Improving Spam Mail Filtering Using Classification Algorithms With Partition Membership Filter
C. Neelavathi, S. M. Jagatheesan · 2016
E-mail is one of the most popular and frequently used ways of communication due to its worldwide accessibility, relatively fast message transfer and low sending cost. The different classification algorithms (JRip, Filtered Classifier, K- star, SGD, Multinomial) which are used for classify the email as spam or not. However these algorithms has number of drawbacks such that lack of useful and relevant features that can distinguish between spam and non-spam email increase data dimensionality that decreases accuracy. To overcome these problems, Random Tree algorithm is used. In the proposed algorithm, Random Tree classifier generates the best outcome in terms of accuracy, kappa statistics and less error rate.