Enhancing clustering quality of fuzzy geographically weighted clustering using Ant Colony optimization

Fuzzy Geographically Weighted Clustering (FGWC) is recognized as one of the most efficient methods for geo-demographic analysis problem. FGWC uses neighborhood effect to remedy the limitation of classical fuzzy clustering methods in terms of geographic factors. However, there are some drawbacks of FGWC such as sensitivity to cluster initialization phase that is required to overcome. In this paper a new hybrid approach of FGWC based on Ant Colony Optimization (ACO), namely FGWC-ACO is proposed in which the initialization is performed better and in an appropriate manner. Based on the experimental simulation, the proposed method clearly outperforms the standard FGWC and offers a better geo-demographic clustering quality.