CN106055568A - Automatic friend grouping method for social network based on single-step association adding - Google Patents

Abstract

The invention discloses an automatic friend grouping method for a social network based on single-step association adding. The method is characterized in that the social network is described as a binary group; and an overlapped association detection algorithm based on the single-step association adding is used to solve problems in automatic friend grouping of the social network. The method disclosed by the invention has the advantages that friend circles in the social network can be grouped automatically and rapidly; and efficiency and accuracy of the grouping are increased. In this way, more reliable friends can be recommended for a user through the accurate friend grouping; and unnecessary troubles generated when the user searches the friends in the same camp can be reduced.

Description

A method for automatic grouping of friends in a social network based on single-step group addition

technical field

The invention relates to the technical field of automatic grouping of friends in a social network, and specifically proposes a method for automatic grouping of friends in a social network based on single-step group addition.

Background technique

With the advent of the Internet age, social networks have grown rapidly like mushrooms after rain, attracting extensive attention from scholars at home and abroad. The emergence of social networks has changed people's traditional thinking, communication, work and lifestyle. More and more users choose to join online social networks for information exchange. speed. The intensification of network information makes it impossible for users to choose the information they need from the massive data, that is, users face the problem of information overload. If users manually group friends in their online social network, it is not only time-consuming and laborious, but many users do not classify their friends. Therefore, most social networking sites automatically classify users according to some public attributes of users. friend.

At present, the automatic grouping problem in social network friend recommendation contains two types of research methods:

One type of research method is based on the global characteristics of the social network. This type of method needs to detect all the link structures in the social network and obtain the global information of the network. However, such methods are computationally expensive and may affect the efficiency of recommendation algorithms in complex social networks.

One type of research method is based on the local structural characteristics of social networks. This type of method takes the structure of nodes in the network as the research object to realize the friend grouping problem in friend recommendation. This type of method has low computational complexity, but it usually results in inaccurate grouping of friends in social networks due to insufficient information given.

Contents of the invention

Aiming at the deficiencies in the prior art, the present invention provides a method for automatically grouping friends in social networks based on single-step group addition, in order to quickly realize automatic grouping of circles of friends in social networks, improve grouping efficiency and accuracy, and thereby Through accurate friend grouping, more reliable friends can be recommended for users, reducing unnecessary troubles for users when searching for like-minded friends.

The present invention adopts following technical scheme for solving technical problems:

A kind of friend automatic grouping method based on the social network of single-step adding group of the present invention is to carry out according to the following steps:

Step 1. Define the social network representation as a binary group {V, E}, V={v ₁ , v ₂ ,...,v _i ,...,v _n } represents the set of all users in the social network, v _i represents the i-th user; n is the total number of users; E={e _ij |i=1,2,...,n; j=1,2,...,n} represents the set of links between any two users ; e _ij represents the connection between the i-th user v _i and the j-th user v _j ; if e _ij = 1, it means that there is an edge connection between the i-th user v _i and the j-th user v _j , and the i-th user The first user v _i and the jth user v _j are mutually called neighbor users; if e _ij = 0, it means that there is no boundless connection between the i-th user v _i and the j-th user v _j , that is, there is no connection;

Step 2. Divide the social network G into k grouping sets by adopting an overlapping network community detection algorithm based on single-step adding groups, denoted as C={C ₁ ,C ₂ ,...,C _x ,...,C _X }; C _α represents the xth group; x=1,2,...,X; thereby realizing the automatic grouping of friends in the social network.

The feature of the automatic grouping method of friends based on the social network of single-step adding groups according to the present invention is also that,

The overlapping network community detection algorithm based on single-step added groups in the step 2 is carried out as follows:

Step 0, initialization x=1; With the set V of all users of the social network as the candidate user set of the xth preliminary group, denoted as V ^(x) ; Define iteration variable t; Define iteration variable r;

Step 1. Initialize t=1;

Step 2, randomly selecting the user v ^(x)(t) of the tth iteration from the candidate user V ^(x) of the xth preliminary group;

Step 3. Calculate the clustering coefficient G ^(x)(t) of the user v ^(x)(t) of the t-th iteration by using formula (1):

${\mathrm{<span\; class="notranslate">\; G</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 2</span>}{\mathrm{<span\; class="notranslate">\; m</span>}}_{{\mathrm{<span\; class="notranslate">\; v</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}}}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{{\mathrm{<span\; class="notranslate">\; v</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{{\mathrm{<span\; class="notranslate">\; v</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

