the highest core value in the network is a 4-core node;

[0034] Figure 3 is to add a link ("20" - "29") to the dolphin dataset network, and the core values of some nodes in the network are increased to 5 cores. Detailed Embodiments

[0035] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0036] Embodiment 1

[0037] This embodiment provides a network public opinion guidance method based on the method for enhancing the importance level of characters in a social network based on edge growth of the present invention.

[0038] In order to improve the importance level of nodes, the present invention is based on the classical network node importance ranking index, the K-core decomposition algorithm. Through the edge growth method, that is, screening links that meet the restrictive conditions, adding some links to the original network, and after adversarial attacks, the core values of network nodes are changed to increase the core values of nodes, effectively enhancing the importance level of characters in the social network and making them key characters in the network.

[0039] The technical solutions adopted by the present invention to achieve the above-mentioned invention purposes are as follows:

[0040] The method for enhancing the importance level of characters in a social network based on edge growth includes the following steps:

[0041] S1: Assume an undirected unweighted network G. Using the K-core decomposition algorithm, obtain the core values of each node in the network G, and select the subgraph G' with the highest core value as the subgraph whose core value is expected to be increased;

[0042] S2: Define the list of nodes with the highest degree values in the subgraph G' as top. Prune the subgraph G' according to the degree values of the nodes in G', remove the nodes with small degree values, and only retain the largest connected subgraph G" that contains top. Calculate the priority of each node in the subgraph G" according to the path length and degree value between the target nodes, and sort the nodes according to the priority to generate the candidate node list candidate.

[0043] S3: Generate the links to be connected within the candidate node list candidate, sort the links according to the priority by the calculation formula for the nodes, generate the link list, and use the greedy algorithm to sequentially add the links with the highest priority to the original network.

[0044] In step S1: Assume an undirected unweighted social network G = (V, E), where V is the set of nodes in the network and E is the set of links in the network. Use the K-core decomposition algorithm to assign a core value to each node in the network, and select a series of nodes with the highest core values and their links to form the subgraph G'. The graph G' = (V', E') is a subgraph of another graph G = (V, E). Its vertex set V' is a subset of V, and its edge set E' is a subset of E. The subgraph relationship is denoted as

[0045] Step S2 specifically includes:

[0046] S2.1: Define one or several nodes with the highest degree values in the subgraph G' as top, and prune layer by layer according to the degree values of the nodes in the subgraph G', remove the nodes with small degree values until top only exists in the largest connected subgraph G";

[0047] S2.2: Calculate the node distance between any node i in the network and the top node. The node distance refers to the number of edges included in the shortest path between two nodes, denoted as X i ; Normalize the node distance. The normalization formula is as follows in Equation (1):

[0048]

[0049] where X min is the minimum value among all node distances, and X max is the maximum value among all node distances;

[0050] S2.3: Calculate the degree value of each node, denoted as Y i ; Normalize the degree value of the node. The normalization formula is as follows in Equation (2):

[0051]

[0052] where Y min is the minimum value among all degree values, and Y maxis the maximum value among all degree values;

[0053] S2.4: Then the priority of node i can be determined according to formula (3):

[0054] ρ i = x i + y i (3)

[0055] ρ i The smaller ρ is, the higher the priority of node i. Sort the nodes according to the priority to generate the candidate node list candidate.

[0056] In step S3: Generate the links to be connected within the candidate node list candidate. The links do not include the existing links in the original network. Sort the links according to the priority of the nodes. The way to assign the priority is as follows in formula (4),

[0057]

[0058] Generate the link list. Use the greedy algorithm to add the link with the highest priority to the original network. Note that each time a new link is added to the network, the priority of the link needs to be updated before adding another link until the maximum core value of the original network G changes. Thus, the importance levels of some nodes in the original network G have been effectively improved.

[0059] For the method for enhancing the importance level of social network characters based on edge growth of the present invention, the final results show that the method for enhancing the importance level of social network characters based on edge growth has excellent effects and can successfully improve the importance levels of characters in the social network. The final results based on this method can play a crucial role in the information dissemination direction of the social network. On the one hand, in the aspect of public opinion monitoring, the key characters in the social network can effectively promote the direction of network public opinion. Network opinion leaders are a special group that influences netizens through suggestions or opinions in the network. In today's complex Internet environment, some netizens will inevitably be influenced by network opinion leaders. If the positive role of network opinion leaders can be fully exerted, it will help to guide the healthy development of the network environment, be beneficial to the mental health development of netizens, and contribute to building a clear and orderly Internet environment in China.

