Integrating Machine Learning Algorithms into Decision Support Systems for Predicting BBNI Stock

Abstract

This study focuses on integrating machine learning models into a Decision Support System (DSS) for predicting BBNI stock prices and generating actionable investment recommendations. The research aims to address the challenges of stock price prediction by evaluating the performance of four models: Multiple Linear Regression (MLR), Support Vector Machine (SVM), Random Forest (RF), and Artificial Neural Networks (ANN). The methodology involves collecting and preprocessing BBNI stock data from 2019 onward, training and evaluating the models using metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and R-squared, and incorporating the best-performing model into the DSS. The findings reveal that MLR outperformed other models, achieving an MAE of 18.29, RMSE of 23.73, and an R-squared value of 0.9995, indicating high predictive accuracy. RF performed slightly worse but remained competitive, while SVM and ANN exhibited poor results due to limitations in handling complex patterns or tuning issues. The DSS, powered by MLR, successfully generated buy, sell, or hold recommendations based on stock price predictions, with investment simulations confirming its reliability. This study contributes to the field of financial decision-making by demonstrating the effectiveness of MLR in stock price prediction and DSS integration. However, limitations include reliance on historical data and potential model bias. Future research should explore hybrid models and advanced techniques such as deep learning to enhance predictive capabilities. The proposed DSS offers a practical tool for investors, combining robust machine learning insights with user-friendly decision-making support.