Virtual Ear: A Center Based Visual Clustering Approach

摘要: Visual data mining is one of the most interesting research areas in the field of Information Technology. Finding the pattern of a dataset and showing them in an understandable way is a challenging job. There are several methods already proposed and more are coming in recent times. Here we propose a system named as “Virtual Ear”(to resemble with human ear) to find not only the cluster but also the cluster center. The main profit of this approach is that it helps the user to make more concentration on visual presentation for further analysis as the number of point is reduced to cluster centers though the original nature of the data filed is not lost. This method is more deterministic as we concentrate on specific data points and their cluster center rather than clustered region (which is some times not so much separable). Finally the resultant of the data points gives the user some capability to macro level analysis.