AN EFFECT OF DISTANCE MEASURES IN CLASSIFYING LARGE DATASETS

As digital usage increases day to day, a large voluminous data is accumulated in various applications. Bulky data contain useful facts and information which may go unidentified if not processed. Data mining is a promising technology which processes bulky data to extract facts. Data mining technique is better influenced by the classification work. Effective use of classifier with accurate distance measure calculation helps in extracting minute facts available in the bulky data. In this paper, a set of classifiers such as K nearest neighbor, Support vector machine and centroid are implemented. Then a novel classifier, an improved fuzzy soft classifier is implemented and results are produced using two distance measures viz., Euclidean and Jaccard. The proposed classifier shows better results when Euclidean distance measure is used with continuous data and it shows better results when jaccard distance is used with categorical data. Keywords – Fuzzy soft set, Euclidean distance, Jaccard distance, Support Vector Machine, K Nearest Neighbour.