gamclass (version 0.62.5)

ldaErr: Calculate Error Rates for Linear Discriminant Model

Description

Given an lda model object, calculate training set error, leave-one-out cross-validation error, and test set error.

Usage

ldaErr(train.lda, train, test, group = "type")

Value

Vector that holds leave-one-out, training, and test error rates

Arguments

train.lda

Fitted lda model object.

train

Training set data frame.

test

Test set data frame.

group

Factor that identifies groups in training data.

Examples

if (FALSE) {
data(spam, package='kernlab')
spam[,-58] <- scale(spam[,-58])
nr <- sample(1:nrow(spam))
spam01 <- spam[nr[1:3601],]     ## Use for training,
spam2 <- spam[nr[3602:4601],]   ## Test
spam01.lda <- lda(type~., data=spam01)
ldaRates <- ldaErr(train.lda=spam01.lda, train=spam01, test=spam2, group="type")
}