Comparative Study of Advanced Classification Methods

Abstract

Abstract-The availability of huge amounts of data leads to the need for powerful data analysis tools to extract useful knowledge. Several data mining tools exist to improve data analysis on large data sets. There are a number of data mining tools namely Classification and Regression, Association Rules, Cluster Analysis and Outlier Analysis. Classification in data mining is a form of data analysis that extracts model using a training set, whose class label is known. This model is used as a classifier and is used for predicting the class label of unknown data set. This method of working is known as supervised learning. Various types of classifiers are Support Vector Machine, Bayesian Classification, Decision Tree Induction, Artificial Neural Network, K-Nearest Neighbor and Genetic Algorithms etc. Support Vector Machine (SVM) belongs to the class of supervised learning algorithms, SVMs construct a hyperplane or set of hyperplanes in higher dimensional space that separates two classes. Naive Bayes Classifiers (NBC) are statistical classifiers and are based on Bayes Theorem. They can predict the class membership using probabilities. In this paper study of Support Vector Machine and Naive Bayes Classifier are carried out for various training sets and their efficiency for the unknown set are analyzed for Accuracy, AUC, Error Rate, F-measure, Precision, Recall and Specificity results are documented. Most of the results correlated with the literature.