ANDERATION: A New Anti-Community Detection Algorithm and its Application to Explore Incompatibility of Traditional Chinese Medicine

The problem of community detection has attracted great attentions from various fields and community structure is one of the most important characteristics in complex networks. However, there are few researches on the detection of anti-community structure, where the nodes share no or few connections inside their group they belong to but most of their connections outside. In Traditional Chinese Medicine (TCM), the incompatibility problem of herbs is a great challenge to clinical medication safety, which becomes a serious threat to public health. In this paper, a new ANti-community detection algorithm based on the DEgree and the RATio between the Inner degree and the Outer degree of a Node (ANDERATION), is proposed, in which these two factors are firstly introduced to detect anti-community structure by first creating the initial anti-community structure and then maximizing the objective function. Experimental results on 15 synthetic and 14 real-world networks demonstrate that the proposed algorithm can detect better anti-community structures with less running time than the existing algorithms. By applying ANDERATION to the herb network, we find that it is effective in exploring incompatible herb combinations in TCM.