The prediction of the incident duration is an important issue to the advanced traffic incident management. An accurate prediction of incident duration contributes significantly to making appropriate decisions to address incidents for traffic managers and obtaining traffic information for travelers in a timely manner. This paper analyzes several regression methods, such as the principal component regression (PCR), partial least square regression (PLSR), and builds models between incident duration and its influence factors. The models were established for three types of incident, correspondingly, i.e. stopped vehicle, lost load, and accident. The models received the best prediction accuracy of 83.6%, 92.7%, and 88.2% for stopped vehicle, lost-load and accident at error for 20 minutes, respectively. The results indicated that the regression method is a promising application to predict incident duration and the PCR and the PLSR have a better performance.
Analysis of Regression Method on Traffic Incident Duration Prediction
Second International Conference on Transportation Information and Safety ; 2013 ; Wuhan, China
ICTIS 2013 ; 1008-1015
2013-06-11
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Modeling Traffic Incident Duration Using Quantile Regression
Transportation Research Record | 2016
|Big data-driven highway traffic incident duration prediction method
Europäisches Patentamt | 2022
|Application of nonparametric regression in predicting traffic incident duration
DOAJ | 2018
|Text analysis in incident duration prediction
Elsevier | 2013
|Competing risks mixture model for traffic incident duration prediction
Online Contents | 2015
|