Summarize and Plot a Penalized MARS Model
Usage
# S3 method for class 'mars_lasso'
summary(object, ...)
# S3 method for class 'summary.mars_lasso'
print(x, digits = 4L, ...)
# S3 method for class 'mars_lasso'
print(x, digits = 4L, ...)
# S3 method for class 'mars_lasso'
plot(x, type = c("cv", "coefficients"), ...)
Arguments
- object, x
Fitted mars_lasso object or its summary, as appropriate.
- ...
Additional graphical parameters for plot; unsupported summary and
print arguments are rejected.
- digits
Significant digits to print.
- type
cv plots validation scores and approximate fold standard errors;
coefficients plots the supplied-unit coefficient path. Zero lambdas use a
linear horizontal axis; positive-only paths use a logarithmic axis.
Value
Summary returns point estimates, variance components, tuning scores
and diagnostics. Print and plot return their input invisibly.
Details
Selected-model coefficients have no ordinary Wald tests or intervals.
The approximate 1se tuning threshold is not a confidence interval. Variance
components are conditional ML point estimates under the penalized mean model.