<![CDATA[Stock price prediction is a crucial task in financial market analysis due to its impact on investment decision-making. This study aims to apply the Gated Recurrent Unit (GRU) model to forecast the stock price of ANTM.JK using historical time series data. A total of 12 experimental models were developed by varying data split ratios, window sizes, epochs, and batch sizes to identify the optimal model configuration. Model performance was evaluated using Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE). The results show that the GRU model is capable of predicting stock prices with high accuracy, achieving an accuracy of 98.02%. The RMSE values ranged from 57.78 to 91.09, MAE values ranged from 38.20 to 62.87, and MAPE values ranged from 1.98% to 3.22%. The best-performing model was Model 7, with a 70:30 training–testing split, a window size of 30, 50 epochs, and a batch size of 16, which produced the lowest error values among all models. These findings indicate that GRU is an effective and reliable approach for modeling nonlinear and dynamic stock price time series and has strong potential for supporting financial market analysis and investment decision-making.]]>
Copyrights © 2026
