Metode Fuzzy Time Series Average Based untuk Peramalan Tingkat Suku Bunga Acuan Bank Indonesia

Abstract

Forecasting the Bank Indonesia benchmark interest rate (BI Rate) is one of the essential indicator in monetary policy and national economic, as its fluctuations directly affect inflation, the exchange rate, the banking sector, and economic activity in general. Therefore, forecasting the BI Rate is essential for informing decision-making among economic stakeholders. This study aims to implement the Fuzzy Time Series Average Based method to forecast the BI Rate, as this method has the capability to determine the effective interval length, thereby producing forecasts with a high level of accuracy. The data used are secondary BI Rate on a quarterly basis, January, April, July, October during the period April 2016 to April 2025. The results show that the Fuzzy Time Series Average Based method produce a Mean Absolute Percentage Error (MAPE) of 0.08%, which indicates that this method has an accuracy level of 99.92%. The very small MAPE value (<10%) indicates that this method as a highly accurate forecasting techniques. The forecast result for the period of July 2025 is 5.61%, indicates that the model detects a downward trend from the previous period.