CN109034562A - A kind of social networks node importance appraisal procedure and system - Google Patents

Abstract

The invention relates to a social network node importance evaluation method and system, which belongs to the technical field of social network analysis, and solves the problem that the node importance evaluation method based on the H-index or H-like index in the prior art has a low degree of differentiation of node importance, The problem of not being able to effectively sort the importance of nodes with the same H value. The method includes the following steps: obtaining the K index value of a node in a given social network; determining the importance of the node to be evaluated according to the sum of the K index values of all neighbor nodes of the node to be evaluated; determining the importance of the node to be evaluated based on the importance of the node to be evaluated The importance of evaluation nodes is evaluated. The invention makes full use of the influence of neighbor nodes, can effectively distinguish and sort the importance of nodes with the same H index value, can quickly and accurately evaluate the importance of nodes in a social network, and the evaluation results have a high degree of discrimination; at the same time , which can analyze large-scale social networks, facilitate the rapid discovery of important nodes, and has strong adaptability.

Description

A social network node importance evaluation method and system

technical field

The invention relates to the technical field of social network analysis, in particular to a method and system for evaluating the importance of social network nodes.

Background technique

In recent years, the prosperity of social software represented by WeChat, Weibo, FaceBook, and Linkedin has greatly promoted the development of social network research, among which the importance evaluation of social network nodes is an important direction of social network research. Importance assessment is of great significance for further identifying key nodes and analyzing network structure.

Existing node importance measurement methods in social networks can be divided into methods based on network local attributes, network global attributes, random walks, and community structures according to the implementation methods. Among them, the representative method based on local attribute information is the degree centrality method. The degree value of a node indicates the number of nodes connected to the node, which can intuitively reflect the local importance of the node, but cannot well reflect the node's importance in the whole The situation in the network; measures based on global attributes include betweenness centrality, compactness centrality, eigenvector centrality, etc. The time complexity of such methods is relatively high, and they are not suitable for large networks.

The H index was originally used to evaluate the influence of researchers' personal achievements. The H index corresponds to the nodes in the social network. If a node has at least h neighbor nodes with a degree of h, then the node's H index value is h. However, directly applying the H index to the evaluation of the importance of social network nodes will have the same flaws as the k-shell decomposition algorithm, and nodes with the same h value cannot be distinguished. The root cause of this defect is that the importance of a node in a social network not only depends on its own metric value, but also depends on the influence of its neighbors on the node, or the degree of dependence of the neighbors on the node. To solve this problem, H-like indices such as g-index, K-index, and w-index have emerged. Among them, the K index is further subdivided by the sum of the total degrees of neighbor nodes, but it still cannot effectively distinguish some nodes.

Contents of the invention

In view of the above analysis, the present invention aims to provide a social network node importance evaluation method and system to solve the problem that the existing node importance evaluation methods based on H-index or H-like index have low degree of distinction of node importance and cannot The problem of efficiently ranking the importance of nodes with the same H value.

The purpose of the present invention is mainly achieved through the following technical solutions:

On the one hand, a method for assessing the importance of social network nodes is provided, including the following steps:

Find the K-index value of a node in a given social network;

Determine the importance of the node to be evaluated according to the sum of the K index values of all neighbor nodes of the node to be evaluated;

The importance of the node to be evaluated is evaluated based on the importance of the node to be evaluated.

The beneficial effects of the present invention are as follows: the present invention makes full use of the connection information between social network nodes, comprehensively considers the influence of neighbor nodes, not only utilizes the influence of the node itself, but also makes full use of the influence of neighbor nodes, and can make full use of the same H index It can effectively distinguish the importance of nodes in the social network, and can quickly and accurately evaluate the importance of nodes in the social network. At the same time, the evaluation results are highly differentiated, and it can analyze large-scale social networks, which is convenient for quickly discovering important nodes. Adaptability powerful.

On the basis of the foregoing scheme, the present invention has also made the following improvements:

Further, said obtaining the K index value of the node in the social network includes the following steps:

Find the H-index value of a node in a given social network;

Select a neighbor node whose degree value is not less than the H index value of the node in the neighbor node set of the above node;

According to the sum of the degree values of the selected neighbor nodes, the K index value of the node is determined.

