Stock Price Prediction of Bank Rakyat Indonesia Using an Ensemble Stacking Model of K-Nearest Neighbors (KNN) and Support Vector Machine (SVM)

Abstract

Need for future price forecasts by investors faces difficulties in achieving accurate predictions because market changes exist. Standard single models do not accurately model stock market behaviors because of their complex nature. The problem solution implemented by the study involves combining K-Nearest Neighbors (KNN) and Support Vector Machine (SVM) to create ensemble stacking. Research personnel collected Bank Rakyat Indonesia's (BRI) historical stock price data using KNN and SVM models. Studio performance delivers superior predictive results with lower error rates than KNN and SVM models that operate individually. Study results demonstrate stacking technology produces the most desirable results for stock market price prediction.