Research On Community Division Algorithm Based On Complex Network Structure

With the continuous advancement of science and technology,the types and scales of complex systems in the real world are constantly changing.The diversity of species and the rapid growth of the scale make the complexity of complex networks more and more complicated.The acquisition of useful knowledge in the data volume has become a hot topic in the research of complex networks.Community division can reveal the potential laws in the real world.It can be applied to social networks,the World Wide Web,e-commerce and criminal investigation.Therefore,the research of this subject has theoretical value and practical significance.The community-based algorithm based on resistor network can divide the social network in a linear time.However,this kind of algorithm uses the idea of graph segmentation to divide the social network into two communities with similar scales.It needs to iteratively execute the algorithm for the problem of multi-community partitioning,and the algorithm execution efficiency is relatively low;in addition,the algorithm needs to set the voltage in advance.Thresholds are used to divide social networks into communities.Setting thresholds will directly affect the accuracy of community demarcation.Aiming at these two kinds of problems in this kind of algorithm,this paper proposes a community partition algorithm based on voltage mean value,uses the node voltage mean to select the community center node,avoids the setting of voltage threshold,and can be divided into multiple communities;This paper also proposes a community partitioning algorithm based on the number of determined communities,using the community core node as the initial center node to solve the problem of pre-established number of divided communities.Finally,this paper adopts two classical data sets in the field of community division to carry out experimental analysis and verification of the proposed algorithm.The experimental results show that the two algorithms proposed in this paper have good division results and improve the efficiency of the algorithm to some extent.