Comparative Analysis Of Fuzzy Clustering Algorithms In Data Mining
Binsy Thomas, Madhu Nashipudimath · International journal of advanced research in computer science and electronics engineering · 2012
Data clustering acts as an intelligent tool, a method that allows the user to handle large volumes of data effectively. The basic function of clustering is to transform data of any origin into a more compact form, one that represents accurately the original data. Clustering algorithms are used to analyze these large collection of data by means of subdividing them into groups of similar data. Fuzzy clustering extends the crisp clustering technique in such a way that instead of an object belonging to just one cluster at a time, the object belongs to one or more clusters at the same time with appropriate membership values assigned to the object in a cluster. This paper addresses the major issues associated with the conventional partitional clustering algorithms, namely difficulty in determining the cluster centers and handling noise or outlier points. Integration of fuzzy logic in data mining subjugates these traditional methods to handle natural data which are often vague. The study provides an analysis of two fuzzy clustering algorithms videlicet fuzzy c- means and adaptive fuzzy clustering algorithm and its illustration on different fields.