createMultiFoldsPu: createMultiFoldsPu

Description Usage Arguments See Also

Description

Create test/training partition for PU data. The positive data is split into k groups while the unlabeled data is used completely in every fold. If you want to use standard cross-validation use createFolds, which served as template for this function.

Usage

1
2
createMultiFoldsPu(y, k, times = 5, positive = NULL, indepUn = NULL,
  seed = NULL)

Arguments

y

a vector of outcomes for the positive and the negative class

k

an integer for the number of folds (applied to the positive class)

times

the number of repetitions to create

positive

the positive class in y. if empty the label with the smaller frequency is assumed to be the positive class.

indepUn

optional, the elements in y which should always been used in the test group. If not given and the indices are passed to trainOcc they are randomly sampled frm the argument u.

seed

an integer in order to set a seed point

See Also

createMultiFoldsPu, createFolds


benmack/oneClass documentation built on Dec. 15, 2020, 7:38 p.m.