Further Advanced FS Methods

This chapter introduces two promising areas in feature selection. The first, feature grouping, is developed from recent work in the literature where groups of features are selected simultaneously. By reasoning with fuzzy labels, the search process can be made more intelligent allowing various search strategies to be employed. The second, ant-based feature selection, seeks to address the non-trivial issue of finding the smallest optimal feature subsets. This approach to feature selection uses artificial ants and pheromone trails in the search for the best subsets. Both of these developments can be applied within feature selection in general, but are applied to the specific problem of subset search within FRFS in this book.

R. Jensen and Q. Shen. Computational Intelligence and Feature Selection: Rough and Fuzzy Approaches. IEEE Press/Wiley & Sons.