Improved k-medoids clustering based on cluster validity index and object density

Clustering is the process of classifying objects in to different groups by partitioning sets of data into a series of subsets called clusters. Clustering has taken its roots from algorithms like k-means and k-medoids. However conventional k-medoids clustering algorithm suffers from many limitations. Firstly, it needs to have prior knowledge about the number of cluster parameter k. Secondly, it also initially needs to make random selection of k representative objects and if these initial k medoids are not selected properly then natural cluster may not be obtained. Thirdly, it is also sensitive to the order of input dataset. First limitation was removed by using cluster validity index. Aiming at the second and third limitations of conventional k-medoids, we have proposed an improved k-medoids algorithm. In this work instead of random selection of initial k objects as medoids we have proposed a new technique for the initial representative object selection. The approach is based on density of objects. We find out set of objects which are densely populated and choose medoids from each of this obtained set. These k data objects selected as initial medoids are further used in clustering process. The validity of the proposed algorithm has been proved using iris and diet structure dataset to find the natural clusters in this datasets.