Enhancement to Asymmetric Clustering

摘要

Data mining or the knowledge discovery process is a technique for analyzing voluminous amount of data. It is considered important technology in areas like market analysis and management, financial data analysis, Fraud detection, biological data analysis and other various scientific applications. Clustering is a technique in data mining which groups similar objects into clusters. There are various approaches in clustering and performance of clustering depends on ability of algorithms to find hidden useful knowledge. In an asymmetric clustering, data is partitioned in which similar and dissimilar data is separated out. The partitions can be affirmed vigorously and usually run on a single cluster at a time. In the previous model discussed in the paper the time taken to produce clustering results was much lower and decreases the performance of clustering. In this paper we have proposed a model to improve asymmetric clustering results by combining both mean shift and K- means normalization algorithm. Final Clustering results are plotted and performance parameters are compared between previous and proposed method. The proposed model shows performance improvement with increase in accuracy and reduced execution time and noise level.