Feature Selection: A Hybrid Approach Based on Self-Adaptive Ant Colony and Support Vector Machine

Ant colony optimization (ACO) is a kind of bionic swarm intelligence algorithm belongs to artificial intelligence (AI) field and has been successfully applied in resolving complex optimization problems. Support vector machine (SVM) is a new machine learning method with greater generalization performance, and has shown its superiority in classification and regression problems. By combining the advantages of two approaches, this paper proposes a new hybrid method based on self-adaptive ant colony optimization and SVM for feature selection in data mining which exploits novel heuristic information and takes results of feature subset classified by SVM method as a positive feedback adjustment factor for ACO. Experiments and results showed the hybrid method has a higher classification accuracy or smaller feature subset when having the same classification accuracy as other algorithm.