ucar.unidata.data.grid

Class GridUtil.QuantileTransformer

java.lang.Object

ucar.unidata.data.grid.GridUtil.QuantileTransformer

All Implemented Interfaces:

java.io.Serializable, java.lang.Comparable<GridUtil.QuantileTransformer>

Enclosing class:

GridUtil

public static class GridUtil.QuantileTransformer
extends java.lang.Object
implements java.io.Serializable, java.lang.Comparable<GridUtil.QuantileTransformer>

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
QuantileTransformer(int nQuantiles, int subsample) Create an empty quantile transformer with the specified number of quantiles and subsample.
QuantileTransformer(int nQuantiles, int subsample, double[] references, double[] quantiles) Construct a quantile transformer from an existing internal representation.

Method Summary

Modifier and Type	Method and Description
void	clearCache() Clears the cache of transformers.
int	compareTo(GridUtil.QuantileTransformer other)
<I> void	fit(I x)
int	getColInputCount()
int	getColOutputCount()
GridUtil.QuantileTransformer	getQuantileTransformerFromCache(GridUtil.QuantileTransformer transformer) Stores transformers in a private cache.
void	shutdown()
<I,O> O	transform(I x, java.lang.Class<O> returnType, boolean rowMajorOrder)

Methods inherited from class java.lang.Object

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

Constructor Detail

QuantileTransformer

public QuantileTransformer(int nQuantiles,
                           int subsample)

Create an empty quantile transformer with the specified number of quantiles and subsample.

Parameters:

nQuantiles - Number of quantiles

subsample - Number of subsamples

QuantileTransformer

public QuantileTransformer(int nQuantiles,
                           int subsample,
                           double[] references,
                           double[] quantiles)

Construct a quantile transformer from an existing internal representation.

Only not use this constructor if you have the internal state of an existing quantile transformer!

This constructor is useful from porting a scikit-learn sklearn.preprocessing.QuantileTransformer since the implementations share the same internal objects.

Parameters:

nQuantiles - Number of quantiles

subsample - Number of subsamples

references - The references

quantiles - The quantiles

Method Detail

getQuantileTransformerFromCache

public GridUtil.QuantileTransformer getQuantileTransformerFromCache(GridUtil.QuantileTransformer transformer)

Stores transformers in a private cache. When an existing transformer is found that results in compareTo==0 with transformer then the cached object will be returned. Otherwise transformer will be returned.

Instead of making the QuantileTransformer constructor private, set your local transformer to the result of this cache method to prevent duplicating transformers in memory.

The implementation synchronizes on the transformer cache so this method is thread-safe.

Parameters:

transformer - The transformer

Returns:

A cached copy of the transformer

clearCache

public void clearCache()

Clears the cache of transformers.

The implementation synchronizes on the transformer cache so this method is thread-safe.

fit

public <I> void fit(I x)

transform

public <I,O> O transform(I x,
                         java.lang.Class<O> returnType,
                         boolean rowMajorOrder)

getColInputCount

public int getColInputCount()

getColOutputCount

public int getColOutputCount()

shutdown

public void shutdown()

compareTo

public int compareTo(GridUtil.QuantileTransformer other)

Specified by:

compareTo in interface java.lang.Comparable<GridUtil.QuantileTransformer>