USE OF BOX AND JENKINS TIME-SERIES ANALYSIS TO ISOLATE THE IMPACT OF A PAVEMENT IMPROVEMENT POLICY

The impact of a freeway pavement improvement policy, in terms of its actual effects on freeway volumes and the duration of those effects as well as the effect of this intervention on the accuracy of model forecasts, is analyzed. Univariate and multivariate time-series models are used for the impact analysis. Both types of models exhibit accurate volume forecasts. Two approaches to isolating the impact effect are used. One approach is to fit a univariate time-series model to predict trends prior to the intervention. A second approach fits a time-series model with intervention variables to the entire series. Both models yield similar results in terms of measured volume impacts. (Author)