Closed ramhiser closed 8 years ago
A good model to follow is given here.
Although the code is for a different example, it's from the same author: ggplot2 heatmap code.
Example plot:
Not as good as the example above, so I'm gonna leave the ticket open, but a good start.
Example code:
library(sparsediscrim)
iris_x <- iris[, -5]
iris_y <- iris$Species
cv_out <- hdrda_cv(x=iris_x, y=iris_y, shrinkage_type = "convex")
plot(cv_out)
Within
hdrda_cv
, acv_summary
is created with the cross-validation error rates for each tuning-parameter pair. Plot a heatmap or similar of the error rates.