Attribute Reduction Algorithm Based on Improved Information Gain Rate and Ant Colony Optimization
Solving minimal attribute reduction (MAR) in rough set theory is a NP-hard and nonlinear constrained combinatorial optimization problem. Ant colony optimization (ACO), a new intelligent computing method, takes strategies of heuristic search, which is characterized by a distributed and positive feedback and it has the advantage of excellent global optimization ability for handling combinatorial optimization problems. Having considered that the existing information entropy and information gain methods fail to help to select the optimal minimal attribute every time, this paper proposed a novel attribute reduction algorithm based on ACO. Firstly, the algorithm adopts an improved information gain rate as heuristic information. Secondly, each ant solves a problem of minimum attributes reduction and then conduct redundancy test to each selected attribute. What’s more, redundant detection of all non-core attributes in the optimal solution will be perfomed in each generation. The result of the experiment on several datasets from UCI show that the proposed algorithms are more capable of finding the minimum attribute reduction and can faster converge and at the same time they can almost keep the classification accuracy, compared with the traditional attribute reduction based on ACO algorithm.
KeywordsRough set Ant colony optimization Information gain rate Attribute reduction
This work has been supported by National Natural Science Foundation of China (61363029, 61572146, U1711263), Science Foundation of Guangxi Key Laboratory of Trusted Software (kx201515), and the Foundation of Guangxi Educational Committee (KY2015YB105).
- 2.Ding, H., Ding, S.F., Li-Hua, H.U.: Research progress of attribute reduction based on rough sets. Comput. Eng. Sci. 32(6), 92–94 (2010). (in Chinese)Google Scholar
- 5.Zhai, J.H., Liu, B., Zhang, S.: A feature selection approach based on rough set relative classification information entropy and particle swarm optimization. CAAI Trans. Intell. Syst. 12(3), 397–404 (2017). in ChineseGoogle Scholar
- 8.Duan, H., Wang, D., Yu, X.: Review on research progress in ant colony algorithm. Chin. J. Nat. 28(2), 102–105 (2006). in ChineseGoogle Scholar
- 9.Miao, D., Wang, J.: Information representation of the concepts and operations in rough set theory. J. Softw. 10(2), 113–116 (1999). (in Chinese)Google Scholar
- 10.Skowron, A., Rauszer, C.: The discernibility matrices and functions in information systems. In: Slowinski, R. (ed.) Intelligent Decision Support. Handbook of Applications and Advances of the Rough Set Theory, pp. 331–362. Kluwer, Dordrecht (1992)Google Scholar
- 16.Jensen, R., Shen, Q.: Finding rough set reducts with ant colony optimization. In: Proceedings of 2003 UK Workshop on Computational Intelligence, pp. 15–22 (2003)Google Scholar
- 18.Chen, Y., Chen, Y.: Attribute Reduction Algorithm Based on Information Entropy and Ant Colony Optimization. J. Chin. Comput. Syst. 36(3), 586–590 (2015). in ChineseGoogle Scholar
- 20.Chebrolu, S., Sanjeevi, S.G.: Attribute reduction on continuous data in rough set theory using ant colony optimization metaheuristic. In: Proceedings of International Symposium on Women in Computing and Informatics. ACM, New York, pp. 17–24 (2015)Google Scholar