“…-Weather characteristics: sunny, rainy, and snowy -Temporal characteristics: AM peak, PM peak, night, and weekday -Incident characteristics: lanes, property, severity, debris, road repair, and pothole -Involved vehicle characteristics: bus, van, and truck During the past few years, a variety of methods have been applied to develop freeway accident duration estimating/forecasting models. The most representative approaches can be classified into the following categories: multivariate regression (Garib et al, 1997;Smith K. & Smith B., 2001;Valenti et al, 2010), fuzzy logic model (Choi, 1996;Dimitriou & Vlahogianni, 2015), artificial neural network (Wang et al, 2005), and survival (Chung et (Spławińska, 2015;Pamula, 2012). Thus, kNN and ANN were chosen as the key analytical techniques in this study.…”