An Approach for Selective Ensemble Feature Selection Based on Rough Set Theory

摘要

Rough set based knowledge reduction is an important method for feature selection. Ensemble methods are learning algorithms that construct a set of base classifiers and then classify new objects by integrating the prediction of the base classifiers. In this paper, an approach for selective ensemble feature selection based on rough set theory is proposed, which meets the tradeoff between the accuracy and diversity of base classifiers. In our simulation experiments on the UCI datasets, high recognition rates are resulted.