Indofood CBP Sukses Makmur Tbk Stock Price Prediction Using Long Short-Term Memory (LSTM)

Abstract

Fluctuating stock price movements are a challenge in the investment world, so an accurate prediction model is needed to assist decision making. This study aims to evaluate the ability of the LSTM model to predict ICBP stock prices based on historical data and will compare the results of the LSTM model predictions with actual stock price movements to determine the extent to which this model is able to capture trends and patterns of ICBP stock prices. The results show a comparison of the original price and the predicted price indicating that the model can follow market trends, although there are still deviations at some points, especially when volatility is high. Residual analysis shows a distribution of prediction errors that is close to normal, indicating that the model does not experience significant bias. In addition, evaluation of the loss function on the training and validation data confirms that the model has converged well. In the performance evaluation, the model is able to capture stock movement patterns quite well, indicated by the Mean Absolute Error (MAE) value of 0.0231, Root Mean Squared Error (RMSE) of 0.0305, and Mean Absolute Percentage Error (MAPE) of 19.21%.