Quantile Plot for LMS Quantile Regression

Plots quantiles associated with a LMS quantile regression.

qtplot.lmscreg(object, newdata = NULL,
               percentiles = object@misc$percentiles,
               show.plot = TRUE, ...)

Arguments

object: A VGAM quantile regression model, i.e., an object produced by modelling functions such as vglm and vgam with a family function beginning with "lms.", e.g., lms.yjn.
newdata: Optional data frame for computing the quantiles. If missing, the original data is used.
percentiles: Numerical vector with values between 0 and 100 that specify the percentiles (quantiles). The default are the percentiles used when the model was fitted.
show.plot: Logical. Plot it? If FALSE no plot will be done.
...: Graphical parameter that are passed into plotqtplot.lmscreg.

Details

The `primary' variable is defined as the main covariate upon which the regression or smoothing is performed. For example, in medical studies, it is often the age. In VGAM, it is possible to handle more than one covariate, however, the primary variable must be the first term after the intercept.

Value

A list with the following components.

fitted.values: A vector of fitted percentile values.
percentiles: The percentiles used.

References

Yee, T. W. (2004). Quantile regression via vector generalized additive models. Statistics in Medicine, 23, 2295–2315.

Author

Thomas W. Yee

Note

plotqtplot.lmscreg does the actual plotting.

Examples

if (FALSE) { # \dontrun{
fit <- vgam(BMI ~ s(age, df = c(4, 2)), lms.bcn(zero=1), bmi.nz)
qtplot(fit)
qtplot(fit, perc = c(25, 50, 75, 95), lcol = 4, tcol = 4, llwd = 2)
} # }