Performance Analysis of K-NN and Naïve Bayes Classifiers for Spam Filtering Applications

Abstract

Pattern classification is one of the most important and leading aspects of modern image processing systems. By training a classifier
on a set of data, the unseen samples can be categorized as much accurate as training has been done. There are many different classifiers having
varying accuracies, design complexities and performance. With different design strategies these classifiers may have different characteristics. In
this paper a performance analysis of K-NN and Naïve Bayes classifiers have been presented for the classification of spam emails. Different
design aspects of both classifiers have also been presented in terms of computational complexity and classification accuracy against their
performance.

Keywords – K-NN Classifier, Naïve Bays Classifier, Spam Filtering, Performance Analysis of K-NN and Naïve Bays Classifiers