Adaptive Input Selection and Evolving Neural Fuzzy Networks Modeling

Alisson Marques Silva,Andre Lemos,Fernando Gomide,Walmir Caminhas

doi:10.1080/18756891.2015.1129574

Abstract

AbstractThis paper suggests an evolving approach to develop neural fuzzy networks for system modeling. The approach uses an incremental learning procedure to simultaneously select the model inputs, to choose the neural network structure, and to update the network weights. Candidate models with larger and smaller number of input variables than the current model are constructed and tested concurrently. The procedure employs a statistical test in each learning step to choose the best model amongst the current and candidate models. Membership functions can be added or deleted to adjust input space granulation and the neural network structure. Granulation and structure adaptation depend of the modeling error. The weights of the neural networks are updated using a gradient-descent algorithm with optimal learning rate. Prediction and nonlinear system identification examples illustrate the usefulness of the approach. Comparisons with state of the art evolving fuzzy modeling alternatives are performed to evaluate ...

Highlights

Evolving fuzzy systems constitute a class of systems whose structure and parameters can be adapted concurrently in a stepwise manner using data streams
This paper extends the X-eNFN-AFS approach originally introduced in Ref. 9
The parameters and the structure of the models evolve as each data sample is input

Summary

Introduction

Evolving fuzzy systems constitute a class of systems whose structure and parameters can be adapted concurrently in a stepwise manner using data streams. Adaptation proceeds continuously and gradually by means of incremental learning. Incremental learning enables fast processing with low storage cost because samples in data streams are processed only once and can be discarded[1]. While learning enables continuous and gradual knowledge update changing the structure and parameters of models, it maintains the relevant knowledge of objects learned so far[2]. A limitation of the current evolving fuzzy modeling approaches concerns the non-flexibility to select the input variables as the system structure and parameters are adapted. The input variables are chosen using a priori knowledge or a selection technique. The input variables remain the same[3]

Methods

Results

Conclusion