spruce-multiple: Spruce up multi-outcome predictions
In hardhat: Construct Modeling Packages

spruce-multiple

R Documentation

Spruce up multi-outcome predictions

Description

This family of ⁠spruce_*_multiple()⁠ functions converts multi-outcome predictions into a standardized format. They are generally called from a prediction implementation function for the specific type of prediction to return.

Usage

spruce_numeric_multiple(...)

spruce_class_multiple(...)

spruce_prob_multiple(...)

Arguments

...

Multiple vectors of predictions:

For spruce_numeric_multiple(), numeric vectors of equal size.
For spruce_class_multiple(), factors of "hard" class predictions of equal size.
For spruce_prob_multiple(), tibbles of equal size, which are the result of calling spruce_prob() on each matrix of prediction probabilities.

If the ... are named, then this name will be used as the suffix on the resulting column name, otherwise a positional index will be used.

Value

For spruce_numeric_multiple(), a tibble of numeric columns named with the pattern ⁠.pred_*⁠.
For spruce_class_multiple(), a tibble of factor columns named with the pattern ⁠.pred_class_*⁠.
For spruce_prob_multiple(), a tibble of data frame columns named with the pattern ⁠.pred_*⁠.

Examples

spruce_numeric_multiple(1:3, foo = 2:4)

spruce_class_multiple(
  one_step = factor(c("a", "b", "c")),
  two_step = factor(c("a", "c", "c"))
)

one_step <- matrix(c(.3, .7, .0, .1, .3, .6), nrow = 2, byrow = TRUE)
two_step <- matrix(c(.2, .7, .1, .2, .4, .4), nrow = 2, byrow = TRUE)
binary <- matrix(c(.5, .5, .4, .6), nrow = 2, byrow = TRUE)

spruce_prob_multiple(
  one_step = spruce_prob(c("a", "b", "c"), one_step),
  two_step = spruce_prob(c("a", "b", "c"), two_step),
  binary = spruce_prob(c("yes", "no"), binary)
)

hardhat documentation built on March 31, 2023, 10:21 p.m.