Malaysia Stock Price Prediction Using ARIMA Model and Geometric Brownian Motion with Monte Carlo Simulation

Abstract

Stock market acts as a platform for stock trader to issues and deals in stocks where there is offering the probability of financial gains with familiar procedure of buying undervalued stock followed by selling at inflated price. The information of prediction acts an essential consideration to adopt a trading strategy. Thus, this study purposed to make stock closing price prediction by implementing Autoregressive Integrated Moving Average (ARIMA) and geometric Brownian motion (GBM) models, and estimate the highest, average lowest paths predictions for each model by Monte Carlo simulation (MCS). The historical daily closing price of three Malaysian plantation companies stock included of IOI Corporation Berhad, Kuala Lumpur Kepong Berhad and Sime Darby Plantation Berhad are used to fit the models. Predictions are evaluated with mean forecast error (MFE), root mean square error (RMSE) and mean absolute percentage error (MAPE). Based on the results, ARIMA outperformed GBM and the combination of each model with MCS had lower average MAPE values on average path prediction among three stocks compared to that of each single model respectively in the majority. Hence, it is concluded application of MCS improved accuracy of prediction for both methods.