An Efficient Fuzzy Rough Approach for Feature Selection
Rough set theory is a powerful tool for feature selection. To avoid the information loss by discretization in rough sets, fuzzy rough sets are used to deal with the continuous values. However, the cost of computation of the approach is too high to be worked out as the number of selected features increases. In this paper, a new computational method is proposed to approximate the conditional mutual information between the selected features and the decision feature, and thus improve the efficiency and decrease the complexity of the classical fuzzy rough approach based on mutual information. Extensive experiments are conducted on the large-sized coal-fired power units dataset with steady state, and the experimental results confirm the efficiency and effectiveness of the proposed algorithm.
KeywordsFuzzy rough sets Feature selection Mutual information Large-sized coal-fired power units
Unable to display preview. Download preview PDF.
- 1.Yu, L., Liu, H.: Feature Selection for High-Dimensional Data: A Fast Correlation-Based Filter Solution. In: Proceedings of the Twentieth International Conference on Machine Learning (ICML-2003), Washington DC (2003)Google Scholar
- 4.Jensen, R., Shen, Q.: Computational Intelligence and Feature Selection: Rough and Fuzzy Approaches. Wiley-IEEE Press (2008)Google Scholar
- 8.Maji, P., Paul, S.: Rough Set Based Maximum Relevance-Maximum Significance Criterion and Gene Selection from Microarray Data. Int. J. Approx. Reason. (2010) (in press)Google Scholar