Motorcycle Helmet Use Behavior: What Does the Data Tell Us?
Analysis of helmet use behavior is a critical aspect of improving motorcycle safety. Machine learning has made significant progress in recent years, but its potential to analyze helmet use behavior is still largely unexploited. This study thus aimed to compare various machine learning approaches (support vector machine (SVM), random forest (RF), and gradient boosting decision tree (GBDT)) and a parametric statistical model, logistic regression (LR), to derive the accurate motorcycle helmet use behavior model using data from face-to-face interviews and online questionnaire surveys of 624 motorcycle riders in Dhaka, Bangladesh. It also analyzed the most important indicators of helmet use behavior and how they correlated to such behavior based on the predictions of the high-performance model. After reducing the dimensionality of the data with principal component analysis, this study developed and compared the predictive performance of the LR, SVM, RF, and GBDT. With an accuracy of 86%, an F1-score of 0.938, and an AUC value of 0.772, RF outperformed other models in the comparison. Using Gini-based and SHAP-based (SHapley Additive explanation) feature importance techniques, the most significant elements in explaining helmet usage behavior are trip duration, positive attitude toward helmet use, intended driving purpose, and temporal riding features and helmet cost. In addition, this study reported how helmet use behavior is influenced by these key factors.
- Record URL:
-
Supplemental Notes:
- This paper was sponsored by TRB committee ACS40 Standing Committee on Occupant Protection.
-
Corporate Authors:
Transportation Research Board
, -
Authors:
- Raihan, Md Asif
- Anik, B M Tazbiul Hassan
- Ashik, Fajle Rabbi
- Hasan, Md Mahmud
- Mahmud, S M Sohel
-
Conference:
- Transportation Research Board 102nd Annual Meeting
- Location: Washington DC, United States
- Date: 2023-1-8 to 2023-1-12
- Date: 2023
Language
- English
Media Info
- Media Type: Web
- Features: Figures; References; Tables;
- Pagination: 20p
Subject/Index Terms
- TRT Terms: Behavior; Helmet use; Machine learning; Motorcyclists; Surveys; Traffic safety
- Geographic Terms: Dhaka (Bangladesh)
- Subject Areas: Data and Information Technology; Highways; Safety and Human Factors;
Filing Info
- Accession Number: 01874165
- Record Type: Publication
- Report/Paper Numbers: TRBAM-23-03573
- Files: TRIS, TRB, ATRI
- Created Date: Feb 23 2023 9:21AM