Incident duration prediction through integration of uncertainty and risk factor evaluation: A San Francisco incidents case study
2025

Predicting Traffic Incident Duration in San Francisco

Sample size: 16986 publication 20 minutes Evidence: high

Author Information

Author(s): Salehi Amirreza, Babaei Ardavan, Khedmati Majid

Primary Institution: Sharif University of Technology, Tehran, Iran

Hypothesis

Can integrating uncertainty and risk factor evaluation improve the prediction of traffic incident durations?

Conclusion

The study presents a framework that significantly enhances the prediction of traffic incident durations by incorporating street risk factors and fuzzy clustering methods.

Supporting Evidence

  • Traffic congestion is worsened by incidents, which can be predicted more accurately using the proposed framework.
  • The study identifies 'Risk' as a critical factor influencing incident duration.
  • Fuzzy clustering methods allow for better categorization of incidents, especially those that do not fit neatly into defined clusters.
  • Machine learning models demonstrated high accuracy in predicting incident durations.

Takeaway

This study helps predict how long traffic incidents will last by looking at things like street conditions and weather, making it easier for people to plan their routes.

Methodology

The study used a dataset from San Francisco, applying machine learning and fuzzy clustering techniques to predict incident durations based on various risk factors.

Potential Biases

Potential biases may arise from missing data and inaccuracies in incident reporting.

Limitations

The dataset may lack comprehensive information about specific incidents, and variations in driving regulations can affect generalizability.

Digital Object Identifier (DOI)

10.1371/journal.pone.0316289

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication