CODEWITHSUNDEEP
rknnconfusion-matrixcaret
Regis Maria O'Connor
Regis Maria O'Connor

Reputation: 43

confusionMatrix() caret package in R - what is happening under the hood in the Class 3 Example on the help screen?

In the R help page for confusionMatrix() in the caret package there is this 3 class example:

confusionMatrix(iris$Species, sample(iris$Species))

newPrior <- c(.05, .8, .15)
names(newPrior) <- levels(iris$Species)

confusionMatrix(iris$Species, sample(iris$Species))

How is the newPrior value getting incorporated into the Confusion Matrix?

Upvotes: 0

Views: 59

Answers (1)

Marek Fiołka
Marek Fiołka

Reputation: 4949

In this example, newPrior is not included in the error confusion matrices in any way!!

You may have been confused that the two matrices are different. However, this is due to the pseudo-random nature of the sample function. If you add the grain alignment before each call, you will find that in both cases the confusion matrices are identical!

library(caret)
set.seed(123)
cf1 = confusionMatrix(iris$Species, sample(iris$Species))

newPrior <- c(.05, .8, .15)
names(newPrior) <- levels(iris$Species)

set.seed(123)
cf2 = confusionMatrix(iris$Species, sample(iris$Species))

identical(cf1, cf2)

output

[1] TRUE

Upvotes: 1

Related Questions