SVM classifier incorporating feature selection using GA for spam detection

Authors:
Huai-bin Wang;Ying Yu;Zhen Liu
Affiliations:
Dept. of computer science, Tianjin University of Technology, Tianjin, China;School of Computer Science and Technology, Harbin Institute of Technology, Harbin, China;Nagasaki Institute of Applied Science, Japan, Nagasaki, Japan
Venue:
EUC'05 Proceedings of the 2005 international conference on Embedded and Ubiquitous Computing
Year:
2005

Citing 8
Cited 3

Genetic algorithms: A Survey

Computer
Support-Vector Networks

Machine Learning
Wrappers for feature subset selection

Artificial Intelligence - Special issue on relevance
Choosing Multiple Parameters for Support Vector Machines

Machine Learning
An introduction to variable and feature selection

The Journal of Machine Learning Research
Variable selection using svm based criteria

The Journal of Machine Learning Research
A Bayesian Approach to Joint Feature Selection and Classifier Design

IEEE Transactions on Pattern Analysis and Machine Intelligence
Support vector machines for spam categorization

IEEE Transactions on Neural Networks

Behavior-based spam detection using a hybrid method of rule-based techniques and neural networks

Expert Systems with Applications: An International Journal
Combining neural networks and semantic feature space for email classification

Knowledge-Based Systems
Automatic thesaurus construction for spam filtering using revised back propagation neural network

Expert Systems with Applications: An International Journal

Quantified Score

Hi-index	0.01

Visualization

Abstract

The use of SVM (Support Vector Machines) in detecting e-mail as spam or nonspam by incorporating feature selection using GA (Genetic Algorithm) is investigated. An GA approach is adopted to select features that are most favorable to SVM classifier, which is named as GA-SVM. Scaling factor is exploited to measure the relevant coefficients of feature to the classification task and is estimated by GA. Heavy-bias operator is introduced in GA to promote sparse in the scaling factors of features. So, feature selection is performed by eliminating irrelevant features whose scaling factor is zero. The experiment results on UCI Spam database show that comparing with original SVM classifier, the number of support vector decreases while better classification results are achieved based on GA-SVM.