Machine learning interface tasks (dqa.tasks.ml)

BalanceClassificationTrainingSet(label_name, ...)	Resamples a training and label data set such that there is an equal number of samples for each label value.
CategoryToOneHot(**kwargs)	Converts index labels into one-hot vectors.
LabelExtractionTask(main_dataset, label_dataset)	Abstract basis class for a task that extracts labels from a separate dataset into one or multiple main datasets.
LabelsByRanges(main_dataset, label_dataset, ...)	Creates a label metadata key from a table in another dataset specifying a range for each label.
Prediction(algo[, direct_params])	Performs a prediction with a machine learning model.
RandomizeAndSubsample([ratio, size])	Randomly subsamples data to a specified relative ratio or absolute size.
TrainTestSplit(input_name, output_name[, ...])	Splits data into training and test data.
Training(input_name, labels_name, algo[, ...])	Performs training with a machine learning model.
UniqueLabels(default_value, **kwargs)	Transforms a 2-dimensional labels array with multiple values in each row (for example from taking windows of a label time series) into a 1-dimensional label vector.