The Indonesia Stock Exchange Composite Prediction based on Macroeconomic Indicators using ARIMA, LSTM, and ANN

The capital market will make it uncertain to collect investments in the future because it is constantly shifting. This ambiguity shows that there are risks that investors must accept. Investors typically aim to maximize their expected return considering the risk taken. In order to minimize investment risk, it is critical to measure the movement of the stock market. Indonesia Stock Exchange (IDX) indicator values are a reflection of the state of the stock market. The focus of this research is to forecast the IDX using macroeconomic factors that reflect the current state of the economy. Interest rates, inflation rates, money supply, and exchange rates are the macroeconomic variables. Bank Indonesia provided secondary data for this study. ARIMA, LSTM, and ANN are the approaches used in this study. This approach is thought to be useful for handling issues, representing complex patterns of each macroeconomic variable, and predicting values with lowest error. Based on the research results, the Root Mean Square Error (RMSE) for the ARIMA (0.087), LSTM (0.0298), and ANN (0.0466). The best time to predict is 2 months before. The RMSE value proves that macroeconomic variables are good indicators to predict the IDX movement.