Decreasing iteration number of k-medoids algorithm with IFART

K-medoids is a well known and widely used algorithm in data clustering. Performance of the algorithm depends on the initialization of cluster centers as in other centered based clustering techniques. In this article we used an initialization method based on fuzzy art to initialize clusters. The algorithm has been applied to different datasets. The experiments show that our approach can achieve higher or comparable performance when it is compared with conventional k-medoids.