Title :
Hybrid feature selection algorithm based on dynamic weighted ant colony algorithm
Author :
Xiong, Shang-Hua ; Wang, Ji-Yi ; Lin, Huang
Author_Institution :
Coll. of Math. & Inf. Sci., Zhejiang Normal Univ. Jinhua, Jinhua, China
Abstract :
This paper proposes a hybrid feature selection algorithm based on dynamic weighted ant colony algorithm. Features are treated as graph nodes to construct graph model. Ant colony algorithm is used to select features while support vector machine classifier is applied to evaluate the performance of feature subsets, and then feature pheromone is computed and updated based on the evaluation results. At the same time, dynamic weighted is introduced into ant colony algorithm for feature selection in order to keep a good balance between the convergence rate and the stagnant phenomenon. The experimental comparison verifies that the algorithm has good classification accuracy and time efficiency.
Keywords :
graph theory; optimisation; pattern classification; support vector machines; dynamic weighted ant colony algorithm; feature selection; graph model; graph nodes; support vector machine classifier; Accuracy; Algorithm design and analysis; Classification algorithms; Heuristic algorithms; Machine learning; Machine learning algorithms; Support vector machines; Ant colony algorithm; Feature selection; Mutual information(MI); Support vector machine (SVM);
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2010 International Conference on
Conference_Location :
Qingdao
Print_ISBN :
978-1-4244-6526-2
DOI :
10.1109/ICMLC.2010.5581009
