Overlapped cell discovery method based on influence and seed expansion

Abstract

The invention relates to an overlapped community discovery method based on influence and seed expansion. The method comprises the steps of (1) reading a network data set, generating a network structure diagram and acquiring node information, (2) calculating an influence score of a node with the combination of a Jaccard coefficient and node information, (3) finding a core seed node based on the influence score and establishing a core seed cell set, (4) merging cells with a high similarity and obtaining an optimized core seed cell set, (5) marking a candidate node according to similarities of nodes and corresponding cells in a cell neighbor set, (6) add the candidate node satisfying a modularity requirement into the corresponding cell to obtain an overlapped community set, (7) adding a freenode into the overlapped cell with a high similarity and updating the overlapped community set, (8) merging the cells with a high similarity and obtaining an optimized overlapped community set, and (9) outputting a final cell segmentation result. According to the method, an overlapped structure of a complex network can be efficiently and accurately divided.

Description

technical field [0001] The invention relates to the technical field of discovering overlapping communities on a large-scale network, in particular to a method for discovering overlapping communities based on influence and seed expansion. Background technique [0002] With the rapid development of social information networks, many complex network structures have emerged, such as social networks, scientist collaboration networks, literature citation networks, and protein collaboration networks. In a complex network, nodes represent individuals in the network, and edges represent connections between individuals, and some edges also have attribute values. The community structure in a complex network usually shows that the points within the community are tightly connected, while the points between the communities are sparsely connected. Community discovery is one of the key techniques for studying complex network structures. At present, the research results discovered by the co...

AI Technical Summary

Problems solved by technology

[0004] At present, many scholars have conducted research on overlapping community discovery based on seed expansion, and have achieved a lot of results, but there are still several problems: First, the time and space complexity of the algorithm is relatively high, and there are problems when dealing with large-scale networks. Insufficient; secondly, the algorithm fails to fully consider the closeness between nodes when selecting seed nodes, which affects the accuracy of community expansion mining; finally, in the community expansion stage, the closeness of nodes and communities is not considered comprehensively, and the parameter The choice is also not very reasonable, so the accuracy of the community discovery is not high

Images