TRAFFIC VOLUME FORECASTING MODELS FOR RURAL DESERT TOWNS

This paper develops traffic volume forecasting models for rural desert roads that connect rural towns with a major activity center in Jordan. Data on socioeconomic and demographic characteristics from 48 rural towns were collected along with traffic volume data. Regression analysis was performed to estimate average daily traffic volume for roads in each town. Based on the analysis, linear and multiplicative traffic volume forecasting models were developed. The most important predictive variables were town population, employment levels, number of health clinics and number of shops. Both models were found to be reasonable, but the linear model is recommended because it is easy to understand, includes two explanatory variables that could easily be obtained or estimated and its form is consistent with previously developed prediction models.