Effective Similarity Measure with Enhanced K-medoid Partitioned Clustering Algorithm

Now a days, it becomes more difficult for users to find the documents related to their interests, since the number of available web pages grows at large. Clustering is the method of grouping the data objects into classes or clusters so that data objects within a cluster have high similarity as compared to one another, but are very dissimilar to objects in other clusters. Such similarity measures helps users to find the relevant documents much easier and also help users to form an understanding of the different facets of the query that have been given for web search engines. A popular technique for clustering is based on partitioning method i.e. k-medoids, such that the data is partitioned into k clusters and each cluster is represented by one of the objects located near the centre of the cluster. The remaining data objects belongs to the cluster whose value with the center is nearer. Existing algorithms for k-medoids partitioned clustering are slower and do not scale to large number of data objects. A fast k-medoids algorithm can attempt to reduce these drawbacks, but still there is a limitation when the algorithm is used for large number of data objects. So, we have proposed an efficient method to compute the limitation for this algorithm. The experimental results shows that this enhanced k-medoids algorithm is efficient to other partitioned methods and can help users to find the relevant documents of their interest much in easier manner.