Prediction of Stock Price using Autoregressive Integrated Moving Average Filter ((ARIMA (p,d,q))
Abstract – The financial system of any economy is seen to be divided between the financial intermediaries (banks, insurance companies and pension funds) and the markets (bond and stock markets). This study was designed to look at the behavior of stock price of Nigerian Breweries Plc with passage of time and to fit Autoregressive Integrated Moving Average Filter for the prediction of stock price of the Nigerian Breweries Plc. The data were collected from Nigerian Stock exchange and Central Securities Clearing System (CSCS).Time plot was used to detect the presence of time series components in the daily stock prices of Nigerian breweries from 2008 to 2012 and to check if the series is stationary. The structure of dependency was measured by using autoovariance, the auto-correlation and partial autocorrelation. An autoregressive model and moving average model were fitted to stationary series to predict the future stock prices. Alkaike Information Criteria (AIC) was used to determine the order of the fitted autoregressive model. Diagnostic checks were carried out to assess the fit of the fitted autoregressive model. The time plot showed an irregular upward trend. A first difference of the non stationary series made the series stationary. The plots of the Autocorrelation and Partial Autocorrelation showed that stationary has been introduced into the original non-stationary series in which most of the Plotted points decaying to zero sharply. The plot of Akaike Information Criterion showed that the order of the fitted autoregressive model was 8. The ARIMA model diagnostic check showed that the fitted ARIMA model had a reasonable fit for the original series. Predicted stock price ranges from 138.66 to 141.49.
