Comparison of forecasting valuation in commercial banking groups using the ARIMA and ARIMAX Models

Abstract

This study aims to compare stock price forecasting in commercial bank sector by using ARIMA and ARIMAX models. Data of a sample group gathered from Siam Commercial Bank, Kasikorn Bank, and Bangkok Bank from January 2013 to December 2018 in total of 72 months or 6 years and stock prices were forecasted in advance for 1 year from January 2019 to December 2019. Models that used in the study included ARIMA and ARIMAX, by considering Mean Absolute Percentage Error (MAPE) and Root Mean Square Error (RMSE) to compare performance of the models.

Findings of the study found it that stock prices forecasted by ARIMA and ARIMAX models provided minimum MAPE and RMSE that rather higher than actual stock prices. So, the forecast was conducted in additional period that was to include 2018 in order to test whether could ARIMA and ARIMAX models be used to forecast stock prices. The trial suggested that in 2018, both ARIMA and ARIMAX models could forecast stock prices more precisely than it was for 2019 and the stock prices derived from the model were similar to actual stock prices.