In formula (1), Denotes the number of neighbor users of user v ^(x)(t) at iteration t in the preparation group x, ^Denotes all The number of connected edges between neighboring users;

Step 4. Use formula (1) to calculate the user v ^(x)(t) of the t-th iteration in the x-th preliminary group The clustering coefficients of neighboring users are compared with the clustering coefficient G ^(x)(t) of user v ^(x)(t) in the t-th iteration, if the largest clustering coefficient and the largest clustering can be found The neighbor user whose coefficient is greater than G ^{(x) (t)} , then the neighbor user found is used as the user v ^{(x) (t+1)} of the t+1 iteration in the xth preparation group; and perform step 5; Otherwise, retain the user v ^(x)(t) of the tth iteration as the core user in the xth preliminary group;

Step 5. Assign t+1 to t; and repeat step 3;

Step 6, initialize r=1;

Step 7, find a k group connected by the core user v ^(x)(t) of the tth iteration in the xth preliminary group; all the k groups connected by the core user v ^{(x) (t)} The users are all neighbor users; the k group connected with the core user v ^{(x) (t)} is used as the community C ^{(x) (r)} of the rth iteration in the xth preliminary group; k≥3;

Step 8. Calculate the community fitness value F ^(x)(r) of the community C ^(x)(r) of the r-th iteration in the x-th preliminary group according to formula (2):

${\mathrm{<span\; class="notranslate">\; f</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; )</span>}}{{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; u</span>}\mathrm{<span\; class="notranslate">\; t</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

In formula (2), α is an adjustable parameter; respectively represent the internal degree and external degree of the initial community C ^(x)(r) in the rth iteration, and the internal degree Twice the number of connected edges between all users in the community C ^(x)(r) of the rth iteration; the external degree is the number of edges connecting all users and external users in the community C ^(x)(r) of the rth iteration;

Step 9. Find the neighbor users of all users in the community C ^(x)(r) of the r-th iteration and perform union processing to form the community C ^(x)(r) of the r-th iteration in the x-th preliminary group neighbor users;

Step 10, judging whether the neighbor users of the community C ^(x)(r) of the r-th iteration in the x-th preparation group have a k group connected to themselves; A k group of the r-th iteration is added to the community C ^(x)(r) of the r-th iteration to calculate the corresponding community fitness value; if it does not exist, the corresponding neighbor user is directly added to the r-th iteration of the community C ⁽ Calculate the corresponding community fitness value in ^x)(r) ;

Step 11. If it is possible to find the neighbor user or the k clique connected to the neighbor user that makes the community fitness value F ^(x)(r) increase the most; then add the neighbor user or the k clique connected to the neighbor user with the largest increase to the initial community C ^(x)(r) of the rth iteration, and as the community C ^(x)(r+1) of the r+1th iteration, perform step 12; otherwise, keep the xth preliminary The initial community C ^(x)(r) of the rth iteration in the group is used as the xth grouping; and step 13 is performed;

Step 12, assign r+1 to r; and repeat step 8;

Step 13, removing the xth group from the candidate user V ^(x) of the xth preliminary group, thereby obtaining the candidate user V ^(x+1) of the x+1th preliminary group;

Step 14. Assign x+1 to x; and return to step 1 for execution, so as to obtain X groups, and then realize the automatic grouping of friends in social networks.

Compared with the prior art, the beneficial effects of the present invention are reflected in:

1. According to the idea of local community expansion, the present invention adopts a community expansion method based on single-step adding groups when automatically grouping friends in a social network to complete the automatic grouping of social networks. In the community expansion process, the present invention only needs to obtain the local information of users in the social network to complete the automatic grouping of users without knowing the global information and node information of the entire social network, so that each friend group of the social network can be quickly detected, and social More accurate friend grouping results in the network; it solves the problem of high computational cost based on global structural feature methods, and also solves the problem of inaccurate network division in some methods based on local structural features.

2. When performing automatic grouping of friends, the initial friend grouping must first be selected, and the initial friend grouping of the present invention is obtained through core node search. As a group or the most special node in the graph, the core node is the central node of the group and plays a key role in the connection of users in the group, which can avoid the randomness caused by the inaccurate selection of the initial group to a certain extent, thus Improved the accuracy of automatic grouping of social network friends.

