<![CDATA[This study aims to develop a web-based stock price prediction system using the Long Short-Term Memory (LSTM) algorithm to forecast the highest and lowest prices of stocks listed in the LQ45 index. LSTM was chosen for its ability to recognize long-term patterns in time series data and its more stable performance compared to methods such as ARIMA and GRU. The system features an interactive interface and user activity logging to enhance usability and user experience. Evaluation results show that the LSTM model performs well, with MAPE below 3% and RMSE values varying according to stock volatility. The best results were achieved by ACES, with RMSE values of 28,772 (High) and 27,142 (Low), and MAPE of 2,19% and 2,2%, while AMMN had the highest error rates with RMSE values of 247,154 and 281,926, and MAPE of 2,42% and 2,79%. The system successfully delivers real-time predictions through a responsive and user-friendly web interface.]]>
Copyrights © 2025
