DFCM: Density Based Approach to Identify Outliers and to Get Efficient Clusters in Fuzzy Clustering

The task of outlier identification is to find small groups of data objects that are exceptional when compared with rest large amount of data. The identification of outliers can lead to the discovery of truly unexpected knowledge in areas such as electronic commerce, credit card frauds, voting irregularity analysis, data cleansing, network intrusion, severe weather prediction & many more. This paper deals with the identification of outliers and to get efficient clusters in fuzzy clustering. In this paper a new density based definition of outlier and an algorithm dasiaDFCMpsila is proposed; which works in two phases. In first phase, it identifies outliers and separate them from original data-set and in the second phase, it creates clusters from noiseless data. DFCM modifies FCM fuzzy clustering technique to create clusters. But it can also be implemented with any other fuzzy clustering technique. Numerical examples and tests show that proposed algorithm gives better result when compared with FCM.