Further, the K index value of the node is determined according to the sum of the degree values of the selected neighbor nodes, the formula is:

In the formula, K _i represents the K index value of node i in the social network, h _i represents the H index value of node i, Indicates the sum of the degree values of the selected neighbor nodes.

Further, according to the sum of K index values of all neighbor nodes of the node to be evaluated, the importance of the node to be evaluated is determined, and the formula is:

In the formula, Γ _(i) refers to the set of neighbor nodes of node i to be evaluated, K _j is the K index value of neighbor node j, and LK _(i) is the importance of node i to be evaluated.

Further, said obtaining the H-index value of a node in a given social network includes the following steps:

Find the degree value of all nodes in the social network;

The binary search method is used to calculate the H index value of the nodes in the social network.

On the other hand, a social network node importance evaluation system is also provided, including:

The node K index value calculation module is used to obtain the K index value of the node in the given social network;

The node importance determination module is used to determine the importance of the node to be evaluated according to the sum of the K index values of all neighbor nodes of the node to be evaluated, and output to the node importance evaluation module;

The node importance evaluation module evaluates the importance of the node to be evaluated according to the received node importance.

The beneficial effects of the present invention are as follows: the present invention makes full use of the connection information between social network nodes, comprehensively considers the influence of neighbor nodes, not only utilizes the influence of the node itself, but also makes full use of the influence of neighbor nodes, and can make full use of the same H index It can effectively distinguish the importance of nodes in the social network, and can quickly and accurately evaluate the importance of nodes in the social network. At the same time, the evaluation results are highly differentiated, and it can analyze large-scale social networks, which is convenient for quickly discovering important nodes. Adaptability powerful.

On the basis of the foregoing scheme, the present invention has also made the following improvements:

Further, the node K index value calculation module includes an H index value calculation unit, a neighbor node selection unit, and a K index value determination unit;

The H-index value calculating unit is used to calculate the H-index value of a node in a given social network, and output it to a neighbor node selection unit;

The neighbor node selection unit selects a neighbor node whose degree value is not less than the H index value of the node in the neighbor node set of the above-mentioned node;

The K index value determination unit determines the K index value of the node according to the sum of the degree values of the neighbor nodes selected by the neighbor node selection unit.

Further, according to the sum of the degree values of the selected neighbor nodes, determine the K index value of the node, the formula is:

In the formula, K _i represents the K index value of node i in the social network, h _i represents the H index value of node i, Indicates the sum of the degree values of the selected neighbor nodes.

Further, according to the sum of K index values of all neighbor nodes of the node to be evaluated, the importance of the node to be evaluated is determined, and the formula is:

In the formula, Γ _(i) is the set of neighbor nodes of node i to be evaluated, K _j is the K index value of neighbor node j, and LK _(i) is the importance of node i to be evaluated.

Further, the node H index value calculation unit includes a degree value calculation subunit and an H index value calculation subunit:

The degree value calculation subunit is used to calculate the degree values of all nodes in the social network, and output to the H index value calculation subunit;

The H-index calculation subunit is used to calculate the H-index value of the nodes in the social network by using the binary search method.

In the present invention, the above technical solutions can also be combined with each other to realize more preferred combination solutions. Additional features and advantages of the invention will be set forth in the description which follows, and some of the advantages will be apparent from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by what is particularly pointed out in the written description, claims as well as the appended drawings.

Description of drawings

The drawings are for the purpose of illustrating specific embodiments only and are not to be considered as limitations of the invention, and like reference numerals refer to like parts throughout the drawings.

FIG. 1 is a flow chart of the method for evaluating the importance of social network nodes described in Embodiment 1 of the present invention.

FIG. 2 is a schematic structural diagram of the social network node importance evaluation system described in Embodiment 2 of the present invention.

FIG. 3 is a schematic diagram of a topology structure of a simple example network described in Embodiments 3 and 4 of the present invention.

FIG. 4 is a schematic diagram of the complementary cumulative distribution function CCDF of the network shown in FIG. 3 .

FIG. 5 is a schematic diagram of the complementary cumulative distribution function CCDF of the Karate club network described in Embodiment 4 of the present invention.

FIG. 6 is a schematic diagram of the complementary cumulative distribution function CCDF of the Dolphin network described in Embodiment 4 of the present invention.

