Department of Transportation Engineering, The University of Seoul, 163 Seoulsiripdae-ro Dongdaemun-gu, Seoul 02504, South Korea.
Texas A&M Transportation Institute, College Station, TX 77843, USA.
Accid Anal Prev. 2022 Feb;165:106517. doi: 10.1016/j.aap.2021.106517. Epub 2021 Dec 9.
Despite the countless benefits derived from motorcycle usage, it has become a significant public health concern, particularly in developing countries, due to the plateauing number of fatal/serious injuries associated with them. Although it has been well documented that the frequency and fatality rates of intersection-related motorcycle crashes are high, little research efforts have been made to explore the contributory factors influencing motorcycle-involved crashes at these locations. Interestingly, no study has investigated the latent patterns and chains of factors that simultaneously contribute to the injury severity sustained by motorcycle crash casualties at intersections under different traffic control conditions in developing countries. Since motorcycles are mostly used as taxis in developing countries, it is imperative to consider the injury severity sustained by all crash casualties in the motorcycle safety analysis. This study bridges the research gap by employing a plausible data mining tool to explore hidden rules associated with motorcycle crash casualty injury severity outcomes at both signalized and non-signalized intersections in Ghana's most densely populated region, Accra, using three-year crash data spanning 2016-2018. Besides, a binary logit regression model was also employed to explore the impact of crash factors on casualty severity outcomes using the same dataset. The results from both analysis techniques were consistent; however, the data mining technique provided chains of factors which provided additional insights into the groups of factors that collectively influence the casualty injury severity outcomes. From the rule discovery results, while full license status, daytime/daylight, and shoulder presence increased the risk of fatal injuries at signalized intersections, factors such as inattentiveness, good road surface, nighttime, shoulder absence, and young rider were highly likely to increase casualty fatalities at non-signalized intersections. By controlling all or some of these risk factors, the level of injury severity on the roadways could be reduced. Based on the findings, we provide enforcement, education, and engineering-based recommendations to help improve motorcycle safety.
尽管使用摩托车有无数的好处，但由于与摩托车相关的致命/重伤数量趋于平稳，它已成为一个重大的公共卫生问题，尤其是在发展中国家。虽然有大量文献记录表明交叉口相关的摩托车碰撞事故的频率和死亡率很高，但很少有研究致力于探讨影响这些地点涉及摩托车碰撞的因素。有趣的是，没有研究调查过在发展中国家不同交通控制条件下，同时影响交叉口摩托车碰撞伤亡严重程度的潜在模式和因素链。由于摩托车在发展中国家大多用作出租车，因此在进行摩托车安全分析时，必须考虑所有碰撞伤亡人员所承受的伤害严重程度。本研究通过使用合理的数据挖掘工具，来填补这一研究空白，以探索加纳人口最密集的地区阿克拉的信号交叉口和无信号交叉口的摩托车碰撞伤亡人员受伤严重程度的隐藏规则，研究使用了 2016 年至 2018 年为期三年的碰撞数据。此外，还使用相同的数据集，采用二元逻辑回归模型来探讨碰撞因素对伤亡人员严重程度的影响。两种分析技术的结果都是一致的；然而，数据挖掘技术提供了因素链，为共同影响伤亡人员受伤严重程度的因素组提供了更多的见解。从规则发现结果来看，虽然完全的驾照状态、白天/日光和路肩存在增加了信号交叉口致命伤害的风险，但在无信号交叉口，诸如注意力不集中、良好的路面、夜间、无路肩和年轻骑手等因素更有可能增加伤亡人员的死亡风险。通过控制所有或部分这些危险因素，可以降低道路上的伤害严重程度。基于这些发现，我们提供了执法、教育和工程方面的建议，以帮助提高摩托车安全性。
