Skip to contents

Random-Effect Variance Components for a TMB Demand Model

Source: R/tmb-methods.R
VarCorr.beezdemand_tmb.Rd

Extracts the random-effect variance components from a beezdemand_tmb fit in the matrix layout produced by nlme::VarCorr(), so users familiar with nlme or lme4 can introspect a TMB fit with the accessor they already know. The reported values are the same ones returned by summary.beezdemand_tmb: the Q0 and alpha random-effect standard deviations on the log10 scale and the residual standard deviation on the model's likelihood scale. This is a presentation shim — it formats already-computed values and recomputes nothing.

Usage

# S3 method for class 'beezdemand_tmb'
VarCorr(x, sigma = 1, rdig = 3, ...)

Arguments

x

A beezdemand_tmb object.

sigma

Present for signature compatibility with nlme::VarCorr(). The TMB summary reports variance components as absolute standard deviations, so there is no residual scale factor to apply; any value other than the default (1) is an error.

rdig

Integer. Number of significant digits used when formatting the displayed values. Default 3.

...

Unused; present for generic compatibility.

Value

A character matrix of class "VarCorr.lme" with one row per random-effect term plus a final "Residual" row, columns "Variance" and "StdDev", and — for fits with correlated random effects (pdSymm) — a "Corr" column. print() dispatches to nlme's print.VarCorr.lme().

Note

The Corr column is placed using nlme's convention — each correlation on the row of its higher-indexed random effect. For multi-block pdBlocked fits the correlations are placed on the correct global rows (each correlated block's off-diagonals are offset by the cumulative random-effect dimension of the earlier blocks); summary(x)$correlations remains available for the labelled values.

See also

VarCorr, summary.beezdemand_tmb

Examples

# \donttest{
data(apt)
fit <- fit_demand_tmb(apt, equation = "exponential", verbose = 0)
#>   equation='exponential': Dropped 14 zero-consumption observations (146 remaining).
VarCorr(fit)
#>          Variance StdDev Corr  
#> Q0       0.0281   0.167        
#> alpha    0.0397   0.199  -0.436
#> Residual 0.0202   0.142        
# }