Application of Cheng’s Fuzzy Time Series in World Crude Oil Price Prediction

Abstract

Fuzzy time series is a new concept that can be used to predict an event using historical data. Historical data is processed using the principles and logic of fuzzy sets. The aim of this study is to predict world oil prices. The historical data used was from 3 January 2022 to 30 June 2023. This article discusses Cheng’s Fuzzy Time Series application. Determining the number of fuzzy class intervals uses 3 approaches, namely using the Sturges formula, Average-based and Partial Frequency Density. The 3 approaches used will be compared. Fuzzy Time Series with the Sturges formula produces a MAPE of 10.54% and a MSE of 9.13. Average-based Fuzzy Time Series produces a MAPE of 7.64% and a MSE of 5.15. Partial Frequency Density Fuzzy Time Series produces a MAPE of 8.09% and a MSE of 5.99. The results of this study state that Cheng’s Average based Fuzzy Time Series has the best accuracy in predicting world oil prices.