Open AccessPredicting Severity of Accidents in Malaysia By Ordinal Logistic Regression Models
Author(s) -
Uneb Gazder,
Ashar Ahmed,
Umaira Shahid
Publication year - 2021
Publication title -
international journal of traffic and transportation management
Language(s) - English
Resource type - Journals
ISSN - 2371-5782
DOI - 10.5383/jttm.03.01.002
Subject(s) - logistic regression , ordered logit , ordinal regression , road accident , accident (philosophy) , regression analysis , transport engineering , geography , statistics , environmental health , medicine , mathematics , engineering , philosophy , epistemology
This study was aimed at determining the relationships of accident severity using road environment and traveller characteristics. Ordinal logistic regression models were used in this study. The accident data was provided by Malaysian Research Institute of Road Safety (MIROS) for all accidents which occurred in Penang state during 2006-2011. It was observed that motorbikes were predominantly involved in these accidents, hence, it was decided to develop three separate models; one for the overall data, and others for accidents with and without motorbikes. Logistic regression models showed that commercial land use, road width and experience of driver are important factors that may increase severity of accidents. Shoulder width was found to decrease the severity of motorbike accidents. Commercial land use, road width and driver experience have more impact on motorbike accidents as compared to accidents of other vehicles.