Analyzing Factors Associated with Fatal Road Crashes: A Machine Learning Approach

Area of Science:

• Public Health
• Transportation Safety
• Data Science

Background:

• Road traffic injuries impose a significant global human and economic burden.
• Identifying risk factors for fatal road injuries is essential for developing effective prevention strategies.

Purpose of the Study:

• To propose and validate a hybrid ensemble machine learning model to identify risk factors for fatal road injuries.
• To analyze the Lebanese Road Accidents Platform (LRAP) database to understand factors contributing to road crash fatalities.

Main Methods:

• A hybrid ensemble machine learning classifier combining sequential minimal optimization and decision trees was developed.
• The model was trained, tested, and validated using a dataset of 8482 road crash incidents from the LRAP database.
• Sensitivity analysis was performed to assess the influence of various factors on fatality occurrence.

Main Results:

• Seven out of nine independent variables were significantly associated with fatality occurrence.
• Key risk factors identified include crash type, injury severity, spatial cluster ID, and crash time (hour).

Conclusions:

• The developed machine learning model effectively identifies significant risk factors for fatal road injuries.
• Findings provide evidence-based insights for policymakers to enhance road safety programs and policies.
• Understanding these factors is vital for reducing the human and economic toll of road traffic injuries.