Abstract This paper presents a methodology for estimating the incident duration and identifying its critical contributing factors in the state of Maryland. The incident database from the Maryland State Highway (MDSHA) and Police Accident Report database between years 2003 and 2005 were used for model development. This study employed a hybrid model to develop the primary estimation system, consisting of a Rule-Based Tree Model (RBTM), Multinomial Logit Model (MNL), and Naïve Bayesian Classifier (NBC). Through the extensive data analysis and model estimation, we have identified some critical relationships between the set of key factors and the resulting incident duration. The proposed model along with research findings can play a vital role for traffic agencies to establish an advanced traveler information system, and to provide the incident-induced delay to both pre-trip and the en-route drivers.
Development of a Hybrid Prediction Model for Freeway Incident Duration: A Case Study in Maryland
International Journal of Intelligent Transportation Systems Research ; 10 , 1 ; 22-33
2011-09-20
12 pages
Article (Journal)
Electronic Resource
English
Incident duration , Tree model , Logit model , Naïve Bayesian Classifier , Advanced traveler information Engineering , Civil Engineering , Computer Imaging, Vision, Pattern Recognition and Graphics , User Interfaces and Human Computer Interaction , Automotive Engineering , Electrical Engineering , Robotics and Automation
Predicting Freeway Incident Duration Using Machine Learning
Springer Verlag | 2020
|Hazard based models for freeway traffic incident duration
Online Contents | 2013
|