Stock Price Forecasting Based on ARIMA Model an Example of Cheung Kong Hutchison Industrial Co.

Abstract

Previous studies in stock prediction mainly researched the ARIMA model, however, scant studies take Hong Kong stock price prediction as a data source for this model. The present study takes Cheung Kong Hutchison Industrial Co. as an example. With conducting descriptive analysis and the ARIMA model, this study goes through four steps: 1) Stationarity Processing: the time series data are pre-processed with differential operations to be stationary, and the series are tested for stationarity by the unit root (ADF) test. 2) Order determination: the order can be observed by drawing autocorrelation and partial autocorrelation plots. The ARIMA model's p and q values can then be calculated through SPSS. 3) Validity tests: A Q-value Table is set to determine whether the residual term conforms to the white noise distribution. Moreover, a Ljung-Box test is performed to determine whether the residual series correlated. 4) Prediction: The stock prices in the next few trading days are predicted and analyzed according to the model. The main finding of this research suggests that the price of the stock will decline at a relatively gradual pace from 48.55 to 48.51 over the next five trading days. Moreover, the ARIMA model predicts preciously this time with an error of no more than 3%. The findings can inspire researchers with the application of the ARIMA model, which can also offer suggestions for traders to obtain returns in the HK stock market.