Road safety measurement with reliability using an advanced hybrid decision model

Area of Science:

• Decision Sciences
• Public Health
• Data Science

Background:

• Effective road safety analysis requires robust decision-making frameworks.
• Existing methods may lack robustness and global optimality in identifying performance patterns.
• High-dimensional data in road safety presents challenges for traditional analysis.

Purpose of the Study:

• To introduce and validate a novel hybrid multi-criteria decision-making (MCDM) framework, the HDVP-BeVarMax model.
• To apply this model for measuring national road safety performance using comprehensive indicators.
• To provide a reliable decision-support tool for policy development and resource allocation in road safety.

Main Methods:

• Developed a hybrid MCDM framework combining High-Dimensional Vector Projection (HDVP) and Between-class Variance Maximization (BeVarMax).
• Applied the HDVP-BeVarMax model to longitudinal road safety data from 13 East Asia Summit (EAS) countries (2012-2023) using 15 safety performance indicators (SPIs).
• Conducted sensitivity, benchmarking, and validation analyses to assess model reliability and robustness.

Main Results:

• The HDVP-BeVarMax model demonstrated superior reliability, robustness, and discriminative power compared to conventional methods.
• The model successfully quantified national road safety performance and identified latent structures among countries.
• Analyses confirmed the model's effectiveness across various normalization and weighting schemes.

Conclusions:

• The HDVP-BeVarMax model offers a trustworthy approach for decision-making in road safety.
• It enables effective benchmarking of national performance and informs targeted interventions and resource allocation.
• The model supports regional cooperation and progress towards global road safety goals.