Automatic Clustering Approaches Based On Initial Seed Points

-Since clustering is applied in many fields, a number of clustering techniques and algorithms have been proposed and are available in the literature. This paper proposes a novel approach to address the major problems in any of the partitional clustering algorithms like choosing appropriate K-value and selection of K-initial seed points. The performance of any partitional clustering algorithms depends on initial seed points which are random in all the existing partitional clustering algorithms. To overcome this problem, a novel algorithm called Weighted Interior Clustering (WIC) algorithm to find approximate initial seed-points, number of clusters and data points in the clusters is proposed in this paper. This paper also proposes another novel approach combining a newly proposed WIC algorithm with K-means named as Weighted Interior K-means Clustering (WIKC). The novelty of this WIKC is that it improves the quality and performance of K-means clustering algorithm with reduced complexity. The experimental results on various datasets, with various instances clearly indicates the efficacy of the proposed methods over the other methods. Keywords-Clustering, partitioning, data mining, unsupervised learning, hierarchical clustering, kmeans.