GenSoFNN: a generic self-organizing fuzzy neural network

Abstract

Existing neural fuzzy (neuro-fuzzy) networks proposed in the literature can be broadly classified into two groups. The first group is essentially fuzzy systems with self-tuning capabilities and requires an initial rule base to be specified prior to training. The second group of neural fuzzy networks, on the other hand, is able to automatically formulate the fuzzy rules from the numerical training data. No initial rule base needs to be specified prior to training. A cluster analysis is first performed on the training data and the fuzzy rules are subsequently derived through the proper connections of these computed clusters. However, most existing neural fuzzy systems (whether they belong to the first or second group) encountered one or more of the following major problems. They are (1) inconsistent rule-base; (2) heuristically defined node operations; (3) susceptibility to noisy training data and the stability-plasticity dilemma; and (4) needs for prior knowledge such as the number of clusters to be computed. Hence, a novel neural fuzzy system that is immune to the above-mentioned deficiencies is proposed in this paper. This new neural fuzzy system is named the generic self-organizing fuzzy neural network (GenSoFNN). The GenSoFNN network has strong noise tolerance capability by employing a new clustering technique known as discrete incremental clustering (DIC). The fuzzy rule base of the GenSoFNN network is consistent and compact as GenSoFNN has built-in mechanisms to identify and prune redundant and/or obsolete rules. Extensive simulations were conducted using the proposed GenSoFNN network and its performance is encouraging when benchmarked against other neural and neural fuzzy systems.

SONFIS: Structure Identification and Modeling with a Self-Organizing Neuro-Fuzzy Inference System

“…We compare the results of our proposal with the results obtained in Ref. 38. SONFIS GA is the same proposed algorithm, applying a Genetic Algorithm to find the optimal parameters.…”

“…Nowadays, constructive methods for flexible modeling and identification have attracted the atention 1,19,32,38 . Several authors have extended the neurofuzzy models in order to endow them with some constructive capabilities.…”

“…Bundesanstalt, Braunschweig, Germany. These data were registred from a Far-Infrared-Laser in a chaotic state 38 ; The Wastewater dataset 32 , Boston dataset and Building dataset were obtained from UCI Machine Learning repository 4 . The first real dataset 38 was obtained from the Forecasting and Time Series Analysis Santa Fe Institute Competition.…”

“…This makes it hard to estimate the overall mapping directly from the consequent of each rule output. Finally, a self-organizing partitioning solves the input space partitioning problem by means of a learning algorithm which automates structure and parameters identification simultaneously 1,19,32,38 . This type of partitioning has a strong dependence on the selforganizing operations.…”

System Identification Using Fuzzy Cerebellar Model Articulation Controllers

“…In order to tackle these disadvantages, several methods, such as online-based clustering (Kasabov & Song, 2002;Tung & Quek, 2002) for the above-mentioned first drawback, Bspline functions (Lane et al, 1992;Wu & Pratt, 1999) and fuzzy concepts (Jou, 1992;Chen, 2001;Guo et al, 2002;Ker et al, 1997;Lai & Wong, 2001;Zhang & Qian, 2000) for the second one, and competitive learning (Chow & Menozzi, 1994), clustering (Hwang & Lin, 1998) and Shannon's entropy and golden-section search (Lee et al, 2003) for the third one, were proposed. Among these approaches, further improvements were implemented by with self-constructing algorithm and Gaussian basis functions.…”