FIG. 7 is a schematic diagram of the complementary cumulative distribution function CCDF of the Celegan network described in Embodiment 4 of the present invention.

FIG. 8 is a schematic diagram of changes in the discrimination index of different node importance evaluation methods when the n parameter of the LFR network generator described in Embodiment 4 of the present invention changes.

FIG. 9 is a schematic diagram of changes in the discrimination index of different node importance evaluation methods when the μ parameter of the LFR network generator described in Embodiment 4 of the present invention changes.

FIG. 10 is a schematic diagram of changes in the discrimination index of different node importance evaluation methods when the k parameter of the LFR network generator described in Embodiment 4 of the present invention changes.

FIG. 11 is a schematic diagram of the variation of the discrimination index of different node importance evaluation methods when the λ parameter of the LFR network generator described in Embodiment 4 of the present invention changes.

Detailed ways

Preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of the application and together with the embodiments of the present invention are used to explain the principle of the present invention and are not intended to limit the scope of the present invention.

Embodiment one

A social network node importance evaluation method is disclosed. As shown in Figure 1, the following steps are included:

Step S1, obtain the K index value of the node in the given social network;

Step S2. Determine the importance of the node to be evaluated according to the sum of the K index values of all neighboring nodes of the node to be evaluated;

Step S3, evaluating the importance of the node to be evaluated based on the importance of the node to be evaluated.

Compared with the prior art, the social network node importance evaluation method provided by this embodiment makes full use of the connection information between social network nodes and comprehensively considers the influence of neighbor nodes, not only using the influence of the node itself, but also making full use of It can effectively distinguish the importance of nodes with the same H index value, and can quickly and accurately evaluate the importance of nodes in social networks. At the same time, the evaluation results are highly differentiated, and can be used for large-scale social Network analysis facilitates quick discovery of important nodes and strong adaptability.

Specifically, in step S1, calculating the K index value of the nodes in the social network includes the following steps:

Step S101, obtaining the H-index value of a node in a given social network;

First calculate the degree value of all nodes in a given social network; then according to the definition of the H index, calculate the H index value h of the node in the social network according to the obtained degree value (the H index value h of the node indicates that the node has at least The degrees of h neighboring nodes are all greater than h). Preferably, the H-index value h is calculated using a quadratic search method.

Step S102, selecting a neighbor node whose degree value is not less than the H index value of the node in the neighbor node set of the node;

Step S103, according to the sum of the degree values of the neighbor nodes selected in step S102, determine the K index value of the node.

Among them, the formula for calculating the K index value of a node is:

In the formula, K _i represents the K index value of node i in the social network, h _i represents the H index value of node i, Indicates the sum of the degree values of the selected neighbor nodes.

In step S2, the importance of the node to be evaluated is determined according to the sum of the K index values of all neighbor nodes of the node to be evaluated, the formula is:

In the formula, Γ _(i) refers to the set of neighbor nodes of node i to be evaluated, K _j is the K index value of neighbor node j, and LK _(i) is the importance of node i to be evaluated.

In step S3, the importance of the node is evaluated based on the determined importance of the node. After obtaining the importance of all nodes in the social network, all nodes are sorted according to the importance, and then the importance of the node is evaluated (the greater the value of the importance, the higher the importance).

Embodiment two

A social network node importance evaluation system is disclosed. As shown in Figure 2, it includes: a node K index value calculation module, a node importance determination module, and a node importance evaluation module; wherein,

The node K index value calculation module is used to obtain the K index value of the node in the given social network;

The node importance determination module is used to determine the importance of the node to be evaluated according to the sum of the K index values of all neighbor nodes of the node to be evaluated, and output to the node importance evaluation module;

The node importance evaluation module evaluates the importance of the node to be evaluated according to the received node importance.

Compared with the prior art, the social network node importance evaluation system provided by this embodiment makes full use of the connection information between social network nodes and comprehensively considers the influence of neighbor nodes, not only using the influence of the node itself, but also making full use of It can effectively distinguish the importance of nodes with the same H index value, and can quickly and accurately evaluate the importance of nodes in social networks. At the same time, the evaluation results are highly differentiated, and can be used for large-scale social networks. It is easy to quickly find important nodes through analysis, and has strong adaptability.

