vglmer_predict {vglmer}

Predict after vglmer

Description

These functions calculate the estimated linear predictor using the variational distributions. predict.vglmer draws predictions using the estimated variational distributions; predict_MAVB does so using the MAVB procedure described in Goplerud (2022a).

Usage

## S3 method for class 'vglmer'
predict(
  object,
  newdata,
  samples = 0,
  samples_only = FALSE,
  summary = TRUE,
  allow_missing_levels = FALSE,
  ...
)

predict_MAVB(
  object,
  newdata,
  samples = 0,
  samples_only = FALSE,
  var_px = Inf,
  summary = TRUE,
  allow_missing_levels = FALSE
)

Arguments

object

Model fit using vglmer.

newdata

Dataset to use for predictions. It cannot be missing.

samples

Number of samples to draw. Using 0 (default) gives the expectation of the linear predictor. A positive integer draws samples samples from the variational distributions and calculates the linear predictor.

samples_only

Default (FALSE) returns the samples from the variational distributions, not the prediction. Each row is a sample and each column is a parameter.

summary

Default (TRUE) returns the mean and variance of the samples for each observation. FALSE returns a matrix of the sampled linear predictor for each observation. Each row is a sample and each column is an observation.

allow_missing_levels

Default (FALSE) does not allow prediction for levels not observed in the original data. TRUE allows for prediction on unseen levels; the value of 0 (with no uncertainty) is used for the corresponding random effect.

...

Not used; included to maintain compatibility with existing methods.

var_px

Variance of working prior for marginal augmentation. Default (Inf) is a flat, improper, prior.

Value

This function returns an estimate of the linear predictor. The default returns the expected mean, i.e. E_{q(\alpha,\beta)}[x_i^T \beta + z_i^T\alpha]. If samples > 0, these functions return a summary of the prediction for each observation, i.e. the estimated mean and variance. If summary = FALSE, the sampled values of the linear predictor are returned as a matrix. predict_MAVB performs MAVB as described in Goplerud (2022a) before returning the linear predictor.

If allow_missing_levels = TRUE, then observations with a new (unseen) level for the random effect are given a value of zero for that term of the prediction.

Examples

set.seed(123)
sim_data <- data.frame(
  x = rnorm(100),
  y = rbinom(100, 1, 0.5),
  g = sample(letters, 100, replace = TRUE)
)

# Run with defaults
est_vglmer <- vglmer(y ~ x + (x | g), data = sim_data, family = "binomial")

# Simple prediction
predict(est_vglmer, newdata = sim_data)
# Return 10 posterior draws of the linear predictor for each observation.
predict_MAVB(est_vglmer, newdata = sim_data, summary = FALSE, samples = 10)
# Predict with a new level; note this would fail if 
# allow_missing_levels = FALSE (the default)
predict(est_vglmer,
  newdata = data.frame(g = "AB", x = 0),
  allow_missing_levels = TRUE
)