symbox first transforms x to each of a series
of selected powers, with each transformation standardized to mean 0 and
standard deviation 1. The results are then displayed side-by-side in boxplots, permiting
a visual assessment of which power makes the distribution reasonably symmetric.
For the "lm" method, the response variable in the model is successively
transformed.
symbox(x, ...)
# S3 method for formula
symbox(formula, data=NULL, subset, na.action=NULL, ylab, ...)
# S3 method for default
symbox(x, powers = c(-1, -0.5, 0, 0.5, 1), start,
trans=bcPower, xlab="Powers", ylab, ...)
# S3 method for lm
symbox(x, powers = c(-1, -0.5, 0, 0.5, 1), start, trans=bcPower,
xlab, ylab="Studentized residuals", ...)as returned by boxplot.
a numeric vector.
a one-sided formula specifying a single numeric variable.
as for statistical modeling functions (see, e.g., lm).
axis labels; if ylab is missing, a label will be supplied. For the "lm" method, if xlab is missing, a label will also be supplied.
a vector of selected powers to which x is to be raised.
For meaningful comparison of powers, 1 should be included
in the vector of powers.
a constant to be added to x. If start is missing and
trans is bcPower (the default) or bcnPower, then
a start will be automatically generated if there are zero or negative values in
x, and a warning will be printed; the auto-generated start is the
absolute value of the minimum x plus 1 percent of the range of x.
a transformation function whose first argument is a numeric vector and whose second
argument is a transformation parameter, given by the powers argument; the default
is bcPower, and another possibility is yjPower.
bcnPower may also be used, in which case the gamma parameter
is set to the value of start.
arguments to be passed down.
Gregor Gorjanc, John Fox jfox@mcmaster.ca.
Fox, J. and Weisberg, S. (2019) An R Companion to Applied Regression, Third Edition. Sage.
symbox(~ income, data=Prestige)
symbox(lm(wages ~ education + poly(age, 2) + sex, data=SLID))
Run the code above in your browser using DataLab