Specifically, the node K index value calculation module includes an H index value calculation unit, a neighbor node selection unit, and a K index value determination unit; wherein,

The H index value calculation unit is used to obtain the H index value of the node in the given social network, and output to the neighbor node selection unit;

It should be noted that the H-index value calculation unit includes a degree value calculation subunit and an H-index value calculation subunit: the degree value calculation subunit is used to calculate the degree values of all nodes in the social network and output them to H The subunit for calculating the index value; the subunit for calculating the H index value is used to calculate the H index value of the node in the social network by using the binary search method.

The neighbor node selection unit selects the neighbor node whose degree value is not less than the H index value of the node in the neighbor node set of the above-mentioned node;

The K index value determination unit determines the K index value of the node according to the sum of the degree values of the neighbor nodes selected by the neighbor node selection unit.

It should be emphasized that in the K-index value determination unit, the formula for calculating the K-index value of a node is:

In the formula, K _i represents the K index value of node i in the social network, h _i represents the H index value of node i, Indicates the sum of the degree values of the selected neighbor nodes.

The node importance determination module is used to determine the importance of the node to be evaluated, and output to the node importance evaluation module; specifically, according to the sum of the K index values of all neighbor nodes of the node to be evaluated, determine the importance of the node to be evaluated , the formula is:

In the formula, Γ _(i) refers to the set of neighbor nodes of node i to be evaluated, K _j is the K index value of neighbor node j, and LK _(i) is the importance of node i to be evaluated.

The node importance evaluation module is used to obtain the importance degree of the node of the node importance degree determination module, and evaluate the importance of the node. After obtaining the importance of all nodes in the social network, all nodes are sorted according to the importance, and then the importance of the node is evaluated (the greater the value of the importance, the higher the importance).

Embodiment Three

In this embodiment, a simple example network is taken as an example, and the social network node importance evaluation method in Embodiment 1 is used to evaluate the importance of the network node. The topology of the example network is shown in Figure 3, which contains 17 nodes and 21 edges. Specifically include the following steps:

1) For a given example network, according to the definition of H-index, the binary search method is used to calculate the H-index value of nodes in the social network, as shown in Table 1.

Table 1: H-index values for example network nodes

node number H index value node number H index value 1 1 10 3 2 1 11 3 3 1 12 3 4 1 13 3 5 1 14 1 6 2 15 2 7 3 16 1 8 2 17 1 9 1

2) Select the neighbor node whose degree value is not less than the H index value of the node in the neighbor node set of the node; and calculate the sum of the degree values of the selected neighbor nodes to obtain the K index value of the node. See Table 2 for the K-index values of example network nodes.

Table 2: K-index values for example network nodes

node number K index value node number K index value 1 1.75 10 3.47 2 1.67 11 3.36 3 1.67 12 3.44 4 1.83 13 3.44 5 1.75 14 1.83 6 2.33 15 2.60 7 3.31 16 1.75 8 2.56 17 1.50 9 1.67

3) Calculate the sum of the K index values of all neighbor nodes of the node as the importance of the node. According to the node importance calculation formula described in Embodiment 1 of the present invention, the calculated importance of an example network node is shown in Table 3.

Table 3: Importance of example network nodes

node number Importance node number Importance 1 2.33 10 17.93 2 1.83 11 10.35 3 1.83 12 12.87 4 5.67 13 12.87 5 2.33 14 3.47 6 8.64 15 8.63 7 8.36 16 4.10 8 8.44 17 1.75 9 2.56

It can be seen from Table 3 that almost all example network nodes are given different degrees of importance. The importance of the example network nodes is well distinguished, which in turn makes the importance evaluation results more differentiated.

