cascadesvm: cascade SVM

Description Usage Arguments Details Value

Description

we do not restrict the number of subsets. in case at any point there is an uneven number of splits we combine all of them normally (two by two), except for the last three subsets. these we combine directly.

Usage

1
cascadesvm(X, Y, k, epochs = 1, train.param = NULL, verbose = FALSE, ...)

Arguments

X

training input

Y

labels

k

number of initial subsets.

epochs

number of epochs to run

Details

FIXME: -stratified sampling,

Value

svm models, containing: -final model (libsvm model), -?


aydindemircioglu/crushSVM documentation built on May 11, 2019, 4:14 p.m.