com.databricks.labs.automl.model.tools.split
Dataset that contains feature vector, out of DataPrep phase, ready to be split into
number of 'copies' of the split to perform in order to fulfill the number of kFold models to be built
The type of split being performed (i.e. 'stratified', 'random', 'kSample')
Name of the label column
Source directory to use to build the delta persisted data sets if using 'delta' mode in persistMode
'cache', 'persist' or 'delta' - how to retain each of the kFold train/test splits.
The model family in order to determine how many parts in which to repartition the train and test splits for optimal performance.
Wrapper interface for performing the splits, dependent on mode
Wrapper interface for performing the splits, dependent on mode
Array[TrainSplitReferences] from the above methods.
Train / Test split handler class
0.7.1