Clustering is used to organize data for efficient retrieval. One of the problems in clustering is the identification of clusters in given data. A popular technique for clustering is based on K-means such that the data is partitioned into K clusters. In this method, the number of clusters is predefined and the technique is highly dependent on the initial identification of elements that represent the clusters well. A large area of research in clustering has focused on improving the clustering process such that the clusters are not dependent on the initial identification of cluster representation. In this paper, I advance an adaptive technique that grows the clusters without regard to initial selection of cluster representation. As such, the technique can identify K clusters in an input data set by merging existing clusters and by creating new ones while keeping the number of clusters constant. The technique has been used to achieve an impressive speedup of a search process when other efficient search techniques may not be available.
Jan-11-2006, 12:34:12 GMT