mmrm_tmb_methods {mmrm}

Methods for mmrm_tmb Objects

Description

Usage

## S3 method for class 'mmrm_tmb'
coef(object, complete = TRUE, ...)

## S3 method for class 'mmrm_tmb'
fitted(object, ...)

## S3 method for class 'mmrm_tmb'
predict(
  object,
  newdata,
  se.fit = FALSE,
  interval = c("none", "confidence", "prediction"),
  level = 0.95,
  nsim = 1000L,
  conditional = TRUE,
  ...
)

## S3 method for class 'mmrm_tmb'
model.frame(
  formula,
  data,
  include = c("subject_var", "visit_var", "group_var", "response_var"),
  full,
  na.action = "na.omit",
  ...
)

## S3 method for class 'mmrm_tmb'
model.matrix(object, data, include = NULL, ...)

## S3 method for class 'mmrm_tmb'
terms(x, include = "response_var", ...)

## S3 method for class 'mmrm_tmb'
logLik(object, ...)

## S3 method for class 'mmrm_tmb'
formula(x, ...)

## S3 method for class 'mmrm_tmb'
vcov(object, complete = TRUE, ...)

## S3 method for class 'mmrm_tmb'
VarCorr(x, sigma = NA, ...)

## S3 method for class 'mmrm_tmb'
deviance(object, ...)

## S3 method for class 'mmrm_tmb'
AIC(object, corrected = FALSE, ..., k = 2)

## S3 method for class 'mmrm_tmb'
BIC(object, ...)

## S3 method for class 'mmrm_tmb'
print(x, ...)

## S3 method for class 'mmrm_tmb'
residuals(object, type = c("response", "pearson", "normalized"), ...)

## S3 method for class 'mmrm_tmb'
simulate(
  object,
  nsim = 1,
  seed = NULL,
  newdata,
  ...,
  method = c("conditional", "marginal")
)

Arguments

object

(mmrm_tmb)
the fitted MMRM object.

complete

(flag)
whether to include potential non-estimable coefficients.

...

mostly not used; Exception is model.matrix() passing ... to the default method.

newdata

(data.frame)
optional new data, otherwise data from object is used.

se.fit

(flag)
indicator if standard errors are required.

interval

(string)
type of interval calculation. Can be abbreviated.

level

(number)
tolerance/confidence level.

nsim

(count)
number of simulations to use.

conditional

(flag)
indicator if the prediction is conditional on the observation or not.

formula

(mmrm_tmb)
same as object.

data

(data.frame)
object in which to construct the frame.

include

(character)
names of variable types to include. Must be NULL or one or more of c("subject_var", "visit_var", "group_var", "response_var").

full

(flag)
indicator whether to return full model frame (deprecated).

na.action

(string)
na action.

x

(mmrm_tmb)
same as object.

sigma

cannot be used (this parameter does not exist in MMRM).

corrected

(flag)
whether corrected AIC should be calculated.

k

(number)
the penalty per parameter to be used; default k = 2 is the classical AIC.

type

(string)
unscaled (response), pearson or normalized. Default is response, and this is the only type available for use with models with a spatial covariance structure.

seed

unused argument from stats::simulate().

method

(string)
simulation method to use. If "conditional", simulated values are sampled given the estimated covariance matrix of object. If "marginal", the variance of the estimated covariance matrix is taken into account.

Details

include argument controls the variables the returned model frame will include. Possible options are "response_var", "subject_var", "visit_var" and "group_var", representing the response variable, subject variable, visit variable or group variable. character values in new data will always be factorized according to the data in the fit to avoid mismatched in levels or issues in model.matrix.

Value

Depends on the method, see Functions.

Functions

• coef(mmrm_tmb): obtains the estimated coefficients.

• fitted(mmrm_tmb): obtains the fitted values.

• predict(mmrm_tmb): predict conditional means for new data; optionally with standard errors and confidence or prediction intervals. Returns a vector of predictions if se.fit == FALSE and interval == "none"; otherwise it returns a data.frame with multiple columns and one row per input data row.

• model.frame(mmrm_tmb): obtains the model frame.

• model.matrix(mmrm_tmb): obtains the model matrix.

• terms(mmrm_tmb): obtains the terms object.

• logLik(mmrm_tmb): obtains the attained log likelihood value.

• formula(mmrm_tmb): obtains the used formula.

• vcov(mmrm_tmb): obtains the variance-covariance matrix estimate for the coefficients.

• VarCorr(mmrm_tmb): obtains the variance-covariance matrix estimate for the residuals.

• deviance(mmrm_tmb): obtains the deviance, which is defined here as twice the negative log likelihood, which can either be integrated over the coefficients for REML fits or the usual one for ML fits.

• AIC(mmrm_tmb): obtains the Akaike Information Criterion, where the degrees of freedom are the number of variance parameters (n_theta). If corrected, then this is multiplied with m / (m - n_theta - 1) where m is the number of observations minus the number of coefficients, or n_theta + 2 if it is smaller than that (Hurvich and Tsai 1989; Burnham and Anderson 1998).

• BIC(mmrm_tmb): obtains the Bayesian Information Criterion, which is using the natural logarithm of the number of subjects for the penalty parameter k.

• print(mmrm_tmb): prints the object.

• residuals(mmrm_tmb): to obtain residuals - either unscaled ('response'), 'pearson' or 'normalized'.

• simulate(mmrm_tmb): simulate responses from a fitted model according to the simulation method, returning a data.frame of dimension ⁠[n, m]⁠ where n is the number of rows in newdata, and m is the number nsim of simulated responses.

References

• Hurvich CM, Tsai C (1989). “Regression and time series model selection in small samples.” Biometrika, 76(2), 297–307. doi:10.2307/2336663.

• Burnham KP, Anderson DR (1998). “Practical use of the information-theoretic approach.” In Model selection and inference, 75–117. Springer. doi:10.1007/978-1-4757-2917-7_3.

• Gałecki A, Burzykowski T (2013). “Linear mixed-effects model.” In Linear mixed-effects models using R, 245–273. Springer.

See Also

mmrm_methods, mmrm_tidiers for additional methods.

Examples

formula <- FEV1 ~ RACE + SEX + ARMCD * AVISIT + us(AVISIT | USUBJID)
object <- fit_mmrm(formula, fev_data, weights = rep(1, nrow(fev_data)))
# Estimated coefficients:
coef(object)
# Fitted values:
fitted(object)
predict(object, newdata = fev_data)
# Model frame:
model.frame(object)
model.frame(object, include = "subject_var")
# Model matrix:
model.matrix(object)
# terms:
terms(object)
terms(object, include = "subject_var")
# Log likelihood given the estimated parameters:
logLik(object)
# Formula which was used:
formula(object)
# Variance-covariance matrix estimate for coefficients:
vcov(object)
# Variance-covariance matrix estimate for residuals:
VarCorr(object)
# REML criterion (twice the negative log likelihood):
deviance(object)
# AIC:
AIC(object)
AIC(object, corrected = TRUE)
# BIC:
BIC(object)
# residuals:
residuals(object, type = "response")
residuals(object, type = "pearson")
residuals(object, type = "normalized")