<![CDATA[In risk analysis and investment decision-making, stock market volatility is very important, especially in the Composite Stock Price Index (IHSG), which is constantly changing and influenced by many economic factors. To predict IHSG volatility, this study uses a Hybrid GARCH–Random Forest approach. This approach combines the strength of Random Forest in modelling non-linear relationships with the ability of the GARCH model to identify financial data heteroscedasticity. To predict volatility on the following day, data from the IHSG closing price is processed into logarithmic returns and then used to estimate volatility using the GARCH(1,1) model. According to performance evaluations, this hybrid model is capable of providing highly accurate predictions. In addition, the model was tested in a volatility classification scheme into three categories: low, medium, and high. In regression testing, an MSE value of 0.000386 and an RMSE value of 0.01965 were obtained, indicating that the volatility prediction error was very low. The accuracy, recall, and f1-values were between 0.99 and 1.00. The results show that the Hybrid GARCH–Random Forest approach is very effective in modelling IHSG volatility. This approach can also be a reliable tool to support risk analysis and decision-making strategies in financial markets.]]>
Copyrights © 2026