3. In the current method based on local expansion, only one user node is added each time in the friend grouping process, and the local information of the user node is not fully considered, which will lead to inaccurate grouping. The present invention refers to the concept of clique: a clique is a fully connected subgraph, and they are very likely to belong to the same group during the group division process, so if a user node belongs to a certain group, then the clique where the user node is located is also more likely Be part of this group. When the present invention adds a group each time, it not only considers the connection strength between the added user point and the existing group, but also better considers the internal connection situation of the added user, and can more accurately find the group belonging to multiple groups at the same time. user.

Description of drawings

Fig. 1 is the algorithm flowchart of the present invention;

Fig. 2 is a schematic diagram of a simple network example of the present invention.

detailed description

In an embodiment, a method for automatically grouping friends in a social network based on one-step group addition converts the problem of automatic grouping of friends in a social network into a complex network community detection problem, by using a complex network community detection algorithm based on one-step group addition Solve the problem of automatic grouping of friends in the social network, so as to obtain the grouping of friends in the social network; specifically, proceed as follows:

Step 1. Define the social network representation as a binary group {V, E}, V={v ₁ , v ₂ ,...,v _i ,...,v _n } represents the set of all users in the social network, v _i represents the i-th user; n is the total number of users; E={e _ij |i=1,2,...,n; j=1,2,...,n} represents the set of links between any two users ; e _ij represents the connection between the i-th user v _i and the j-th user v _j ; if e _ij = 1, it means that there is an edge connection between the i-th user v _i and the j-th user v _j , and the i-th user A user v _i and a jth user v _j are called neighbor users; if e _ij = 0, it means that there is no boundless connection between the i-th user v _i and the j-th user v _j , that is, there is no connection; as shown in 2 shows a simple social network structure diagram containing 11 users, where each node represents a user in the social network, and each edge represents a connection between users;

Step 2. Divide the social network G into k grouping sets by adopting an overlapping network community detection algorithm based on single-step adding groups, denoted as C={C ₁ ,C ₂ ,...,C _x ,...,C _X }; C _α represents the xth group; x=1,2,...,X; thereby realizing the automatic grouping of friends in the social network.

Specifically, as shown in Figure 1, the overlapping network community detection algorithm based on single-step addition of cliques is carried out as follows:

Step 0, initialization x=1; With the set V of all users of the social network as the candidate user set of the xth preliminary group, denoted as V ^(x) ; Define iteration variable t; Define iteration variable r;

Step 1. Initialize t=1;

Step 2, randomly selecting the user v ^(x)(t) of the tth iteration from the candidate user V ^(x) of the xth preliminary group;

Step 3. Calculate the clustering coefficient G ^(x)(t) of the user v ^(x)(t) of the t-th iteration by using formula (1):

${\mathrm{<span\; class="notranslate">\; G</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 2</span>}{\mathrm{<span\; class="notranslate">\; m</span>}}_{{\mathrm{<span\; class="notranslate">\; v</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}}}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{{\mathrm{<span\; class="notranslate">\; v</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{{\mathrm{<span\; class="notranslate">\; v</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

In formula (1), Denotes the number of neighbor users of user v ^(x)(t) at iteration t in the preparation group x, ^Denotes all The number of connected edges between neighboring users;

Step 4. Use formula (1) to calculate the user v ^(x)(t) of the t-th iteration in the x-th preliminary group The clustering coefficients of neighboring users are compared with the clustering coefficient G ^(x)(t) of user v ^(x)(t) in the t-th iteration, if the largest clustering coefficient and the largest clustering can be found The neighbor user whose coefficient is greater than G ^{(x) (t)} , then the neighbor user found is used as the user v ^{(x) (t+1)} of the t+1 iteration in the xth preparation group; and perform step 5; Otherwise, retain the user v ^(x)(t) of the tth iteration as the core user in the xth preliminary group;

Step 5. Assign t+1 to t; and repeat step 3;

Step 6, initialize r=1;

Step 7, find a k group connected by the core user v ^(x)(t) of the tth iteration in the xth preliminary group; all the k groups connected by the core user v ^{(x) (t)} The users are all neighbor users; the k group connected by the core user v ^{(x) (t)} is used as the community C ^{(x) (r)} of the rth iteration in the xth preliminary group; k≥3; for example, in In Fig. 2, it is assumed that the initial preparation group is composed of user nodes {1, 2, 3, 4, 5}, at this time, k=3 is set;

