The Adaptive Method for Closely Communicating Community Detection Based on Ant Colony Clustering

The investigation of community structures in networks is an important issue in many domains and disciplines.Closely communicating community is different from the traditional community which emphasize particularly on structure or context. The definition of closely communicating community and measuring method are introduced firstly. Based on the previous work, the closely communicating community detection method is improved in some aspects of adaptive adjusting, such as entropy based weight modulation, combining historical paths and random wandering to select next coordination, the strategy of forcing unloading and the adaptive change of ant´s eyesight. The value selection of parameters is discussed in the portion of experiments, and the results also reveal the improvement of our algorithm in adaptive adjusting.