Embodiment four

In this embodiment, real networks and artificial networks are taken as examples, and the social network node importance evaluation method described in Embodiment 1 is used for node importance evaluation of the above network, and compared with other existing node importance evaluation methods. Typical methods selected include: H index (abbreviated as H), K index (abbreviated as K), PageRank algorithm (abbreviated as PR), classic K kernel decomposition algorithm (abbreviated as KS), KS-IF algorithm (abbreviated as KSIF), MDD algorithm (abbreviated as MDD), the method of the present invention (abbreviated as LK). In order to better evaluate the performance of various importance evaluation methods, the discrimination index M is introduced here. The discrimination index is defined as follows:

Among them, R is the level vector of the importance of network nodes, n is the total number of levels of the vector R, and n _r is the number of nodes in the rth level. If all nodes are in the same importance level, the value of the discrimination index M is 0, and the corresponding evaluation method cannot distinguish the importance of each node. If there is only one node in each importance level, and the value of the discrimination index M is 1, the corresponding evaluation method can effectively distinguish the importance of each node and has the strongest discrimination ability.

First, select the example network shown in Example 3 (Figure 3), use the above seven methods to evaluate the importance of the example network nodes, and sort the nodes according to the importance, the sorting results are shown in Table 4 (Table 4 Each column of corresponds to an importance evaluation method, nodes of the same level have the same importance, and "other" means all remaining nodes). As can be seen from Table 4, compared with the existing 6 typical methods, the method disclosed in the present invention can accurately and finely distinguish the importance of network nodes. Exemplarily, the number of nodes in each importance level is at most 2 indivual.

Table 4: Ranking results of node importance in an example network

In order to further illustrate the beneficial effect of the inventive method, choose 11 real networks of different scales (comprising: Karate Club network, Dolphin network, Jazz network, Prison network, NetScience network, Book network, Celegan network, E-mail network, Blogs network , PGP network and Enron network), analyze and compare the discrimination index M of the above seven importance evaluation methods. Table 5 shows the discriminative ability of 7 importance evaluation methods on the importance of 11 real network nodes. It can be seen that for the selected 11 real networks, the methods disclosed in the present invention can all obtain the maximum discrimination value. Compared with the other six node importance evaluation methods, the method of the present invention can more carefully and accurately identify the importance of real network nodes.

Table 5: Discrimination ability of different importance evaluation methods on the importance of real network nodes

network name node Number of sides M(PR) M(KS) M(H) M(K) M(KSIF) M (MDD) M(LK) Karate Club 34 78 0.9542 0.4958 0.5766 0.9542 0.9542 0.7536 0.9542 Dolphins 62 159 0.9979 0.3769 0.6841 0.9748 0.9979 0.9041 0.9979 prison 67 182 0.9964 0.3070 0.6031 0.9722 0.9928 0.8672 0.9964 book 105 441 1.0000 0.4949 0.7067 0.9952 1.0000 0.9077 1.0000 Football 115 613 1.0000 0.0003 0.2349 0.9316 0.9991 0.6089 1.0000 Jazz 198 2742 0.9993 0.7944 0.9383 0.9990 0.9993 0.9882 0.9993 Celegan 379 914 0.9951 0.6421 0.6825 0.9848 0.9944 0.8748 0.9950 NetScience 453 2025 0.9992 0.6962 0.7311 0.9959 0.9975 0.8215 0.9983 E-mail 1133 10903 0.9999 0.8088 0.8583 0.9979 0.9996 0.9229 0.9999 Blogs 1490 16718 0.9993 0.9058 0.9264 0.9991 0.9992 0.9443 0.9993 PGP 10680 24316 0.9997 0.4806 0.5172 0.9942 0.9935 0.6678 0.9981

In order to demonstrate the beneficial effect of the method of the present invention more intuitively, the data in Table 5 is displayed by using a complementary cumulative distribution function (CCDF). Figures 4 to 7 respectively show the CCDFs of the four networks (the network in Embodiment 3, the Karate Club network, the Dolphin network and the Celegan network). According to the principle of CCDF, if the number of nodes at the same importance level is more, the CCDF will decrease faster, otherwise, CCDF will decrease slowly along the diagonal. It can be seen from Fig. 4 to Fig. 7 that the CCDF of the method of the present invention decreases slowly along the diagonal line, indicating that the method of the present invention can well distinguish the importance differences between nodes in the social network.

