Abstract
Abstract: The machine learning technique such as random forests and regression trees, arenonparametric methods that it recently used for regression estimation. In these methods, the variance of randomerrors needs to be constant, but that’s not true always, especially, in the financial data. Unfortunately, the financialtime series suffer from the volatility that happens during different periods where the researchers were an effort tosolve this problem by combining the random forest and the regression trees with the GARCH model. In this paper, weuse these methods to estimate the conditional variance of the GARCH model to forecast the exchange rate ofIQD/USD.
