Fuzzy inference based autoregressors for time series prediction using nonparametric residual variance estimation

Abstract

We apply fuzzy techniques for system identification and supervised learning in order to develop fuzzy inference based autoregressors for time series prediction. An automatic methodology framework that combines fuzzy techniques and statistical techniques for nonparametric residual variance estimation is proposed. Identification is performed through the learn from examples method introduced by Wang and Mendel, while the Marquard-Levenberg supervised learning algorithm is then applied for tuning. Delta test residual noise estimation is used in order to select the best subset of inputs as well as the number of linguistic labels for the inputs. Experimental results for three time series prediction benchmarks are compared against LS-SVM based autoregressors and show the advantages of the proposed methodology in terms of approximation accuracy, generalization capability and linguistic interpretability.