Extreme Learning ANFIS for classification problems

This paper compares Extreme Learning ANFIS (ELANFIS) with conventional ANFIS for classification problems. ELANFIS is a hybrid Fuzzy System based on Extreme Learning Machines. It combines the linguistic knowledge representation of a Fuzzy System with the fast learning speed of ELMs. This paper also proposes the use of a zero order ELANFIS for classification tasks and compares it with first order Fuzzy Systems. The results show that zero order ELANFIS gives lesser classification error with faster learning speed as compared to the mentioned methods since the number of parameters to tune is lesser.