Swarm Intelligence Techniques for Optimization in Data Clustering

Clustering represents the large datasets by a structured well defined number of clusters or prototypes. K-Means is a useful technique to data clustering which partitions the data into K-Clusters. However, the results of k-means algorithm are based on the selection of initial seeds and converge to local optimum solution. The Swarm Intelligence (SI) is an algorithm to apply many simple agents behaviour which inn terns lead to an emergent global behaviour solution. Data mining tasks require fast and accurate partitioning of large datasets, which may come with a number of attributes or features. The Swarm Intelligence Optimization Techniques like ACO and PSO has successfully been applied to a number of real world clustering problems in order to meet the clustering requirement. This paper surveys the research work in the area of swarm intelligence for solving clustering problems. The new clustering model is proposed using the merits of ACO and PSO to overcome the drawback of k-means clustering algorithm.