A novel interval type-2 fuzzy K-nearest neighbor classifier for remotely sensed hyperspectral image classification

Type-1 fuzzy set (T1 FS) theory was first introduced by Zadeh in 1965 and has been successfully applied to many fields. In recent years type-2 fuzzy set (T2 FS) system has been attracting research interests and some good results were reported. Type-2 fuzzy sets as those sets whose membership grades themselves type-1 fuzzy sets. Therefore, a type-2 fuzzy system can model the randomness and fuzziness of data set simultaneously. In this paper, a novel interval type-2 fuzzy K-nearest neighbor (IT2 FKNN) classifier, namely NIT2 FKNN, is proposed and applied to the classification of hyperspectral images. Experimental results show that the proposed IT2 FKNN classifier can obtain better and more stable results than FKNN and KNN classifiers.