| summary.earth {earth} | R Documentation |
Summary method for earth objects.
## S3 method for class 'earth':
summary(object = stop("no 'object' arg"),
details = FALSE, decomp = "anova",
style = c("h", "pmax"),
digits = getOption("digits"), fixed.point=TRUE, ...)
## S3 method for class 'summary.earth':
print(x = stop("no 'x' arg"),
details = x$details, decomp = x$decomp,
digits = x$digits, fixed.point = x$fixed.point, ...)
object |
An earth object.
This is the only required argument for summary.earth.
|
x |
A summary.earth object.
This is the only required argument for print.summary.earth.
|
details |
Default is FALSE.
Use TRUE for earth
to print more information about earth–glm models.
But note that the displayed P-values of the GLM coefficients
are meaningless because of the amount of preprocessing
by earth to select the regression terms.
|
decomp |
See format.earth for legal decomp values.
Default is "anova".
Use "none" to order the terms as created by the forward.pass.
|
style |
Formatting style. One of"h" (default) more compact"pmax" for those who prefer it and for compatibility with old versions of earth.
|
digits |
The number of significant digits.
For summary.earth, the default is getOption("digits").
For print.summary.earth, the default is the $digits component of object.
|
fixed.point |
Method of printing numbers in matrices.
Default is TRUE which prints like this
(making it easier to compare coefficients):
(Intercept) 15.029
h(temp-58) 0.313
h(234-ibt) -0.046
...
whereas fixed.point=FALSE prints like this (which is more usual in R):
(Intercept) 1.5e+01
h(temp-58) 3.1e-01
h(234-ibt) -4.6e-02
...
Matrices with two or fewer rows are never printed with a fixed point.
|
... |
Extra arguments are passed to format.earth.
|
The value is the same as that returned by earth
but with the following extra components.
strings |
String(s) created by format.earth.
For multiple response models, a vector of strings.
|
digits |
|
details |
|
decomp |
|
fixed.point |
The corresponding arguments, passed on to print.summary.earth.
|
The printed Estimated importance uses evimp
with the nsubsets criterion.
The most important predictor is printed first, and so on.
a <- earth(Volume~ ., data = trees) summary(a, digits = 2) # yields: # Call: earth(formula=Volume~., data=trees) # # Volume # (Intercept) 23.21 # h(Girth-12.9) 5.75 # h(12.9-Girth) -2.87 # h(Height-76) 0.72 # # Selected 4 of 5 terms, and 2 of 2 predictors # Estimated importance: Girth Height # Number of terms at each degree of interaction: 1 3 (additive model) # GCV 11 RSS 213 GRSq 0.96 RSq 0.97