Perbandingan Tingkat Akurasi Metode Average Based Fuzzy Time Series Markov Chain dan Algoritma Novel Fuzzy Time Series

Abstract

Fuzzy time series method can be applied in predicting the situation in food price development data such as rice. The position of rice as a staple food has resulted in this commodity being one of the indicators of economic growth. The importance of suppressing rice prices so that they are stable can be done by forecasting rice prices in Indonesia in the future. The research method used for forecasting is average based fuzzy time series Markov chain and novel algorithms fuzzy time series. Researchers will compare the two methods in the case of rice prices by looking at the level of accuracy that is better. The data used in this study is the average monthly rice price at the wholesale trade level from January 2015 to March 2021 in units of Rp/Kg as much as 75 data. The results of the comparison of the level of accuracy using the value of Mean Absolute Percentage Error (MAPE), obtained the forecast of the average price of rice at the Indonesian wholesale trade level for average based fuzzy time series Markov chain which is 0.36%, while the MAPE value for novel algorithm fuzzy time series is 0.19%. Based on the MAPE results, it can be concluded that the novel algorithm method fuzzy time series produces a better level of accuracy compared to the method average based fuzzy time series Markov chain.