Supplementary Developments and Investigations

This chapter offers initial investigations and ideas for further work, which were developed concurrently with the ideas presented in the previous chapters. Firstly, the utility of using the problem formulation and solution techniques from propositional satisfiability for finding rough set reducts is considered. This is presented with an initial experimental evaluation of such an approach, comparing the results with a standard rough set-based algorithm, RSAR. Secondly, the possibility of universal reducts is proposed as a way of generating more useful feature subsets. Thirdly, fuzzy decision tree induction based on the fuzzy-rough metric developed in this book is proposed. Other proposed areas of interest include fuzzy-rough clustering and fuzzy-rough fuzzification optimization.

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