Analisis dan Prediksi Harga Saham Menggunakan Algoritma Long Short-Term Memory (LSTM)

Stock price prediction is a crucial aspect in the world of capital markets, as it enables investors to make more informed and rational investment decisions. In this study, the Long Short-Term Memory (LSTM) method, which is a form of recurrent neural network (RNN) algorithm, is applied to predict the stock prices of four companies from the S&P 500 index, namely Apple (AAPL), Amazon (AMZN), Google (GOOG), and Microsoft (MSFT). The data is sourced from Yahoo Finance, covering the period from January 2012 to June 2025. The data is divided into 95% training data and 5% test data. The normalization process is then carried out using MinMaxScaler to optimize the model’s learning capabilities.
The results of this study show that the LSTM model can provide fairly accurate predictions. Evaluation is done using Root Mean Squared Error (RMSE), where the RMSE values obtained are 17.03 for Apple stocks, 15.52 for Amazon stocks, 11.08 for Google stocks, and 17.37 for Microsoft stocks. Visualization of the prediction results shows that the model can follow the actual price trend in general, although there are some small deviations. This research contributes to proving that the LSTM
method is reliable in capital market analysis, especially in time series-based stock price prediction.