Academic Journals Database
Disseminating quality controlled scientific knowledge

Presenting a Hybrid Feature Selection Method Using Chi2 and DMNB Wrapper for E-Mail Spam Filtering

ADD TO MY LIST
 
Author(s): Seyed Mostafa Pourhashemi | Alireza Osareh | Bita Shadgar

Journal: International Journal of Computer Science and Network Solutions
ISSN 2345-3397

Volume: 1;
Issue: 2;
Start page: 16;
Date: 2013;
VIEW PDF   PDF DOWNLOAD PDF   Download PDF Original page

Keywords: Feature Selection | Classification | Spam Filtering | Machine Learning

ABSTRACT
The growing volume of spam emails has resulted in the necessity for more accurate and efficient emailclassification system. The purpose of this research is presenting an machine learning approach forenhancing the accuracy of automatic spam detection and filtering and separating them from legitimatemessages. In this regard, for reducing the error rate and increasing the efficiency, the hybrid architecture onfeature selection has been used. Features used in these systems, are the body of text messages. Proposedsystem of this research has used the combination of two filtering models, Filter and Wrapper, with ChiSquared (Chi2) filter and Discriminative Multinomial Naïve Bayes (DMNB) wrapper as feature selectors.In addition, MNB classifier, DMNB classifier, SVM classifier and Random Forest classifier are used forclassification. Finally, the output results of this classifiers and feature selection methods are examined andthe best design is selected and it is compared with another similar works by considering differentparameters. The optimal accuracy of the proposed system is evaluated equal to 99%.
Why do you need a reservation system?      Affiliate Program