An improved ICPACA based K-means algorithm with self determined centroids

The bioinformatics field which is now dealing with a vast amount of data such as the protein patterns and the gene expression data, with a lot more information still to be unraveled, uses the basic techniques and tools for Data mining for retrieving useful information from huge biological databases. Clustering is a popular Data mining technique which is extensively used efficiently. The K-means clustering algorithm, because of its simplicity, is the most widely used clustering algorithm. But it has some inherent drawbacks. This paper discusses about an enhanced algorithm that combines the K-means clustering algorithm with Improved Clustering Process Ant Colony Algorithm (ICPACA). The combined algorithm is capable of determining the optimal number of clusters and their corresponding centroids. It also eliminates the problems due to local optimal solutions and dependence on initial centroids.