Fuzzy-rough Nearest-neighbor Classification Method: An Integrated Framework

H. Bian and W. Shen (PRC)

Keywords

Machine Learning, Data Mining, Fuzzy-Rough Set, Nearest Neighbor Classification

Abstract

This paper proposes a new fuzzy-rough nearest neighbor(NN) classification method based on the fuzzy-rough set theory which is prove to be able to deal with pattern recognition tasks under partially exposed environment. An integrated framework of K-NN classification algorithms are developed, and comparisons among the conventional fuzzy K-NN, rough-fuzzy K-NN, and fuzzy-rough NN are provided under this integrated framework. Simulation results of the fuzzy-rough NN algorithm are analyzed and the detailed method for unexposed pattern identification is presented.

Important Links:



Go Back


IASTED
Rotating Call For Paper Image