Assignment of Seasonal Factor Categories to Urban Coverage Count Stations Using a Fuzzy Decision Tree

Abstract

Seasonal factor is an important parameter for converting coverage counts to annual average daily traffic (AADT). There have been many studies on establishing seasonal factor (SF) categories, but there is a limited understanding of how to assign SF groups to short-term count sites. In the current practice, established factor groups are typically assigned based on function class, the physical proximity of short count sites to a permanent traffic count site, and other engineering judgment. There is no sound theoretical basis or rules to guide the seasonal factor assignment process, which results in inaccurate AADT estimates for coverage count sites. This paper describes a data-driven procedure for assigning a seasonal factor category to a given portable count site by taking into consideration the similarities between the characteristics of permanent count sites in the seasonal factor group and the portable count site. A fuzzy decision tree was constructed based on the known SF groupings of permanent count sites and their land use attributes. The decision tree was then applied to determine the SF category for a given portable count site in Broward, Miami-Dade, and Palm Beach counties in Florida.