Analyzing heterogeneity in motorcycle crashes: a comparative study of senior and young riders using the random forest approach

This study provides a comparative analysis of highway road accidents involving senior and young motorcycle riders in Thailand, utilizing a random forest approach to uncover pivotal factors contributing to accidents within these age groups. The dataset, comprising 33,875 highway accident cases recorded between 2015 and 2020, was sourced from official government agency (Thailand Department of Public Disaster Prevention and Mitigation) reports. Accidents were categorized based on multiple variables, including weather conditions, road infrastructure, and human behaviors. The findings reveal that senior motorcycle riders are significantly more likely to experience fatal outcomes compared to their younger counterparts. The random forest algorithm demonstrated strong predictive capabilities, achieving accuracy rates of 67.5 % (AUC: 0.721) for senior riders and 73 % (AUC: 0.745) for young riders. Key contributing factors to accidents differed notably between the two groups: while human factors such as intoxicated riding and riding during daylight with proper lighting were predominant among young riders, environmental factors, including road and weather conditions, played a more critical role in accidents involving senior riders. This study highlights the effectiveness of the random forest algorithm in predicting accident outcomes and identifying distinct risk factors for different age groups. By uncovering these differences, the research provides valuable insights into the underlying causes of highway accidents involving senior and young motorcycle riders. The results underscore the need for tailored interventions and policies to mitigate risks for these vulnerable populations, thereby enhancing road safety outcomes.