Iterative Fuzzy C-means Clustering Algorithm & K-Nearest Neighbor and Dictionary Data Based Ensemble TSK Fuzzy Classifiers

A new ensemble TSK fuzzy classifier (i,e. IK-D-TSK) is proposed. First, all zero-order TSK fuzzy sub-classifiers are organized in a parallel way, then the output of each sub-classifier is augmented to the original (validation) input space, finally, the proposed Iterative Fuzzy C-Means  (IFCM) clustering algorithm generates dictionary data on augmented validation dataset, and then KNN is used to predict the result for test data. IK-D-TSK has the following advantages: the output of each zero-order TSK subclassifier is augmented to the original input space to open the manifold structure in parallel, according to the principle of stack generalization, the classification accuracy can be improved; Compared with traditional TSK fuzzy classifiers which trains sequentially, IK-D-TSK trains all the sub-classifiers in parallel, so the running speed can be effectively guaranteed; Because IK-D-TSK works based on dictionary data obtained by IFCM & KNN, it has strong robustness. The theoretical and experimental results show that IK-D-TSK has high classification performance, strong robustness and high interpretability.