Clustering voluminous of heterogeneous data

Abstract

Clustering analysis is one of the most commonly used data processing algorithm. In this era of data explosion, clustering large volume of data is very challenging. If the data is heterogeneous, it brings more challenges. K-Prototype is an algorithm which aims at clustering mixed dataset which contains numerical as well as categorical data. This algorithm does not distinguish between nominal data and ordinal data. There is no solution available in literature for dealing with voluminous of data having numerical as well as categorical data. In this paper we extended the K-Prototypes algorithm by feature scaling and behavior of γ value. K-Prototypes algorithm integrates the K-Means and K-Modes algorithms for mixed numeric and categorical attributes. We conducted the experiments on benchmark dataset to demonstrate partitioning clustering algorithm.