Forecasting Market Index of Stock Exchange of Thailand, Malaysia, and Singapore with the Gaussian Process Regression Model

This paper aims at examining the predictability of the Gaussian Process Regression Model with an interest to test its performance in three Asian stock markets. This study compares the forecasting performance of the models with a lag from 1 to 5. The comparison is based on the root-mean-square error (RMSE) and the mean absolute error (MAE). The prediction results from the GPR are then compared to the Autoregressive model (AR). The results show that the Gaussian Process Regression (GPR) is a good model for capturing the stock market index in three Asian stock markets. By testing process, it is shown that the GPR model with lag 1 is the best model for the stock market in Malaysia and Thailand, and the GPR model with lag 5 is suitable for the stock market in Singapore.