Maximum likelihood method of estimating the conflict-crash relationship.

Area of Science:

• Traffic safety engineering
• Transportation research
• Statistical modeling

Background:

• Estimating the relationship between traffic conflicts and crashes is crucial for road safety.
• Existing methods face challenges in accurately converting observed traffic events into expected crash numbers.
• Practical application for safety engineers necessitates reliable and efficient estimation techniques.

Purpose of the Study:

• To compare Ordinary Least Squares (OLS) and Maximum Likelihood (ML) methods for estimating Lomax distribution parameters in traffic safety.
• To evaluate the impact of assuming the scale parameter on crash predictions using traffic conflict data.
• To assess the practicality of these methods for safety engineers in road safety assessments.

Main Methods:

• Recalls the OLS method for estimating the Lomax distribution's shape parameter.
• Introduces and applies the ML method for Lomax-based crash estimations.
• Compares OLS and ML estimates, analyzing the effect of assuming the scale parameter.

Main Results:

• Both OLS and ML methods, when assuming the scale parameter, provide Lomax-based crash estimates.
• The effect of assuming the scale parameter is compared against driver type and limited observations.
• A re-parametrized Lomax distribution is proposed to address simultaneous parameter estimation challenges.

Conclusions:

• The study provides a comparative analysis of OLS and ML methods for traffic safety analysis using conflict data.
• Understanding the impact of parameter assumptions is vital for accurate crash risk estimation.
• The proposed re-parametrization offers a potential solution for improving the estimation of Lomax distribution parameters in traffic safety applications.