Step 8. Calculate the community fitness value F ^{(x )(r)} of the community C ^(x)(r) of the r-th iteration in the x-th preliminary group according to formula (2):

${\mathrm{<span\; class="notranslate">\; f</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; )</span>}}{{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; u</span>}\mathrm{<span\; class="notranslate">\; t</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

In formula (2), α is an adjustable parameter; respectively represent the internal degree and external degree of the initial community C ^(x)(r) in the rth iteration, and the internal degree Twice the number of connected edges between all users in the community C ^(x)(r) of the rth iteration; the external degree is the number of edges connecting all users and external users in the community C ^(x)(r) of the r-th iteration, and calculate the first preliminary group {1,2,3,4,5} according to formula (2) The fitness value of is 0.824;

Step 9. Find the neighbor users of all users in the community C ^(x)(r) of the r-th iteration and perform union processing to form the community C ^(x)(r) of the r-th iteration in the x-th preliminary group Neighbor users, so we know that the neighbor users of the first preliminary group {1,2,3,4,5} are {6,7,8};

Step 10, judging whether the neighbor users of the community C ^(x)(r) of the r-th iteration in the x-th preparation group have a k group connected to themselves; A k group of the r-th iteration is added to the community C ^(x)(r) of the r-th iteration to calculate the corresponding community fitness value; if it does not exist, the corresponding neighbor user is directly added to the r-th iteration of the community C ⁽ Calculate the corresponding community fitness value in ^x)(r) ; then in this embodiment, k The groups are {(6,7,8), (6,7,8), (8,9,11)};

Step 11. If it is possible to find the neighbor user or the k clique connected to the neighbor user that makes the community fitness value F ^(x)(r) increase the most; then add the neighbor user or the k clique connected to the neighbor user with the largest increase to the initial community C ^(x)(r) of the rth iteration, and as the community C ^(x)(r+1) of the r+1th iteration, perform step 12; otherwise, keep the xth preliminary The initial community C ^(x)(r) of the rth iteration in the group is used as the xth group; and step 13 is executed; in this embodiment, it can be known through calculation that the k group where the No. 6 user belongs to join the first preliminary group {1 ,2,3,4,5} The fitness value of the group {1,2,3,4,5,6,7,8} obtained is 0.897, which makes the fitness value of the group increase the most, so choose user 6 The connected k group is added to the preliminary group to get a new group {1, 2, 3, 4, 5, 6, 7, 8}, if adding a new user or the k group connected by the user will not make the group adapt The degree value increases, so the first group of this social network is {1,2,3,4,5,6,7,8};

Step 12, assign r+1 to r; and repeat step 8;

Step 13, removing the xth group from the candidate user V ^(x) of the xth preliminary group, thereby obtaining the candidate user V ^(x+1) of the x+1th preliminary group, at this time, Fig. 2 The candidate users in the example are {9,10,11,12};

Step 14, assign x+1 to x; and return to step 1 for execution, thereby obtaining X groups, and then realize the automatic grouping of friends in social networks. In the example shown in Figure 2, the automatic grouping of friends obtained by applying the present invention is { 1,2,3,4,5,6,7,8} and {9,10,11,12}.

Claims (2)

