Best Estimate Curve for Traffic Volume Forecasting

This paper i nvestigates t he application of two forecasting t echniques, namely regression and logistic r egression, to predi ct daily traffic v olume for t he Egyptian intercity roads. The data of three permanent traffic volume stations was selected in this paper. A total of 14-year traffic data for the selected stations (from 1990 t o 20 03) were u sed i n t he development of the two mo dels to predi ct traffic v olume. All of the d ata set s were best r epresented by both regr ession m odel and logistic regres sion m odel. The data s et containing average annual and ave rage monthly volume for y ears 1990 t hrough 2001 was used to f it a curve model. The resulting models were used to f orecast traffic volumes for year 2002 and 2003. The f orecasted traffic vo lumes f or t he m odels were then compared with t he act ual traffic vo lumes. The best estimate curve is chosen according to minimum coefficient of determination R and standard 2 deviation o for bot h e stimate a nd f orecasting data.