PEMODELAN DATA HARGA CABAI DENGAN PENDEKATAN DERET WAKTU FRAKSIONAL ARFIMA

Abstract

Long-memory is a type of time series data that has a high correlation between long observation times. This can be seen from the autocorrelation function where the lag falls slowly over a long period. Such long-memory data can be modeled in the form of an Autoregressive Fractionally Integrated Moving Average (ARFIMA). One of the data that meets the long-memory criteria is the monthly chili price from March 2017 to April 2023 as much as 73 data. ARFIMA model selection is done by comparing the AIC and BIC values of each candidate model, so that the best model is ARFIMA (1;0.22785;0), this means that the movement of chili prices is influenced by previous prices in the long term

Keywords

• arfima
• geweke and porter-hudak (gph) method
• long-memory