Spatial point process methods for linear networks with applications to road accident analysis

Greg McSwiggan

Research output: ThesisDoctoral Thesis

556 Downloads (Pure)

Abstract

Motivated by problems arising in the study of road traffic accidents, this thesis develops statistical methodology for analysing spatial patterns of points along a network of lines. Weaknesses of the traditional "crash-frequency" approach to traffic accident data are identified, and we develop a new approach using Poisson point process models. A new, statistically-principled, rapidly computable method of kernel density estimation on a linear network based on the heat kernel is developed. Techniques for estimating relative risk for spatial point patterns in two-dimensional space are extended to point patterns on a linear network and we develop improved methods for bandwidth selection.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • The University of Western Australia
Supervisors/Advisors
  • Nair, Gopalan, Supervisor
  • Baddeley, Adrian, Supervisor
Thesis sponsors
Award date25 May 2020
DOIs
Publication statusUnpublished - 2019

Fingerprint

Dive into the research topics of 'Spatial point process methods for linear networks with applications to road accident analysis'. Together they form a unique fingerprint.

Cite this