ARIMA Time Series Analysis in Forecasting Daily Stock Price of Chittagong Stock Exchange (CSE)

International Journal of Research and Innovation in Social Science (IJRISS) | Volume V, Issue VI, June 2021 | ISSN 2454–6186

ARIMA Time Series Analysis in Forecasting Daily Stock Price of Chittagong Stock Exchange (CSE)

Tasnim Uddin Chowdhury¹, Md. Shahidul Islam^2
1Assistant Professor (Finance Discipline), Department of Business Administration, Premier University, Chattogram, Bangladesh,
²Divisional Officer, Service Engineering Division, Bangladesh Forest Research Institute, Chattogram, Bangladesh

Abstract- The aim of the study is to examine the nature of daily share price and select a suitable ARIMA model to forecast the future daily share price from the previous daily share price of Chittagong Stock exchange (CSE). A random sampling method has been followed to collect the closing price of 60 companies for the period of January 2019 to December 2019 (241 trading days). Durbin-Watson test has been conducted to find the autocorrelation in each of the share prices. Then the Augmented Dickey-Fuller test has been applied to test the stationary of data and the Autocorrelation function (ACF) and Partial Autocorrelation function (PACF) has been calculated to determine the lag value of moving average MA(q) and autocorrelation AR(p) based on Ljung-Box Test Q, root mean square error, mean absolute error, mean absolute percent error and R-square values. After selecting ARIMA (p,d,q) model, forecasted values for each of the shares are calculated for the next 22 trading days of January 2020. Then a comparison has been made between the forecasted prices and the actual share prices by using the Goodness-of-fit Test, Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), Mean Square Error (MSE) to validate the model. The result shows that the ARIMA model is applicable to forecast the daily share price of CSE.

Keywords- Time series analysis, Autoregressive Integrated Moving Average (ARIMA), Durbin-Watson test, Augmented Dickey-Fuller test, Goodness of fit test.

I. BACKGROUND

Stock market analysts prefer to conduct both fundamental analysis and technical analysis for taking investment decision. The former one involves analyzing the fundamental factors like- the company’s financial position, operating performance, dividend payment history, market competition etc. that affects the future earning capacity of the company and identifying the mispriced stocks by determining its intrinsic value. On the contrary, the latter one undertakes the analysis of historical price movements to get a forecast of future price. To conduct technical analysis different charts like- bar chart, line chart, candle stick chart etc. is used. By using the charts different patterns of price movement is identified to get a sense of price trend that may prevail in the future. Jarrett and Kyper (2005) find predictable patterns in monthly stock prices examining the daily returns for more than 50 firms from American Stock Exchanges. But, unlike technical analysis, time series analysis is a statistical tool which is applied to forecast the price for a particular period of time in the future based on the historical series of data regarding price at constant time interval.

Read More