Optimization Features Using GA-SVM Approach

Andy, Michael Fernando, Kristanto Halim, Gradiyanto Sanjaya

Abstract: Feature selection often used to choose the feature that maximizes the prediction of classification accuracy. Feature selection is one of the most important factor that influence classification accuracy rate. In this paper we proposed the combination of Genetic Algorithm (GA) and Support Vector Machine for feature optimization. In this research we compare the result with K Nearest Neighbor, Decision Tree, and Linear Discriminant Analysis. For better comparison, the experiment was conducted using 6 different dataset. The result shows that GA-SVM gives better accuracy than using all features or other method on 3 of 6 dataset.

Keywords: Feature Optimization, Genetic Algorithms GAs, Support Vector Machine SVM