HYBRID NEURO-FUZZY CLASSIFIER BASED ON NEFCLASS MODEL

The paper presents hybrid neuro-fuzzy classifier, based on NEFCLASS model, which was modified. The presented classifier was compared to popular classifiers - neural networks and k-nearest neighbours. Efficiency of modifications in classifier was compared with methods used in original model NEFCLASS (learning methods). Accuracy of classifier was tested using 3 datasets from UCI Machine Learning Repository: iris, wine and breast cancer wisconsin. Moreover, influence of ensemble classification methods on classification accuracy was presented.