ggmix (version 0.0.2)

plot.ggmix_gic: Plot the Generalised Information Criteria curve produced by gic

Description

Plots the Generalised Information Criteria curve, as a function of the lambda values used

Usage

# S3 method for ggmix_gic
plot(
  x,
  ...,
  sign.lambda = 1,
  type = c("gic", "QQranef", "QQresid", "predicted", "Tukey-Anscombe"),
  s = "lambda.min",
  newy,
  newx
)
plotGIC(x, sign.lambda, lambda.min, ...)

Arguments

x

fitted linear mixed model object of class ggmix_gic from the gic function

...

Other graphical parameters to plot

sign.lambda

Either plot against log(lambda) (default) or its negative if sign.lambda=-1

type

gic returns a plot of the GIC vs. log(lambda). QQranef return a qqplot of the random effects. QQresid returns a qqplot of the residuals which is \(y - X\beta - b_i\) where b_i is the subject specific random effect. predicted returns a plot of the predicted response (\(X \beta\) + b_i) vs. the observed response, where b_i is the subject specific random effect. Tukey-Anscombe returns a plot of the residuals vs. fitted values (\(X \beta\))

s

Value of the penalty parameter lambda at which predictions are required. Default is the value s="lambda.min". If s is numeric, it is taken as the value of lambda to be used. Must be a single value of the penalty parameter lambda at which coefficients will be extracted via the coef method for objects of class ggmix_gic. If more than one is supplied, only the first one will be used.

newy

the response variable that was provided to ggmix. this is only required for type="QQresis", type="Tukey-Anscombe" and type="predicted"

newx

matrix of values for x at which predictions are to be made. Do not include the intercept. this is only required for type="QQresis", type="Tukey-Anscombe" and type="predicted"

lambda.min

the value of lambda which minimizes the gic

Value

plot depends on the type selected

Details

A plot is produced, and nothing is returned.

See Also

gic

Examples

# NOT RUN {
data("admixed")
fit <- ggmix(x = admixed$xtrain,
             y = admixed$ytrain,
             kinship = admixed$kin_train)
hdbic <- gic(fit)

# plot solution path
plot(fit)

# plot HDBIC curve as a function of lambda
plot(hdbic)
# }