ppm.object: Class of Fitted Point Process Models
In spatstat.core: Core Functionality of the 'spatstat' Family

ppm.object

R Documentation

Class of Fitted Point Process Models

Description

A class ppm to represent a fitted stochastic model for a point process. The output of ppm.

Details

An object of class ppm represents a stochastic point process model that has been fitted to a point pattern dataset. Typically it is the output of the model fitter, ppm.

The class ppm has methods for the following standard generic functions:

generic	method	description
`print`	`print.ppm`	print details
`plot`	`plot.ppm`	plot fitted model
`predict`	`predict.ppm`	fitted intensity and conditional intensity
`fitted`	`fitted.ppm`	fitted intensity
`coef`	`coef.ppm`	fitted coefficients of model
`anova`	`anova.ppm`	Analysis of Deviance
`formula`	`formula.ppm`	Extract model formula
`terms`	`terms.ppm`	Terms in the model formula
`labels`	`labels.ppm`	Names of estimable terms in the model formula
`residuals`	`residuals.ppm`	Point process residuals
`simulate`	`simulate.ppm`	Simulate the fitted model
`update`	`update.ppm`	Change or refit the model
`vcov`	`vcov.ppm`	Variance/covariance matrix of parameter estimates
`model.frame`	`model.frame.ppm`	Model frame
`model.matrix`	`model.matrix.ppm`	Design matrix
`logLik`	`logLik.ppm`	log pseudo likelihood
`extractAIC`	`extractAIC.ppm`	pseudolikelihood counterpart of AIC
`nobs`	`nobs.ppm`	number of observations

Objects of class ppm can also be handled by the following standard functions, without requiring a special method:

name	description
`confint`	Confidence intervals for parameters
`step`	Stepwise model selection
`drop1`	One-step model improvement
`add1`	One-step model improvement

The class ppm also has methods for the following generic functions defined in the spatstat package:

generic	method	description
`as.interact`	`as.interact.ppm`	Interpoint interaction structure
`as.owin`	`as.owin.ppm`	Observation window of data
`berman.test`	`berman.test.ppm`	Berman's test
`envelope`	`envelope.ppm`	Simulation envelopes
`fitin`	`fitin.ppm`	Fitted interaction
`is.marked`	`is.marked.ppm`	Determine whether the model is marked
`is.multitype`	`is.multitype.ppm`	Determine whether the model is multitype
`is.poisson`	`is.poisson.ppm`	Determine whether the model is Poisson
`is.stationary`	`is.stationary.ppm`	Determine whether the model is stationary
`cdf.test`	`cdf.test.ppm`	Spatial distribution test
`quadrat.test`	`quadrat.test.ppm`	Quadrat counting test
`reach`	`reach.ppm`	Interaction range of model
`rmhmodel`	`rmhmodel.ppm`	Model in a form that can be simulated
`rmh`	`rmh.ppm`	Perform simulation
`unitname`	`unitname.ppm`	Name of unit of length

Information about the data (to which the model was fitted) can be extracted using data.ppm, dummy.ppm and quad.ppm.

Internal format

If you really need to get at the internals, a ppm object contains at least the following entries:

`coef`	the fitted regular parameters (as returned by `glm`)
`trend`	the trend formula or `NULL`
`interaction`	the point process interaction family (an object of class `"interact"`) or `NULL`
`Q`	the quadrature scheme used
`maxlogpl`	the maximised value of log pseudolikelihood
`correction`	name of edge correction method used

See ppm for explanation of these concepts. The irregular parameters (e.g. the interaction radius of the Strauss process) are encoded in the interaction entry. However see the Warnings.

Warnings

The internal representation of ppm objects may change slightly between releases of the spatstat package.

Author(s)

\adrian

and \rolf

Examples

  fit <- ppm(cells ~ x, Strauss(0.1), correction="periodic")
  fit
  coef(fit)
  # pred <- predict(fit)
  pred <- predict(fit, ngrid=20, type="trend")
  if(interactive()) {
  plot(fit)
  }

spatstat.core documentation built on May 18, 2022, 9:05 a.m.