Hot-Spot Identification

Ferreira, Sara; Couto, António · 2013 · OpenAlex-citations

DOI: 10.3141/2386-01

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the challenge of identifying traffic safety hot spots, a critical process for reducing future accidents. The authors critique existing methodologies, particularly those outlined in the Highway Safety Manual, which often rely on Empirical Bayesian (EB) or Full Bayesian (FB) approaches. While Bayesian methods account for regression-to-the-mean bias, they suffer from significant limitations: EB methods are criticized for implicitly using data twice to estimate prior parameters, while FB methods require challenging and controversial specifications of prior beliefs. Consequently, these methods can be complex and sometimes infeasible for practical safety decision-making. To overcome these issues, Ferreira and Couto propose an alternative probabilistic methodology based on a categorical binary model. This approach aims to provide a simpler, more intuitive framework that avoids the need for large sample sizes to develop safety performance functions and eliminates the arbitrary distinction between safe and unsafe sites. The proposed method operates in two main steps. First, a threshold for the number of accidents is established to classify sites into two categories: hot spots (1) or safe sites (0). Second, a binary choice model is applied to link this outcome to specific risk factors characterizing the sites. For urban intersections, these factors include traffic volume on minor and major approaches, the number of legs (three or four), and the type of signalization (signalized or unsignalized). The model estimates the probability of a site belonging to either category based on these inputs. The ranking criterion for identification is the probability of a site being a hot spot. This allows for the construction of a site list ordered by this probability, enabling selection strategies that target either a fixed number of highest-probability sites or all sites exceeding a specific probability threshold, such as 0.5. To validate the methodology, the authors applied it to simulated urban intersection data from Porto, Portugal, covering a five-year period. The results demonstrated that the binary model provided a good fit to the data. The study evaluated the performance of this probabilistic method against commonly used techniques by testing their power to detect true hot spots. The findings indicated the superiority of the proposed binary model in this regard. The authors conclude that this method is simple to apply and effectively overcomes critical issues associated with traditional approaches, such as the assumptions required for prior distributions in Bayesian analysis and the regression-to-the-mean phenomenon. By providing a realistic and intuitive perspective on site risk, the proposed framework offers a feasible and effective tool for supporting traffic safety decisions and network screening processes.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• induced exposure