A cluster validity index for fuzzy c-means clustering

This paper presents a new validity index for validation of the fuzzy partitions generated by the fuzzy c-means algorithm. The proposed validity index is based on the compactness and separation measure. The compactness measure is defined as the weighted square deviation of the intra cluster, and the separation measure is defined as the distance for the different fuzzy sets. There are high expectations of a large degree compactness and separation among clusters for a good fuzzy partition. The contrast experimental results with various indices show that the proposed index is more robust to the noise and can identify clusters with different densities and sizes.