Class | Description |
---|---|
CrossValidation |
Splits the training data up into CV folds with a given seed.
|
Default |
The most lame of InstanceGenerators, does nothing
|
MultiLevel |
Experimental InstanceGenerator that takes as input a child classifier, and creates multiple levels of training data.
|
RandomSubSampling |
Experimental InstanceGenerator that splits the data up into random folds, with a fixed percentage used for training
instanceGeneratorArguments: A property string with the following
startingseed - the initial seed of the splits
numSamples - the number of folds to gererate
percent - the percentage of data to use in the training set
bias - the bias towards a uniform partition
instance string format:
seed - the seed used to split up data
percent - the percentage of data to use in the training set
bias - the bias towards a uniform partition
|
TerminationHoldout |
Experimental InstanceGenerator that takes as input a child classifier, and holds back a bunch of data as a 'Termination' set from the SMBO methods.
|