Pemodelan Harga Emas Berdasarkan Kurs Rupiah Terhadap USD dengan Pendekatan Regresi Polinomial Lokal

Abstract

Modeling gold prices based on the exchange rate of the rupiah against the USD is important because it can be used in making investment decisions as well as a reference for formulating economic policy. This study aims to apply local polynomial regression in modeling gold prices in Indonesia based on the rupiah exchange rate against the USD. In this study, gold price modeling was carried out using nonparametric regression with local polynomials. The data used in the study are monthly data of exchange rates as predictor variables (X) and gold prices as response variables (Y) observed from January 2014 to October 2024. Applying local polynomial regression starts with collecting data, analyzing data descriptively, and then modeling and estimating gold price data in Indonesia based on the rupiah exchange rate against the USD using the R program. The results showed that gold price modeling based on the rupiah exchange rate against the USD was obtained on insample data with the best local polynomial estimator of order 2 with an optimal bandwidth of 800 with a MAPE of 9.85% which was classified as very good while for outsample data the MAPE value was 24.87% so that the model estimate for outsample data was classified as sufficient. Overall, the MAPE value related to the prediction of gold prices in January 2014 - October 2024 is 11.01% which is classified as good.