[0060] Embodiment 2

[0061] This embodiment provides an e-commerce product recommendation method based on the method for enhancing the importance level of social network characters based on edge growth of the present invention.

[0062] Application example of the present invention in the direction of network marketing. First, from traditional recommendation systems to emerging social recommendation systems, users and content associated with key figures are given higher weights, which has a subtle impact on the recommendation technologies of various information distribution platforms. Second, in the era of live e-commerce shopping that has emerged today, viral marketing uses the spread among groups to enable people to establish an understanding of services and products and achieve the purpose of publicity. Key figures can often be used to release some messages for publicity, which has a profound impact on users.

[0063] Specifically, this embodiment includes steps S1 to S3 of Embodiment 1. In step S4, users and content associated with key figures are given higher weights to promote the spread and marketing of relevant goods and services.

[0064] The content described in the embodiments of this specification is only 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 stated 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 of the present invention.

Claims

1. A method for improving the importance level of social network characters based on edge growth, characterized by: The following steps are involved: S1: Assume an undirected and unweighted network G. Use the K-core decomposition algorithm to obtain the core value of each node in the network G. Select the subgraph G′ with the highest core value and set it as the subgraph with the desired improved core value. S2: Define the list of nodes with the highest degree in subgraph G' as top, prune subgraph G' according to the degree of the nodes in G', remove nodes with small degree, and only retain the largest connected subgraph G" containing top. Calculate the priority of each node in subgraph G" based on the path length and degree between target nodes, sort the nodes according to the priority, and generate a candidate node list candidate; S3: Generate links to be connected in the candidate node list candidate, sort the links according to the calculation formula according to the node priority, generate a link list, and use the greedy algorithm to add the links with the highest priority to the original network in turn.

2. The method for improving the importance level of social network characters based on edge growth according to claim 1, characterized in that: The step S1 comprises: Assume an undirected and unweighted social network G = (V, E), where V is the set of nodes in the network and E is the set of links in the network. Use the K-core decomposition algorithm to assign a core value to each node in the network, and select a series of nodes and their links with the highest core value to form a subgraph G'. Graph G' = (V', E') is a subgraph of another graph G = (V, E). Its vertex set V' is a subset of V, and its edge set E' is a subset of E. The subgraph relationship is recorded as 3. The method for improving the importance level of social network characters based on edge growth according to claim 1, characterized in that: The step S2 comprises: S2.1: Define the node or nodes with the highest degree in the subgraph G′ as the top. Prune the nodes in the subgraph G′ layer by layer according to their degree, removing nodes with small degrees until the top only exists in the largest connected subgraph G"; S2.2: Calculate the node distance between any node i and the top node in the network. The node distance refers to the number of edges contained in the shortest path between two nodes, denoted as X i ; Normalize the node distance, the normalization formula is as follows (1): Among them, X min is the minimum value of all node distances, X max is the maximum value among all node distances; S2.3: Calculate the degree of each node, denoted as Y i Normalize the node degree. The normalization formula is as follows (2): Among them, Y min is the minimum value among all degree values, Y max is the maximum value among all degree values; S2.4: The priority of node i is calculated according to formula (3): r i =x i +y i (3) ρ i The smaller it is, the higher the priority of node i. The nodes are sorted according to the priority and a candidate node list candidate is generated.

4. The method for improving the importance level of social network characters based on edge growth according to claim 1, characterized in that: The step S3 comprises: Generate links to be connected in the candidate node list candidate, which does not include existing links in the original network. Sort the links according to the priority of the nodes. The priority is assigned in the following formula (4): Generate a link list and use the greedy algorithm to add the link with the highest priority to the original network. Note that each time a new link is added to the network, the link priority needs to be updated before the link can be added again, until the maximum core value of the original network G changes. At this point, the importance level of some nodes in the original network G has been effectively improved.

Images