Structured network regression for spatial point patterns

TL;DR

This paper introduces a novel regression model linking network-based spatial point pattern intensities to covariates and graph statistics, enabling analysis of complex dependencies in urban spatial data.

Contribution

It presents the first regression approach specifically designed for network intensity functions of spatial point patterns, bridging a gap in spatial network analysis.

Findings

Model effectively captures dependencies of network intensity on covariates.

Application to urban disturbance data demonstrates practical utility.

Provides a new tool for spatial network data analysis.

Abstract

The analysis of spatial point patterns that occur in the network domain have recently gained much attraction and various intensity functions and measures have been proposed. However, the linkage of spatial network statistics to regression models has not been approached so far. This paper presents a new regression approach which treats a generic intensity function of a planar point pattern that occurred on a network as the outcome of a set of different covariates and various graph statistics. Different to all alternative approaches, our model is the first which permits the statistical analysis of complex regression data in the context of network intensity functions for spatial point patterns. The potential of our new technique to model the structural dependencies of network intensity functions on various covariates and graph statistics is illustrated using call-in data on neighbour and community disturbances in an urban context.