Weighted fuzzy kernel-clustering algorithm with adaptive differential evolution and its application on flood classification

Flood classification is the fundamental problem of flood risk analysis and plays an important role in flood disaster risk management. Considering the fact that flood classification is a problem of multi-attribute and multi-stage fuzzy synthetically evaluation, this paper mainly proposed the weighted fuzzy kernel-clustering algorithm (WFKCA) with adaptive differential evolution algorithm (ADE) to solve this problem. Firstly, WFKCA is detailed introduced, and then the differential evolution algorithm (DE) is applied for the fuzzy clustering, thus to obtain the better results. Taking into consideration the disadvantage of DE, ADE is present after the introduction of DE. Finally, the combination of WFKCA and ADE is applied for flood classification, and the results demonstrated the methodology is reasonable and reliable, thus provide a new effective approach for flood classification. Copyright Springer Science+Business Media Dordrecht 2013