Stock Price Prediction of TLKM and BBCA Using SVR, Random Forest, and LSTM

Abstract

Stock price prediction is a significant challenge in financial market analysis due to the influence of many dynamic and nonlinear factors. This study compares three machine learning methods, namely Support Vector Regression (SVR), Random Forest (RF), and Long Short-Term Memory (LSTM), in predicting the stock prices of PT Telekomunikasi Indonesia Tbk (TLKM) and PT Bank Central Asia Tbk (BBCA) for the period 2019–2023. Data were obtained from Yahoo Finance (15). Model Performance Comparison for TLKM Stock, the testing results show that LSTM achieved RMSE of 0.040093, MAE of 0.029849, and R² of 0.905330, indicating the best performance compared to SVR and Random Forest. Random Forest had an RMSE of 0.138295 and an MAE of 0.114423, while SVR had the highest RMSE at 0.238130. Model Performance Comparison for BBCA Stock: The testing results show that LSTM achieved RMSE of 0.055023, MAE of 0.048056, and R² of 0.657730, indicating the best performance compared to SVR and Random Forest. Random Forest had an RMSE of 0.252568 and an MAE of 0.236129, while SVR had the highest RMSE at 0.6211363. This confirms the LSTM model's effectiveness in capturing temporal patterns in stock data (5), (9), (13).