Type-2 Fuzzy SARIMA system for real-valued uncertain non-stationary data-intensive seasonal time series

Following the concepts of Mpdf and J̃-plane, in this paper, the authors propose a new Type-2 Fuzzy Probabilistic System - Type-2 Fuzzy SARIMA System for real-valued uncertain non-stationary data-intensive seasonal time series forecasting. The system is implemented in Wireless Soft-Switch (WSS) communication network CAPS forecasting and is compared with the statistical model SARIMA to show that Type-2 Fuzzy SARIMA System can forecast real-valued uncertain non-stationary data-intensive seasonal time series more accurately, timely, along with the infinite maximum supported seasonality and less errors. As the series of algorithms and results prove, Type-2 Fuzzy Probabilistic System is viable in practice, which realises Type-2 Fuzzy Logic Systems evolving from rule-based fuzzy systems to the systems based on Type-2 Fuzzy Probabilistic Model.