Particle swarm optimization-based type-2 fuzzy modeling for ECG signal prediction

Abstract

A hybrid prediction framework based on Particle Swarm Optimization (PSO) and the Interval Type-2 Fuzzy Inference System (IT2FIS) is proposed in this paper. The major contribution is a systematic methodology for calculating the ideal parameter vector to tune the membership functions of a type-2 fuzzy controller. In contrast to type-1 fuzzy controllers, the interval type-2 formulation provides better robustness in uncertain operating conditions by explicitly modelling uncertainty through a Footprint of Uncertainty (FOU) bounded by upper and lower membership functions. Four performance indices mean square error (MSE), root mean square error (RMSE), error mean, and error standard deviation were used to assess the suggested IT2FIS-PSO framework on typical ECG arrhythmia signal data. The baseline IT2FIS produced an RMSE of 0.1198. This improved to 0.0429, a reduction of almost 75.98%, after PSO-based optimization of both membership function bounds. Experimental data show that the proposed optimization technique significantly improves ECG prediction accuracy relative to the unoptimized baseline, and simulation results confirm that PSO consistently finds the optimal parameters for a type-2 fuzzy controller.