1. a kind of friend automatic grouping method based on the social network of single-step adding group, it is characterized in that carry out as follows: Step 1. Define the social network representation as a binary group {V, E}, V={v ₁ , v ₂ ,...,v _i ,...,v _n } represents the set of all users in the social network, v _i represents the i-th user; n is the total number of users; E={e _ij |i=1,2,...,n; j=1,2,...,n} represents the set of links between any two users ; e _ij represents the connection between the i-th user v _i and the j-th user v _j ; if e _ij = 1, it means that there is an edge connection between the i-th user v _i and the j-th user v _j , and the i-th user The first user v _i and the jth user v _j are mutually called neighbor users; if e _ij = 0, it means that there is no boundless connection between the i-th user v _i and the j-th user v _j , that is, there is no connection; Step 2. Divide the social network G into k grouping sets by adopting an overlapping network community detection algorithm based on single-step adding groups, denoted as C={C ₁ ,C ₂ ,...,C _x ,...,C _X }; C _α represents the xth group; x=1,2,...,X; thereby realizing the automatic grouping of friends in the social network. 2. the friend automatic grouping method based on the social network of single-step adding group according to claim 1, it is characterized in that: the overlapping network community detection algorithm based on single-step adding group in described step 2 is to carry out as follows: Step 0, initialization x=1; With the set V of all users of the social network as the candidate user set of the xth preliminary group, denoted as V ^(x) ; Define iteration variable t; Define iteration variable r; Step 1. Initialize t=1; Step 2, randomly selecting the user v ^(x)(t) of the tth iteration from the candidate user V ^(x) of the xth preliminary group; Step 3. Calculate the clustering coefficient G ^(x)(t) of the user v ^(x)(t) of the t-th iteration by using formula (1): ${\mathrm{<span\; class="notranslate">\; G</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 2</span>}{\mathrm{<span\; class="notranslate">\; m</span>}}_{{\mathrm{<span\; class="notranslate">\; v</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}}}{{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{{\mathrm{<span\; class="notranslate">\; v</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; the\; s</span>}}_{{\mathrm{<span\; class="notranslate">\; v</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$ In formula (1), Denotes the number of neighbor users of user v ^(x)(t) at iteration t in the preparation group x, ^Denotes all The number of connected edges between neighboring users; Step 4. Use formula (1) to calculate the user v ^(x)(t) of the t-th iteration in the x-th preliminary group The clustering coefficients of neighboring users are compared with the clustering coefficient G ^(x)(t) of user v ^(x)(t) in the t-th iteration, if the largest clustering coefficient and the largest clustering can be found The neighbor user whose coefficient is greater than G ^{(x) (t)} , then the neighbor user found is used as the user v ^{(x) (t+1)} of the t+1 iteration in the xth preparation group; and perform step 5; Otherwise, retain the user v ^(x)(t) of the tth iteration as the core user in the xth preliminary group; Step 5. Assign t+1 to t; and repeat step 3; Step 6, initialize r=1; Step 7, find a k group connected by the core user v ^(x)(t) of the tth iteration in the xth preliminary group; all the k groups connected by the core user v ^{(x) (t)} The users are all neighbor users; the k group connected with the core user v ^{(x) (t)} is used as the community C ^{(x) (r)} of the rth iteration in the xth preliminary group; k≥3; Step 8. Calculate the community fitness value F ^(x)(r) of the community C ^(x)(r) of the r-th iteration in the x-th preliminary group according to formula (2): ${\mathrm{<span\; class="notranslate">\; f</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; )</span>}}{{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; +</span>{\mathrm{<span\; class="notranslate">\; K</span>}}_{\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; u</span>}\mathrm{<span\; class="notranslate">\; t</span>}}^{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; \&alpha;</span>}}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$ In formula (2), α is an adjustable parameter; respectively represent the internal degree and external degree of the initial community C ^(x)(r) in the rth iteration, and the internal degree Twice the number of connected edges between all users in the community C ^(x)(r) of the rth iteration; the external degree is the number of edges connecting all users and external users in the community C ^(x)(r) of the rth iteration; Step 9. Find the neighbor users of all users in the community C ^(x)(r) of the r-th iteration and perform union processing to form the community C ^(x)(r) of the r-th iteration in the x-th preliminary group neighbor users; Step 10, judging whether the neighbor users of the community C ^(x)(r) of the r-th iteration in the x-th preparation group have a k group connected to themselves; A k group of the r-th iteration is added to the community C ^(x)(r) of the r-th iteration to calculate the corresponding community fitness value; if it does not exist, the corresponding neighbor user is directly added to the r-th iteration of the community C ⁽ Calculate the corresponding community fitness value in ^x)(r) ; Step 11. If it is possible to find the neighbor user or the k clique connected to the neighbor user that makes the community fitness value F ^(x)(r) increase the most; then add the neighbor user or the k clique connected to the neighbor user with the largest increase to the initial community C ^(x)(r) of the rth iteration, and as the community C ^(x)(r+1) of the r+1th iteration, perform step 12; otherwise, keep the xth preliminary The initial community C ^(x)(r) of the rth iteration in the group is used as the xth grouping; and step 13 is performed; Step 12, assign r+1 to r; and repeat step 8; Step 13, removing the xth group from the candidate user V ^(x) of the xth preliminary group, thereby obtaining the candidate user V ^(x+1) of the x+1th preliminary group; Step 14. Assign x+1 to x; and return to step 1 for execution, so as to obtain X groups, and then realize the automatic grouping of friends in social networks.