USING VISUAL DATA MINING TECHNIQUES IN CLUSTERING ANALYSIS AND AN APPLICATION

Abstract

Cluster analyses play a very important role in data mining process. A critical and important issue is to decide outliers, potential cluster structures, optimal number of clusters, to choose suitable cluster algorithms and to effectively evaluate the cluster results in cluster analysis. Various quantitative methods can overcome the problems. However, using the quantitative methods, some details can be missed that the details would be important. In the article, dealing with visual data mining techniques, with human visual system' supports, is showed how to effectively decide outliers, potential cluster structures, optimal number of clusters, choose suitable cluster algorithms and evaluate the cluster results in cluster analysis. In this way, more efficient the result of clustering can found in the field of data mining.