<![CDATA[This scientific paper uses three different models, namely the baseline model, hyperparameter tuning with 4 layers, and hyperparameter tuning with 5 layers, to evaluate the effectiveness of GRU in predicting future stock prices. The performance of each model was measured using Mean Squared Error (MSE), Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE). The baseline model, configured with 50 GRU units, 100 lookbacks, 100 timestamps, batch size 64, and 50 epochs, showed the best performance, achieving MSE 5601.44, MAE 54.06, RMSE 74.84, and MAPE 1.90%. In comparison, the 4-layer model showed slightly higher errors, with MSE 10842.89, MAE 79.65, RMSE 104.13, and MAPE 2.77%, while the 5-layer model produced higher errors, with MSE 14687.76, MAE 91.94, RMSE 121.19, and MAPE 3.22%.]]>
Copyrights © 2026
