Fuzzy Adaptive Recurrent Counterpropagation Neural Networks: A Tool for Efficient Implementation of Qualitative Models of Dynamic Processes

François E. Cellier
Institute of Computational Science
ETH Zürich
CH-8092 Zürich
Switzerland
Email: FCellier@Inf.ETHZ.CH
YaDung Pan
Associate Professor
Department of Mechanical Engineering
LungHwa University of Science and Technology
300, Wan Shou Road, Section 1
Kueishan, Taoyuan County 333
Taiwan
Email: YDPan@Mail.LHU.Edu.TW

Keywords

Modeling
Simulation
Time-Series Analysis
Forecasting
Mixed Quantitative and Qualitative Models
Neural Networks
Fuzzy Systems
Learning Systems
Artificial Intelligence

Abstract

In this paper, a new method for efficient implementation of qualitative and mixed quantitative/qualitative models of time-varying (dynamic) processes is shown. It involves a special brand of recurrent neural networks coined Fuzzy Adaptive Recurrent Counterpropagation Neural Network (FARCNN). This is the first paper on FARCNNs ever written. It explains the methodology in detail, and ends with an illustrative example of their use.

Interested in reading the full paper? (16 pages, 1,127,212 bytes, pdf)

Homepage