In addition, the artificial social network is generated by means of the LFR network generator, and the method of the present invention is evaluated by using the artificial social network. The LFR network generator has four important parameters, which are node size n (number of nodes), average node degree k (average degree of nodes), community structure mixing parameter μ (mixing parameter of community structure) and degree power law distribution λ ( power-law of degree distribution). The changes of the above four parameters will affect the topology of the artificial social network. Figures 8 to 11 respectively show the changes of the discrimination index M of different node importance evaluation methods when one of the four parameters is kept changing and the other three parameters remain unchanged. It can be seen that: for the artificial social network, the method of the present invention can obtain the maximum discrimination value. It shows that compared with other five node importance evaluation methods (H, K, KS, KS-IF, MDD), the method of the present invention can identify the node importance of artificial social network more carefully and accurately.

Those skilled in the art can understand that all or part of the processes of the methods in the above embodiments can be implemented by computer program instructions related hardware, and the program can be stored in a computer-readable storage medium. Wherein, the computer-readable storage medium is a magnetic disk, an optical disk, a read-only memory or a random access memory, and the like.

The above is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present invention can easily think of changes or Replacement should be covered within the protection scope of the present invention.

Claims (10)

1. A social network node importance evaluation method is characterized by comprising the following steps:

solving a K index value of a node in a given social network;

determining the importance of the node to be evaluated according to the sum of the K index values of all the neighbor nodes of the node to be evaluated;

and evaluating the importance of the node to be evaluated based on the importance of the node to be evaluated.

2. The method of claim 1, wherein said deriving a K-index value for a node in a given social network comprises:

solving an H index value of a node in a given social network;

selecting neighbor nodes with the values not less than the H index value of the node from the neighbor node set of the node;

and determining the K index value of the node according to the sum of the values of the selected neighbor nodes.

3. The method according to claim 2, wherein the K index value of the selected neighbor node is determined according to the sum of the values of the selected neighbor node, and the formula is as follows:

in the formula, K_iK index value, h, representing node i in social network_iRepresents the H index value of the node i,representing the sum of the values of the selected neighbor nodes.

4. The method according to one of claims 1 to 3, characterized in that the importance of the node to be evaluated is determined according to the sum of the K index values of all the neighbor nodes of the node to be evaluated, and the formula is:

in the formula, gamma_(i)Set of neighbor nodes, K, for node i to be evaluated_jIs the K index value, LK, of the neighbor node j_(i)Is the importance of the node i to be evaluated.

5. The method of claim 4, wherein said deriving the H index value for a node in a given social network comprises the steps of:

solving the values of all nodes in the social network;

and calculating the H index value of the node in the social network by adopting a binary search method.

6. A social network node importance evaluation system, comprising:

the node K index value obtaining module is used for obtaining a K index value of a node in a given social network;

the node importance determining module is used for determining the importance of the node to be evaluated according to the sum of the K index values of all the neighbor nodes of the node to be evaluated and outputting the importance to the node importance evaluating module;

and the node importance evaluation module evaluates the importance of the node to be evaluated according to the received importance of the node.

7. The system according to claim 6, wherein said node K index value obtaining module comprises an H index value obtaining unit, a neighbor node selecting unit, and a K index value determining unit;

the H index value obtaining unit is used for obtaining the H index value of the node in the given social network and outputting the H index value to the neighbor node selecting unit;

the neighbor node selection unit selects neighbor nodes with the values not less than the H index value of the node from the neighbor node set of the node;

and the K index value determining unit determines the K index value of the node according to the sum of the values of the neighbor nodes selected by the neighbor node selecting unit.

8. The system of claim 7, wherein the K index value of the selected neighbor node is determined according to the sum of the values of the selected neighbor node, and the formula is:

in the formula, K_iK index value, h, representing node i in social network_iRepresents the H index value of the node i,representing the sum of the values of the selected neighbor nodes.

9. The system according to any of claims 6-8, wherein the importance of the node to be evaluated is determined according to the sum of the K index values of all neighboring nodes of the node to be evaluated, and the formula is:

in the formula, gamma_(i)A set of neighbor nodes, K, of a node i to be evaluated_jIs the K index value, LK, of the neighbor node j_(i)Is the importance of the node i to be evaluated.

10. The system according to claim 9, wherein said node H-exponent value evaluation unit comprises a value evaluation subunit, an H-exponent value evaluation subunit:

the value solving subunit is used for solving the values of all nodes in the social network and outputting the values to the H index value solving subunit;

and the H index value obtaining subunit is used for calculating the H index value of the node in the social network by adopting a binary search method.