Spam detection using hybrid of artificial neural network and genetic algorithm

Abstract

Spam detection is a significant problem which is considered by many researchers by various developed strategies. In this study, the popular performance measure is a classification accuracy which deals with false positive, false negative and accuracy. These metrics were evaluated under applying two supervised learning algorithms: hybrid of Artificial Neural Network (ANN) and Genetic Algorithm (GA), Support Vector Machine (SVM) based on classification of Email spam contents were evaluated and compared. In this study, a hybrid machine learning approach inspired by Artificial Neural Network (ANN) and Genetic Algorithm (GA) for effectively detect the spams. Comparisons have been done between classical ANN and Improved ANN-GA and between ANN-GA and SVM to show which algorithm has the best performance in spam detection. These algorithms were trained and tested on a 3 set of 4061 E-mail in which 1813 were spam and 2788 were nonspam. Results showed that the proposed ANN-GA technique gave better result compare to classical ANN and SVM techniques. The results from proposed ANNGA gave 93.71% accuracy, while classical ANN gave 92.08% accuracy and SVM technique gave the worst accuracy which was 79.82. The experimental result suggest that the effectiveness of proposed ANN-GA model is promising and this study provided a new method to efficiently train ANN for spam detection.