An Interval Type‐3 Fuzzy Elman Neural Network Model for Chaotic Time‐Series Prediction

Chaotic time series prediction plays a crucial role in various fields, including economics, meteorology, and nonlinear control. However, the strong nonlinearity and high sensitivity to initial conditions make this problem difficult. In this study, we propose a new predictive model based on an Elman neural network combined with interval Type‐3 fuzzy Elman neural network (IT3FENN). The IT3FENN structure is designed to leverage the dynamic state memory capability of the Elman network while enhancing its ability to handle uncertainty and noise through the membership function mechanism of interval Type‐3 fuzzy logic. An adaptive learning algorithm optimizes the system parameters to ensure fast convergence and low forecast errors. Simulation results on benchmark chaotic time‐series datasets as well as a real‐world traffic dataset demonstrate that the proposed approach outperforms the T1FLS‐ENN and IT2FLS‐ENN models in terms of forecasting accuracy and noise robustness. The results demonstrate the potential application of IT3FENN in complex nonlinear forecasting problems.