PERBANDINGAN METODE FUZZY TIME SERIES MARKOV CHAIN DAN FUZZY TIME SERIES CHENG UNTUK PERAMALAN DATA INFLASI

Abstract

The value of inflation can determine decision-making for economic actors. Therefore, in order for entrepreneurs
to plan their business well, accurate inflation forecasting is necessary. Fuzzy Time Series (FTS) is a concept to
solve forecasting problems if historical data is formed into linguistic values. This method has advantages, namely
the calculation process does not require a complex system and is able to solve the problem of forecasting historical
data in the form of linguistic values. Fuzzy Time Series Cheng (FTS-Cheng) method has a slightly different
method of determining intervals, while the interval determination in Fuzzy Time Series Markov Chain (FTS-MC)
method is the same as other FTS methods. FTS-MC is a combined method of FTS with Markov chain stochastic
processes. In this paper, we discuss forecasting inflation data using FTS-MC and FTS-Cheng methods. This study
uses monthly data on Indonesian inflation from January 2017 to December 2021. FTS-MC method has a MAPE
value of 9.41% and FTS-Cheng method has a MAPE value of 32.25%. Based on the criteria for the accuracy of
MAPE, the forecasting value using FTS-MC method meets the very good forecasting results and the forecasting
results using FTSC method meet the sufficient forecasting results. Based on the MAPE value obtained, a better
forecasting method for the case study of Indonesian inflation data in 2017-2022 is FTS-MC method.