<![CDATA[The Jakarta Composite Index (JCI) is a key indicator in assessing the performance of the Indonesian capital market. Dynamic stock price fluctuations require accurate prediction methods to assist investors in decision making. This study aims to compare the performance of four prediction algorithms, namely Support Vector Machine (SVM), K-Nearest Neighbors (K-NN), Extreme Gradient Boosting (XGBoost), and Autoregressive Integrated Moving Average (ARIMA) in predicting the closing price of JCI. The data used is the JCI daily historical data for the 2019-2024 period obtained from Yahoo Finance. The research process includes data pre-processing, prediction model implementation, model training and testing, and performance evaluation using Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) metrics on a 0-1 normalization scale.The results showed that the ARIMA model provided the most stable results with an RMSE value of 0.4038 and MAE of 0.3945, followed by K-NN and SVM. Although SVM has the lowest MAE value, its RMSE is still higher than ARIMA and K-NN. SVM showed the lowest performance in this experiment. Based on the evaluation results, ARIMA is recommended as the best algorithm in predicting JCI closing price based on historical data. The Jakarta Composite Index (JCI) is a key indicator in assessing the performance of the Indonesian capital market. Dynamic stock price fluctuations require accurate prediction methods to assist investors in decision making. This study aims to compare the performance of four prediction algorithms, namely Support Vector Machine (SVM), K-Nearest Neighbors (K-NN), Extreme Gradient Boosting (XGBoost), and Autoregressive Integrated Moving Average (ARIMA) in predicting the closing price of JCI. The data used is the JCI daily historical data for the 2019-2024 period obtained from Yahoo Finance. The research process includes data pre-processing, prediction model implementation, model training and testing, and performance evaluation using Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) metrics on a 0-1 normalization scale.The results showed that the ARIMA model provided the most stable results with an RMSE value of 0.4038 and MAE of 0.3945, followed by K-NN and SVM. Although SVM has the lowest MAE value, its RMSE is still higher than ARIMA and K-NN. SVM showed the lowest performance in this experiment. Based on the evaluation results, ARIMA is recommended as the best algorithm in predicting JCI closing price based on historical data.]]>
