Forecasting the Jakarta Composite Index (IHSG) Using the Moving Average Method

Abstract

Financial market indices play a crucial role in reflecting economic conditions and supporting investment decision-making. In Indonesia, the Jakarta Composite Index (IHSG) serves as a key benchmark for evaluating overall stock market performance. Due to its dynamic and volatile nature, accurate forecasting of IHSG movements remains a challenging task in financial time series analysis. Many recent studies employ complex machine learning and deep learning models, which often require substantial computational resources and lack interpretability, limiting their practical adoption. Motivated by the need for transparent and easily implementable forecasting approaches, this study investigates the use of the Simple Moving Average (SMA) method as a baseline model for forecasting the IHSG. The main contribution of this research lies in providing a systematic evaluation of the moving average method using different window sizes and standard error metrics. Historical IHSG data are preprocessed, analyzed descriptively, and divided into training and testing datasets. Short-term forecasts are generated by applying the SMA model with varying window configurations. The performance of the proposed approach is evaluated using Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE). The results demonstrate that the moving average method is capable of capturing the general trend of the IHSG, with forecasting accuracy strongly influenced by the choice of window size. Future work may focus on integrating additional forecasting techniques, incorporating exogenous variables, and developing hybrid or adaptive models to further enhance prediction accuracy and robustness.