<![CDATA[Abstract Gold has two functions, jewelry, and investment. Investments made by the community are expected to be lucky, therefore it is necessary to predict when gold is bought and sold. This research is a follow-up study on using a forecasting model that is easy, simple, and has high accuracy. This study compares the single exponential smoothing (SES) model in previous studies and the exponential smoothing-state space or better known as ETS (error, trend, and seasonal). This study uses the same data as previous studies. The comparison criteria used for the accuracy of the SES and ETS models are AIC and BIC, while the comparison of forecasting accuracy uses MAPE and RMSE. The results of this study conclude that the ETS model (M, N, N) is more accurate than the SES model and can be used for short-term forecasting and then the model should be updated so that the model gets the latest information about the data so that in predicting the daily gold price for the coming period. can be accurate. The ETS model used has an exponential coefficient (α) of 0.9999 with AIC and BIC values of 2902,143 and 2912,882 with MAPE and RMSE forecasting accuracy values of 0.6513446 and 15.01525, respectively. Forecasts for the next three periods with the ETS(M, N, N) model fluctuate around the price of 1600 ($ 1600 USA per troy ounce).]]>
Copyrights © 2024
