期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2015
卷号：6
期号：3
页码：2023-2026
出版社：TechScience Publications
摘要：In this paper, we proposed to select the optimum number of features from the datasets using genetic programming. For selecting the optimal number of features we run the GP life cycle for 50 percent of the generation and select the features present in the best classifiers and form their optimal set. After that we remove the other features in the classifier from the features present in the optimal set and run the process till the last generation. Then we select the features present in the best classifier after the last generation and called those features the optimal feature set.
关键词：feature selection; genetic programming;multiobjective fitness function
