jsat.datatransform

Interface InPlaceTransform

All Superinterfaces:

Cloneable, DataTransform, Serializable

All Known Subinterfaces:

InPlaceInvertibleTransform

All Known Implementing Classes:

AutoDeskewTransform, Imputer, InsertMissingValuesTransform, LinearTransform, PNormNormalization, StandardizeTransform, UnitVarianceTransform, WhitenedZCA, ZeroMeanTransform

public interface InPlaceTransform
extends DataTransform

An In Place Transform is one that has the same number of categorical and numeric features as the input. This means it can mutableTransform the input data point instead of allocating a new one, which can reduce overhead on memory allocations. This can be useful when performing many data transforms in cross validation or when processing new examples in an environment that is applying an already learned model.

This interface is assumed that it will be applied to numeric features. Incase this is not true, a mutatesNominal() method is provided for the implementation to indicate otherwise.

Author:

Edward Raff

Method Summary

Modifier and Type	Method and Description
void	mutableTransform(DataPoint dp) Mutates the given data point.
boolean	mutatesNominal() By default returns false.

Methods inherited from interface jsat.datatransform.DataTransform

clone, fit, transform

Method Detail

mutableTransform

void mutableTransform(DataPoint dp)

Mutates the given data point. This causes side effects, altering the data point to have the same value as the output of DataTransform.transform(jsat.classifiers.DataPoint).

Parameters:

dp - the data point to alter

mutatesNominal

boolean mutatesNominal()

By default returns false. Only returns true if this transform will mutableTransform the nominal feature values of a data point.

Returns:

true if nominal feature values are mutated, false otherwise.