Identification of Type-2 Fuzzy Models for Time-Series Forecasting Using Particle Swarm Optimization

This paper presents the fuzzy model identification framework, where Particle Swarm Optimization (PSO) algorithm has been used as an optimization engine for building Type-2 fuzzy models from the available chaotic Mackeyâ-Glass time-series data. The presented framework is capable of evolving the Membership Functions parameters, Footprint of Uncertainty (FOU) and the rule set to obtain an optimized Type-2 fuzzy model. Four experiments are reported for differently corrupted chaotic time-series data sets. Root Mean square error (RMSE), between the outputs of the designed T2 FLS and the target is used as the performance criterion to rate the quality of solutions and hence demonstrate the performance of the proposed framework.