Adjustable overlapped community discovery method based on label propagation

Abstract

The invention discloses an adjustable overlapped community discovery method based on label propagation. The method comprises the following steps: 1) inputting an adjustable parameter according to user needs, and selecting an overlapped community detection mode or a non-overlapped community detection mode according to the input value of the parameter; 2) initializing all of node labels; 3) carrying out label propagation; and 4) judging whether reaching a termination condition, if so, outputting a result and if not, returning to the step 3) for iteration propagation. The method can carry out judgment according to the parameter input by the user, and can adopts two different community detection modes, that is, the overlapped community detection mode and the non-overlapped community detection mode, can select the different community detection modes according to the user needs, and has a certain flexibility and generality; and the user can input one adjustable parameter as needed, wherein the larger the parameter is, the smaller the community overlapping degree is, or otherwise, the smaller the parameter is, the larger the community overlapping degree is.

Description

technical field

[0001] The invention relates to the technical field of social networks, in particular to an adjustable overlapping community discovery method based on tag propagation. Background technique

[0002] With the continuous development of computer technology and the Internet, devices such as computers, smartphones, and sensors generate data all the time. Ordinary people have more and more opportunities to access the Internet, and all walks of life are also taking advantage of the convenience brought by the Internet. These data grow exponentially, resulting in massive amounts of data. Most of the various information in life and production can be transformed into a network structure or a graph (Graph) structure, that is to say, network structures are everywhere. Newman summarized the networks in real life into four categories; communication networks, such as the World Wide Web of Information, the collection retrieval network of the University of Electronic Science ...

Examples

Embodiment 1

[0056] An example is given below to describe the label propagation process in detail. figure 2 is an intermediate state of an example network in the label propagation process, figure 2 The shown network contains 21 nodes and 21 edges, and the input parameter K=0.6. Propagate labels in a certain order, (7,17,18,3,2,7,11…). The label propagation of node 7 is performed first, and the labels contained in the neighbor nodes of node 7 are {C,A,A,A,A,A,A,B}. The label with the largest number is A, the number is 6; the number of labels B and C are both 1. because Where |B| represents the number of B labels, and |A| is the same, so label B does not meet the conditions and cannot be added to the label group of node 7. Therefore, it cannot join the label group of node 7 either. The label of node 7 is updated to A, such as image 3 shown. Continue to execute the label propagation of the node 17 in a specific order, and delete the labels in the label group of the node 17. The l...

technical field

[0001] The invention relates to the technical field of social networks, in particular to an adjustable overlapping community discovery method based on tag propagation. Background technique

[0002] With the continuous development of computer technology and the Internet, devices such as computers, smartphones, and sensors generate data all the time. Ordinary people have more and more opportunities to access the Internet, and all walks of life are also taking advantage of the convenience brought by the Internet. These data grow exponentially, resulting in massive amounts of data. Most of the various information in life and production can be transformed into a network structure or a graph (Graph) structure, that is to say, network structures are everywhere. Newman summarized the networks in real life into four categories; communication networks, such as the World Wide Web of Information, the collection retrieval network of the University of Electronic Science ...

Examples

Embodiment 1

[0056] An example is given below to describe the label propagation process in detail. figure 2 is an intermediate state of an example network in the label propagation process, figure 2 The shown network contains 21 nodes and 21 edges, and the input parameter K=0.6. Propagate labels in a certain order, (7,17,18,3,2,7,11…). The label propagation of node 7 is performed first, and the labels contained in the neighbor nodes of node 7 are {C,A,A,A,A,A,A,B}. The label with the largest number is A, the number is 6; the number of labels B and C are both 1. because Where |B| represents the number of B labels, and |A| is the same, so label B does not meet the conditions and cannot be added to the label group of node 7. Therefore, it cannot join the label group of node 7 either. The label of node 7 is updated to A, such as image 3 shown. Continue to execute the label propagation of the node 17 in a specific order, and delete the labels in the label group of the node 17. The l...

Images