Keywords
How to Cite
Abstract
Outliers could be additive, innovative or mixed, depending on whether it affects a single observation; where it serves as an unusual innovation in the generating process affecting all later observations or when only one characteristic of the series is changed by the innovation. (Hawkins (1980), Denby & Martin (1979) and Draper & Smith (1984)).Modification of Outliers generally improves the forecasting performance of a time series model because it is believed that outliers occur at places where the process generating the series has temporarily broken down which call for the modification of the outliers to compensate for the forecast calculations. Forty-one years data set was collected from the Central Bank of Nigeria; including records on the Gross Domestic Product at current price, Real Estate, Information and Communication, Servicing, Mining and Quarrying and Agriculture sector of the economy..The time series data is subjected to stationarity test and thereafter we determined the level at which the data set is useful for analysis. High sigma*2 estimate and a high AIC coefficient adopted as an accuracy measurement indicate an improvement after adjusting for outliers in the model for the prediction of GDP in Nigeria.