<![CDATA[This study aims to analyze and compare the performance of the Linear Regression and Long Short-Term Memory (LSTM) models in predicting gold prices. Gold price data was processed through a preprocessing stage that included data normalization and dataset splitting using three skenarios: 70:30, 80:20, and 90:10. Model performance was evaluated using Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), and the coefficient of determination (R²). The results show that both models are capable of producing predictions with high accuracy across all testing skenarios. However, the Linear Regression model performs better than the LSTM. The best skenario for Linear Regression was obtained with a 70:30 data split, yielding an RMSE of 24.92, a MAPE of 0.76%, and an R² of 0.9986. Meanwhile, the Long Short performed best in the 80:20 and 90:10 skenarios, with the lowest Root Mean Square Error of 35.84 in the 80:20 skenario, as well as the lowest Mean Absolute Percentage Error of 0.98% and the highest coefficient of determination of 0.9967 in the 90:10 skenario. The results of the study indicate that the gold price data patterns during the study period can be well represented by the Linear Regression model, resulting in a lower prediction error rate compared to the LSTM model. Therefore, Linear Regression can be recommended as the more optimal model for predicting gold prices using the dataset employed in this study]]>
Copyrights © 2026
