Particle of Swarm Based Automatic Variable Weighting Clustering Algorithm for Large Database

The clustering techniques play an important role in data mining process. For the mining of large data faced a lot of problem of noise and large number of iteration. The process of pattern generation used two type of technique such as supervised learning and unsupervised learning. In unsupervised learning clustering process is used. The varieties of clustering technique are used such as k-means, FCM and weighted clustering technique. The weighted clustering technique gives the two solution approach one is seed selection and another is mapping of seed in terms of weight of centre. In this dissertation modified the seed selection process using particle of swarm optimization technique. The particle of swarm optimization process select variable value one is seed value and another is weight of centre value. In weighted cluster techniques used some value of centre and generate new centre value of new cluster for the better generation of cluster. For more improvement of weighted clustering technique used two level weighted clustering techniques for better improvement of cluster technique. For the validation of clustering technique used various methods such as weighted clustering technique and multi-level weighted clustering technique. In the last decade, several approaches able to notice several clustering resolutions have been presented. According to the review, they can briefly be characterized into approaches effective on the novel data-space, methods performing space transformations, and methods analysing subspace projections. The main conception is to reflect every subspace as a state in a fuzzy inference, with rule allowed.