The Analysis About Complex Communication Network Topology And The Research Of The Community Extraction Algorithm

As a modern social communication tool, Mobile Phone largely depicts the interpersonal social relationships between people. Consequently, the individuals can built a complex network structure according to these social relationships, and it will be of great value to the analysis of social network structure and human dynamics.The community structure is one of the most widespread topological characteristic of complex network. It has sparse external links between the communities, dense internal links inside the community, as well as overlapping among different communities, which means one person can simultaneously belong to multiple communities. With the development of online social networks, the scale and range of application is more and more big, how to effectively analyze the characteristics of the network is a problem which urgently needs to be solved. However, the classical community detection algorithm needs to analyze all of the topology characteristics of the network, which makes it unable to be applied on the large-scale network. In many cases, we only concern about one particular community, so there is no need to find the whole community structure.Being aimed at the above problems and based on the theories of complex networks, this paper completes the following works which can be started from two aspects.First, the multiple time-varying communication network model comes up, through multiple edges, its model can dynamically and in real time record users’call in many times. The value of edge’s attribute weights represents the time of call, after data preprocessing, we can read users’complete communication behaviors through the network. Also about the overall of network and individual aspects, degrees, average degrees and evolutions are respectively analysed. At the same time the distribution of calling time, talking time’s interval are explored, so we can get the universal law of the human behaviors, this results will provide the realistic basis for community extraction.Second, for one person, who can have close contact with his or her family and friends, By the fact, some significant associations can be formed. Colleagues and work partners will also provide the possibility for the formation of communties. In this paper, based on a communication network of ultra large scale, somebody’s different groups will be extracted. So the community extraction algorithm comes up, there is no need to know all of the network topology in advance, it combines the community search and community determining, a special social communication network can be selected from a unknown structure effectively, this makes the analysis of large scale network community structure become possible. At first, In order to verify the effectiveness and feasibility of the algorithm, an experiment has been done on the network with prior knowledge. The results show that the accuracy and efficiency of this algorithm. Then further inquiry is done about selecting the initial node which may influence the accuracy of algorithm, the results prove that the initial nodes have a great influence on the result of the algorithm. In order to better adapt to multiple time-varying communication network, this algorithm is improved, by the timing relationship on the edge weights, different societies are divided. Then an individual comunities which may be different communities also can be got, through analysising the results of extraction, the corresponding conclusion can be elicited.