Fuzzy clustering in drug design: Application to cocaine abuse

This paper describes a fuzzy clustering methodology for selecting representative clusters for conformational analysis. A novel feature extraction procedure based on the structural properties of the molecule is presented. A composite feature vector is developed and the practical applicability of two fuzzy relational clustering techniques, NERF c-means and FRC, is demonstrated. Several cluster validity measures are applied and are shown to provide reliable information about the validity of the partitions produced by the clustering techniques.