This paper examined three (3) different forecasting techniques in the modeling of twenty (20) years time series data of accident that occurred on Nigeria roads. Road accident, being a rare event is considered for this forecasting process using the techniques of Least Square, Quadratic and Single Exponential Smoothening. Mean Absolute Percentage Error (MAPE), Mean Absolute Deviation (MAD) and Mean Square Deviation (MSD) were adopted as the basis of comparison. Results revealed that the MAPE of 93.10, MAD of 672.55 and MSD of 912.82 attached to the Least Square Model is smaller than that of Quadratic and Exponential Smoothening techniques and hence adjudged the later to be the best model for the forecast.
Article PDFPDF (For Download)