Research On Community Discovery Algorithm Based On Information Communication

Recent studies of complex networks no doubt have become a focus of academic circles in various fields, such as interpersonal relationship networks, the world wide web, global traffic networks and so on. The researchers by a large number of experiments showed that in addition to small world and scale-free property, community structure is also an important characteristic of complex networks. Community is a collection of nodes which are similar to each other and are different from other nodes or networks. Finding the community structure in complex network is contribute to a deeper understanding of the structure and characteristics in complex networks, revealing the hidden information and forecasting further works in complex networks. So the discovery of community structure in networks has important guiding significance for people to explore the mysteries of science, and has gradually become the research focus in recent years.This paper studied deeply of the existing community detection algorithm in complex networks, putting the information transmission method into the community detection algorithm, conveyed information on complex networks at first, and then transformed this information into the appropriate data for clustering, so as to obtain the corresponding community structure.The main work of this paper include the following aspects:(1) We proposed a community detection algorithm based on signal adaptive processing(SAC algorithm):letting the signal transmission adaptively in the complex network at first, so that each node in the network could obtain signal,and then getting the effect of vector corresponding to each other, transforming the topologies of the network nodes into the geometric relationship between vectors, combining the clustering properties for dividing community.In order to obtain the more reasonable space vector, this paper was based on the adaptive making all nodes have access to information, and put forward a method to determine the optimal transmission frequency,improving the accuracy and precision of algorithm.(2) We proposed a community detection based on the effect degree and label propagation algorithm (ILAP algorithm):First,all nodes was assigned a unique label, then found out the key nodes in the network, from the core node to start propagating label, and the influence degree was introduced to the label update strategy, finally, when the community modularity tended to be stable then stoped updating the label. Added the effect degree in the process of label propagation could reduce the random of label propagation, also avoided the oscillation phenomenon that may occur in the label propagation process, to a certain extent, improve the stability of community structure.(3) The SAC algorithm and ILAP algorithm were tested on the computer generated network and the real network, obtaining the relatively good community classification effect, finally compared with other several classical community detection algorithms in the experiment.