Generalized criteria for evaluating hotspot identification methods.

Area of Science:

• Transportation Engineering
• Traffic Safety
• Data Analysis

Background:

• Hotspot identification (HSID) is crucial for highway safety management.
• Existing HSID methods lack consistency and temporal stability in identifying hazardous sites.
• Current evaluation criteria are limited and may not accurately assess HSID performance.

Purpose of the Study:

• To propose and validate generalized criteria for evaluating the performance of HSID methods.
• To address the limitations of existing evaluation criteria, particularly regarding temporal instability.
• To enhance the reliability and consistency of HSID method assessments.

Main Methods:

• Development of three generalized criteria: High Crashes Consistency Test (HCCT), Common Sites Consistency Test (CSCT), and Absolute Rank Differences Test (ARDT).
• Application of these criteria to evaluate three common HSID methods using eight years of crash data from Texas rural two-lane roadways.
• Comparison of the proposed generalized criteria against existing evaluation methods.

Main Results:

• The generalized criteria effectively evaluate HSID methods over multiple time periods.
• The proposed criteria provide more consistent results with less score discrepancy compared to existing methods.
• HCCT, CSCT, and ARDT demonstrate reliability in assessing HSID method performance.

Conclusions:

• The generalized criteria offer a more robust and comprehensive approach to evaluating HSID methods.
• These new criteria improve the assessment of temporal instability and consistency in hotspot identification.
• The study validates the superiority of the proposed generalized criteria for highway safety management.