Toward a novel art inspired incremental community mining algorithm in dynamic social network

摘要

The problem of community detection has a long tradition in the area of data mining research and has many challenging aspects, particularly with regard to community detection in a time-varying context. Temporal smoothness is an important concept in this domain treated specifically by Evolutionary clustering. However, this approach is mostly designed for networks with low changes during different time steps. In this research, we present a new dynamic community detection algorithm in a representative-based category called ARTISON, which is inspired by the Adaptive Resonance Theory technique- a famous adaptive clustering model in neural networks. The proposed model is able to capture both low and high changes in the networks in binary. Further, our approach recognizes the number of communities automatically in both weighted and binary networks. Detailed experiments utilizing the MATLAB platform have yielded encouraging results for different measures and further motivate the application of such bio-inspired models in a dynamic social networking context.