Perbandingan Kinerja Long Short-Term Memory dan Gated Recurrent Unit dalam Prediksi Harga Saham McDonald’s

Muhammad Ridho Alfarid (Universitas Islam Indonesia)
Qonita Husnia Rahmah (IPB University)

JAMBURA JOURNAL OF PROBABILITY AND STATISTICS
Vol 7, No 1 (2026): Jambura Journal of Probability and Statistics

Social movements occurring at the global level can influence public perspectives and actions toward a company and affect its stock value. This study aims to analyze the impact of the social boycott movement on McDonald’s (McD) stock price by comparing the performance of the Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) models, with hyperparameter tuning conducted using Optuna. The data used consist of McD’s daily closing stock prices from January 31, 2015, to January 31, 2025, obtained from www.finance.yahoo.com. The results show that the LSTM model without hyperparameter tuning provides the most optimal performance, achieving a Mean Absolute Percentage Error (MAPE) of 1.79% on the training data and 1.47% on the test data. This model is effective in identifying changes and forecasting McD’s stock price before and after the boycott