Plot the fitted error distribution.
# S3 method for smoothSurvReg
plot(x, plot = TRUE, resid = TRUE, knots = TRUE,
compare = TRUE, components = FALSE, standard = TRUE,
by, toler.c = 1e-5,
xlim, ylim,
xlab = expression(epsilon), ylab = expression(paste("f(",epsilon,")", sep = "")),
type = "l", lty = 1, main, sub, bty = "n", ...)A dataframe with columns named x and y where x gives the grid
and y the values of the density at that grid.
Object of class smoothSurvReg.
If TRUE the plot is directly produced, otherwise only a data.frame
with information used for later plotting is returned.
If resid & plot residuals are added
to the plot in the following way. Residuals based on exact observations
are plotted by pch = 3 (+), residuals based on right censored
observations using pch = 4 (x), residuals based on left censored
observations using pch = 2 (traingle) and midpoints of
residuals based on interval censored observations using
pch = 5 (diamond).
If knots & plot bullets indicating (transformed)
knots corresponding to 'c' G-spline coefficients higher than toler.c
are added to the x-axis of the plot.
If compare & !components & standard & plot plots of the three
parametric distributions (st. normal, st. logistic and st. minimum extreme value)
are added to the plot for comparison.
If components & standard & plot dashed lines with weighted
G-spline components are added to the plot.
If standard data for plotting the fitted error distribution
are created, otherwise data for plotting the distribution
of \(\alpha + \sigma\varepsilon\) are created.
In the case that \(\log(\sigma)\) depends on covariates
and standard is FALSE, all covariates equal to zero are used
to compute \(\log(\sigma)\).
Step for a ploting grid. If NULL it is automatically computed.
Tolerance to indicate zero 'c' G-spline coefficients used if knots == TRUE.
Arguments passed to the plot function.
Arguments passed to the plot function.
Arguments passed to the plot function.
Arguments passed to the plot function.
Argument passed to the plot function.
Arguments passed to the plot function.
Arnošt Komárek arnost.komarek@mff.cuni.cz
smoothSurvReg, plot