lgcp_graph: Create a log-Gaussian Cox process model for metric graphs

In MetricGraph: Random Fields on Metric Graphs

Create a log-Gaussian Cox process model for metric graphs

Description

This function creates a log-Gaussian Cox process model for point pattern data on metric graphs. It handles the creation of integration points and prepares the data for fitting with INLA.

Usage

lgcp_graph(
  formula,
  graph,
  interpolate = TRUE,
  manual_integration_points = NULL,
  manual_covariates = NULL,
  use_current_mesh = TRUE,
  new_h = NULL,
  new_n = NULL,
  repl = ".all",
  repl_col = ".group",
  ...
)

Arguments

formula	A formula object specifying the model structure
graph	A metric_graph object containing the network and point pattern data
interpolate	Logical; if TRUE, interpolate covariates from the graph data to integration points
manual_integration_points	Data frame with columns edge_number, distance_on_edge, and E (integration weights) for manually specified integration points, or NULL to use automatic integration points
manual_covariates	Named vector of covariates at integration points if interpolate is FALSE and covariates are used
use_current_mesh	Logical; if TRUE, use the existing mesh in the graph as integration points
new_h	Numeric; mesh size for creating a new mesh if use_current_mesh is FALSE
new_n	Integer; alternative to new_h, specifies the approximate number of mesh points
repl	Vector of replicates to be used in the model. For all replicates, one must use ".all".
repl_col	Name of the column in the data that contains the replicates. Default is ".group".
...	Additional arguments to be passed to inla

Value

An object containing the fitted LGCP model

MetricGraph documentation built on June 8, 2025, 1:55 p.m.