Hotspot Detection on a Linear Network in the Presence of Covariates: A Case Study on Road Crash Data

24 Pages Posted: 15 Nov 2023

See all articles by Tomáš Mrkvička

Tomáš Mrkvička

University of South Bohemia

Stanislav Kraft

University of South Bohemia

Vojtěch Blažek

University of South Bohemia

Mari Myllymäki

Natural Resources Institute Finland

Multiple version iconThere are 2 versions of this paper

Abstract

A novel method for detecting hotspots of point events on a linear network is introduced. Traditional hotspot detection methods can be spoiled by the dependence of point intensity on covariates. This issue is addressed by the proposed method. It is based on fitting an inhomogeneous point pattern model and simulating from the fitted model. The estimate of the nonparametric intensity function is then used in a Monte Carlo test, where the multiple hypothesis testing problem is resolved using a False Discovery Rate envelope. The method is illustrated using road crash data from South Bohemia, Czech Republic. Although this method is computationally intensive due to its reliance on simulations, the chosen data study demonstrates its feasibility for large data sets. We developed a brand new method for identifying hotspots from the perspective of three different types of road drivers (risky drivers, ordinary drivers, and
1
cautious drivers). These three models is the main scientific added value of the paper. They can be used for assessing the road drivers’ behaviour, for mitigation the riskiness of particular road segments, or for traffic education.

Keywords: False Discovery Rate envelope, Inhomogeneous cluster point process, Monte Carlo method, Simulations of point pattern, South Bohemia

Suggested Citation

Mrkvička, Tomáš and Kraft, Stanislav and Blažek, Vojtěch and Myllymäki, Mari, Hotspot Detection on a Linear Network in the Presence of Covariates: A Case Study on Road Crash Data. Available at SSRN: https://ssrn.com/abstract=4627591 or http://dx.doi.org/10.2139/ssrn.4627591

Tomáš Mrkvička (Contact Author)

University of South Bohemia ( email )

Stanislav Kraft

University of South Bohemia ( email )

Vojtěch Blažek

University of South Bohemia ( email )

Mari Myllymäki

Natural Resources Institute Finland ( email )

P.O. Box 18 (Jokiniemenkuja 1)
Vantaa, FI-01301
Finland

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
22
Abstract Views
92
PlumX Metrics