Overlapping Community Based Relationship Discovery Algorithm for Herbs

The relationships discovery of herbs is an active research focus in data mining of traditional Chinese medicine (TCM). The existing methods are difficult to effectively find herbs relationships in TCM network. TCM doctors often give prescriptions according to the characteristics of herbs to combination. In this study, we propose an overlapping community based herbs relationship discover algorithm, through calculating the similarity of the characteristics of herbs and discovering the relationship between nodes based on the relationship between neighboring nodes. By comparing the proposed algorithm with the state-of-the-art methods, we conduct experiments in real TCM networks, the results show that the proposed algorithm can effectively partition the communities than traditional methods and can find common herbal combinations. It is interesting to find that searching similar relationships between nodes in a TCM networks can benefit partitioning the network community structures.