Fuzzy Time Series Forecasting: Chen, Markov Chain and Cheng Models

Abstract

This paper studies and reviews several procedures for Fuzzy Time Series analysis. Even though forecasting methods have advanced applications in the last few decades, Fuzzy Time Series are common and have a lot of interest because they do not require any statistical assumptions on time series data. Previous research has employed Fuzzy Time Series models to forecast enrollment statistics, stock prices, exchange rates, etc. The major goal of This work is a comparative study of some different methods of forecasting the Fuzzy Time Series among which are the Markov Chain, Chen, and Cheng for Ghabbour Autocars data. Seven statistical criteria have been used for investigating the accuracy of the models. All the calculations were performed using the R software system using the AnalyzeTS R package. The Markov-chain fuzzy time series model showed the highest performance (in all metrics); for instance, in RMSE, MAPE, and U-statistics are 0.013, 0.116, and 1.05 respectively.