Ensembles of Fuzzy Classifiers

The use of bagging is explored to create an ensemble of fuzzy classifiers. The learning algorithm used was ANFIS (Adaptive Neuro-Fuzzy Inference Systems). We compare results from bagging to those of a single classifier using both crisp and fuzzy classifier combination methods. Results on 20 data sets show that bagging results in a significantly more accurate classifier.

• Type: Conference
• Year: 2007
• Where: FUZZIEEE
• Authors: Juana Canul-Reich, Larry Shoemaker, Lawrence O. Hall
• Keywords: Adaptive Neuro-fuzzy Inference, FUZZIEEE 2007, Fuzzy Classifier, Fuzzy Classifier Combination, Fuzzy Logic