org.ojalgo.data

Class DataProcessors

java.lang.Object

org.ojalgo.data.DataProcessors

public class DataProcessors
extends Object

Various data processors that could be useful when doing data science or similar. With ojAlgo it is highly advantageous to store data in columns (rather than rows). All the Transformation2D instances in this class assume columns represent variables, and rows samples.

Author:

apete

Field Summary

Fields

Modifier and Type	Field and Description
static Transformation2D<Double>	CENTER Variables centered so that their average will be 0.0
static Transformation2D<Double>	CENTER_AND_SCALE Variables will be centered around 0.0 AND scaled to be [-1.0,1.0].
static Transformation2D<Double>	SCALE Variables scaled to be within [-1.0,1.0] (divide by largest magnitude regardless of sign).
static Transformation2D<Double>	STANDARD_SCORE Will normalise each variable - replace each value with its standard score.

Constructor Summary

Constructors

Constructor and Description
DataProcessors()

Method Summary

Modifier and Type	Method and Description
static <D extends Access2D<?> & Access2D.Sliceable<?>,M extends Mutate2D> M	covariances(Factory2D<M> factory, D data) Variables in columns and samples in rows
static <M extends PhysicalStore<Double>> M	covariances(Factory2D<M> factory, SingularValue<Double> svd)
static <M extends PhysicalStore<Double>> M	covariances(Factory2D<M> factory, SingularValue<Double> svd, double threshold)
static <M extends PhysicalStore<Double>> M	covariances(Factory2D<M> factory, SingularValue<Double> svd, int complexity)
static Transformation2D<Double>	newTransformation2D(Function<SampleSet,UnaryFunction<Double>> definition)

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

CENTER

public static final Transformation2D<Double> CENTER

Variables centered so that their average will be 0.0

CENTER_AND_SCALE

public static final Transformation2D<Double> CENTER_AND_SCALE

Variables will be centered around 0.0 AND scaled to be [-1.0,1.0]. The minimum value will be transformed to -1.0 and the maximum to +1.0.

SCALE

public static final Transformation2D<Double> SCALE

Variables scaled to be within [-1.0,1.0] (divide by largest magnitude regardless of sign). If all values are positive the range will within [0.0,1.0]. If all are negative the range will be within [-1.0,0.0]

STANDARD_SCORE

public static final Transformation2D<Double> STANDARD_SCORE

Will normalise each variable - replace each value with its standard score.

Constructor Detail

DataProcessors

public DataProcessors()

Method Detail

covariances

public static <D extends Access2D<?> & Access2D.Sliceable<?>,M extends Mutate2D> M covariances(Factory2D<M> factory,
                                                                                               D data)

Variables in columns and samples in rows

covariances

public static <M extends PhysicalStore<Double>> M covariances(Factory2D<M> factory,
                                                              SingularValue<Double> svd)

See Also:

covariances(Factory2D, SingularValue, int)

covariances

public static <M extends PhysicalStore<Double>> M covariances(Factory2D<M> factory,
                                                              SingularValue<Double> svd,
                                                              double threshold)

See Also:

covariances(Factory2D, SingularValue, int)

covariances

public static <M extends PhysicalStore<Double>> M covariances(Factory2D<M> factory,
                                                              SingularValue<Double> svd,
                                                              int complexity)

Parameters:

factory - A factory that will produce the returned covariance matrix

svd - A pre-decomposed SVD instance. The original matrix is assumed to have centered data in its columns

complexity - The maximum number of singular values that should be considered

newTransformation2D

public static Transformation2D<Double> newTransformation2D(Function<SampleSet,UnaryFunction<Double>> definition)