errors: Obtain the test errors result.

In liquidSVM: A Fast and Versatile SVM Package

Description

After calculating the result in test.liquidSVM if labels were given liquidSVM also calculates the test error.

Usage

errors(y, showall = FALSE)

Arguments

y	the results of test.liquidSVM
showall	show the more detailed errors as well.

Details

Depending on the learning scenario there can be multiple errors: usually there is one per task, and mcSVM adds in front the global classification error. In the latter case the names give an information for what task the error was computed.

For each error also the positive and negative validation error can be shonw using showall for example in rocSVM.

Value

for all tasks the global and optionally also the positive/negative errors. Depending on the learning scenario there can be also a overall error (e.g. in multi-class classification).

See Also

test.liquidSVM

Examples

modelTrees <- svm(Height ~ Girth + Volume, trees[1:10, ])  # least squares

y <- test(modelTrees,trees[-1:-10,])
errors(y)

## Not run: 
banana <- liquidData('banana-bc')
s_banana <- rocSVM(Y~., banana$test)
result <- test(s_banana, banana$train)
errors(result, showall=TRUE)

## End(Not run)

liquidSVM documentation built on Sept. 15, 2019, 1:02 a.m.