Performance Evaluation of Partition Based Clustering Algorithms in Grid Environment Using Design of Experiments

Clustering is one of the most important research areas in the field of data mining. Clustering means creating groups of objects based on their features in such a way that the objects belonging to the same groups are similar and those belonging to different groups are dissimilar. Here K Means, K Medoids are basic partition based clustering algorithms. One of the disadvantages of using these algorithms its unsuitability for larger data sets. To solve this problem Grid environment has been selected. The main objective of this paper is to implement the partition based clustering algorithms in the Grid environment on Grid Gain middleware and analyze their performance for large datasets with Design of Experiment (DOE) framework. K-means cluster data faster than K-medoids when tested with large data sets and the results are found to be satisfactory. Keywords— Grid Gain, K-Means, K-Medoids, DOE .