A Performance Assessment on Various Data mining Tool Using Support Vector Machine

Abstract

Data mining is essentially the discovery of valuable information and patterns from huge chunks of available data. Two indispensable techniques of data mining are clustering and classification, where the latter employs a set of pre-classified examples to develop a model that can classify the population of records at large, and the former divides the data into groups of similar objects. In this paper we have proposed a new method for data classification by integrating two data mining techniques, viz. clustering and classification. Then a comparative study has been carried out between the simple classification and new proposed integrated clustering-classification technique. Four popular data mining tools were used for both the techniques by using six different classifiers and one clustered for all sets. It was found that across all the tools used, the integrated clustering-classification technique was better than the simple classification technique. This result was consistent for all the six classifiers used. For both of the techniques, the best classifier was found to be SVM. From the four tools used, KNIME found to be the best in terms of flexibility of algorithm. All comparisons were drawn by comparing the percentage accuracy of each classifier used.