CONTACT US: Contact info
|
Citation |
Mohamad I, Jomnonkwao S, Ratanavaraha V. Case Stud. Transp. Policy 2022; 10(4): 2165-2174. |
|
Copyright |
(Copyright © 2022, World Conference on Transport Research Society, Publisher Elsevier Publishing) |
|
DOI |
|
PMID |
unavailable |
|
Abstract |
Thailand ranks first in Asia and ninth in the world in term of road accident. As of 2020, the number of vehicles registered in Thailand was over 41 million, with motorcycles accounting for half of all vehicles. This study aimed to determine the cause of fatalities to reduce motorcycle accidents. The research entailed separating the accidents and fatalities into those occurring on highways (HWs) versus those occurring on rural roadways (RRs) and focused solely on rider at fault accidents to involve any confounding factors related to passengers or others involved. In Thailand, HWs have higher speed limits and allow more vehicle types than some RRs. Thailand's Department of Public Disaster Prevention and Mitigation recorded 115,154 motorcycle accidents from 2015 to 2020. Decision trees allow for processing large amounts of data to drill down into associations between the individual variables in a large data set; in this study, the tree also separated accidents into whether or not the driver was exceeding the speed limit. The model's performance for HWs, predicted misclassifications were found to be 28.3% (fatality to nonfatality) and a 44.5% (nonfatality to fatality) while predicted misclassification for RRs were 15.5% (fatality to nonfatality) and 60% (nonfatality to fatality). At all ages, the most fatalities were among male riders on dry straightaways in clear daytime weather; notably, however, on RRs, even when the rider was driving responsibly, fatalities were high at night on roads with no light. Following the presentation of the study findings, suggestions are made for ways the Thai government can improve the motorcycle accident and fatality statistics, including increasing the age limit for a motorcycle license, with engine size limits further divided according to age; proper enforcement of the existing rules will also improve the country's accident statistics. It will also be highly effective to improve road lighting, particularly on RRs. Language: en |
|
Keywords |
Accidents; Machine learning; Transportation; Big data; Comparative analysis; Decision tree |