Design and Development of Naive Bayes Classifier

Abstract

The naïve Bayes classifier is a simple form of Bayesian classifiers which assumes all the features are independent of each other. Despite this assumption, the naïve Bayes classifier’s accuracy is comparable to other sophisticated classifiers. In this paper we designed and developed a naïve Bayes classifier for a better understanding of the algorithm. The classifier is tested on two different data sets from the University of California at Irvine machine learning repository. Different cross validation methods are used to calculate the accuracy of the developed classifier. The different accuracies obtained are compared to get the best accuracy of the classifier. This value is also compared with accuracies obtained for the same data sets using different algorithms reported in other papers. It was observed from the comparisons that the naïve Bayes classifier’s results are very comparable to other algorithms.