MIXED VARIABLE ANT COLONY OPTIMIZATION TECHNIQUE FOR FEATURE SUBSET SELECTION AND MODEL SELECTION

This paper presents the integration of Mixed Variable Ant Colony Optimization and Support Vector Machine (SVM) to enhance the performance of SVM through simultaneously tuning its parameters and selecting a small number of features. The process of selecting a suitable feature subset and optimizing SVM parameters must occur simultaneously, because these processes affect each other which in turn will affect the SVM performance. Thus producing unacceptable classification accuracy. Five datasets from UCI were used to evaluate the proposed algorithm. Results showed that the proposed algorithm can enhance the classification accuracy with the small size of features subset.