like {inlabru}

Observation model construction for usage with bru()

Description

Observation model construction for usage with bru()

Usage

like(
  formula = . ~ .,
  family = "gaussian",
  data = NULL,
  response_data = NULL,
  mesh = deprecated(),
  E = NULL,
  Ntrials = NULL,
  weights = NULL,
  scale = NULL,
  samplers = NULL,
  ips = NULL,
  domain = NULL,
  include = NULL,
  exclude = NULL,
  include_latent = NULL,
  used = NULL,
  allow_latent = deprecated(),
  allow_combine = NULL,
  control.family = NULL,
  options = list(),
  .envir = parent.frame()
)

like_list(...)

## S3 method for class 'list'
like_list(object, envir = NULL, ...)

## S3 method for class 'bru_like'
like_list(..., envir = NULL)

## S3 method for class 'bru_like'
c(..., envir = NULL)

## S3 method for class 'bru_like_list'
c(..., envir = NULL)

## S3 method for class 'bru_like_list'
x[i]

Arguments

formula

a formula where the right hand side is a general R expression defines the predictor used in the model.

family

A string identifying a valid INLA::inla likelihood family. The default is gaussian with identity link. In addition to the likelihoods provided by inla (see names(INLA::inla.models()$likelihood)) inlabru supports fitting latent Gaussian Cox processes via family = "cp". As an alternative to bru(), the lgcp() function provides a convenient interface to fitting Cox processes.

data

Likelihood-specific data, as a data.frame or SpatialPoints[DataFrame] object.

response_data

Likelihood-specific data for models that need different size/format for inputs and response variables, as a data.frame or SpatialPoints[DataFrame] object.

mesh

Deprecated.

E

Exposure parameter for family = 'poisson' passed on to INLA::inla. Special case if family is 'cp': rescale all integration weights by a scalar E. For sampler specific reweighting/effort, use a weight column in the samplers object, see fmesher::fm_int(). Default taken from options$E, normally 1.

Ntrials

A vector containing the number of trials for the 'binomial' likelihood. Default taken from options$Ntrials, normally 1.

weights

Fixed (optional) weights parameters of the likelihood, so the log-likelihood⁠[i]⁠ is changed into weights[i] * log_likelihood[i]. Default value is 1. WARNING: The normalizing constant for the likelihood is NOT recomputed, so ALL marginals (and the marginal likelihood) must be interpreted with great care.

scale

Fixed (optional) scale parameters of the precision for several models, such as Gaussian and student-t response models.

samplers

Integration domain for 'cp' family.

ips

Integration points for 'cp' family. Overrides samplers.

domain

Named list of domain definitions.

include

Character vector of component labels that are used as effects by the predictor expression; Default: the result of ⁠[all.vars()]⁠ on the predictor expression, unless the expression is not ".", in which case include=NULL, to include all components that are not explicitly excluded. The bru_used() methods are used to extract the variable names, followed by removal of non-component names when the components are available.

exclude

Character vector of component labels that are not used by the predictor expression. The exclusion list is applied to the list as determined by the include parameter; Default: NULL (do not remove any components from the inclusion list)

include_latent

character vector. Specifies which the latent state variables are directly available to the predictor expression, with a ⁠_latent⁠ suffix. This also makes evaluator functions with suffix ⁠_eval⁠ available, taking parameters main, group, and replicate, taking values for where to evaluate the component effect that are different than those defined in the component definition itself (see component_eval()). Default NULL auto-detects use of ⁠_latent⁠ and ⁠_eval⁠ in the predictor expression.

used

Either NULL or a bru_used() object, overriding include, exclude, and include_latent.

allow_latent

logical, deprecated. Use include_latent instead.

allow_combine

logical; If TRUE, the predictor expression may involve several rows of the input data to influence the same row. Default FALSE, but forced to TRUE if response_data is non-NULL, data is a list, or the likelihood construction requires it.

control.family

A optional list of INLA::control.family options

options

A bru_options options object or a list of options passed on to bru_options()

.envir

The evaluation environment to use for special arguments (E, Ntrials, weights, and scale) if not found in response_data or data. Defaults to the calling environment.

...

For like_list.bru_like, one or more bru_like objects

object

A list of bru_like objects

envir

An optional environment for the new bru_like_list object

x

bru_like_list object from which to extract element(s)

i

indices specifying elements to extract

Value

A likelihood configuration which can be used to parameterise bru().

Functions

• like_list(): Combine bru_like likelihoods into a bru_like_list object

• like_list(list): Combine a list of bru_like likelihoods into a bru_like_list object

• like_list(bru_like): Combine several bru_like likelihoods into a bru_like_list object

• c(bru_like): Combine several bru_like likelihoods and/or bru_like_list objects into a bru_like_list object

• c(bru_like_list): Combine several bru_like likelihoods and/or bru_like_list objects into a bru_like_list object

Author(s)

Fabian E. Bachl [e-mail]

Finn Lindgren [e-mail]

Examples

if (bru_safe_inla() &&
    require(ggplot2, quietly = TRUE)) {

  # The like function's main purpose is to set up models with multiple likelihoods.
  # The following example generates some random covariates which are observed through
  # two different random effect models with different likelihoods

  # Generate the data

  set.seed(123)

  n1 <- 200
  n2 <- 10

  x1 <- runif(n1)
  x2 <- runif(n2)
  z2 <- runif(n2)

  y1 <- rnorm(n1, mean = 2 * x1 + 3)
  y2 <- rpois(n2, lambda = exp(2 * x2 + z2 + 3))

  df1 <- data.frame(y = y1, x = x1)
  df2 <- data.frame(y = y2, x = x2, z = z2)

  # Single likelihood models and inference using bru are done via

  cmp1 <- y ~ -1 + Intercept(1) + x
  fit1 <- bru(cmp1, family = "gaussian", data = df1)
  summary(fit1)

  cmp2 <- y ~ -1 + Intercept(1) + x + z
  fit2 <- bru(cmp2, family = "poisson", data = df2)
  summary(fit2)

  # A joint model has two likelihoods, which are set up using the like function

  lik1 <- like("gaussian", formula = y ~ x + Intercept, data = df1)
  lik2 <- like("poisson", formula = y ~ x + z + Intercept, data = df2)

  # The union of effects of both models gives the components needed to run bru

  jcmp <- ~ x + z + Intercept(1)
  jfit <- bru(jcmp, lik1, lik2)

  # Compare the estimates

  p1 <- ggplot() +
    gg(fit1$summary.fixed, bar = TRUE) +
    ylim(0, 4) +
    ggtitle("Model 1")
  p2 <- ggplot() +
    gg(fit2$summary.fixed, bar = TRUE) +
    ylim(0, 4) +
    ggtitle("Model 2")
  pj <- ggplot() +
    gg(jfit$summary.fixed, bar = TRUE) +
    ylim(0, 4) +
    ggtitle("Joint model")

  multiplot(p1, p2, pj)
}