org.tribuo.regression

Class RegressionInfo

java.lang.Object

org.tribuo.regression.RegressionInfo

All Implemented Interfaces:

Serializable, OutputInfo<Regressor>

Direct Known Subclasses:

ImmutableRegressionInfo, MutableRegressionInfo

public abstract class RegressionInfo
extends Object
implements OutputInfo<Regressor>

The base class for regression information using Regressors.

Stores the observed min, max, mean and variance for each dimension.

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
protected Map<String,com.oracle.labs.mlrg.olcut.util.MutableLong>	countMap
protected Map<String,com.oracle.labs.mlrg.olcut.util.MutableDouble>	maxMap
protected Map<String,com.oracle.labs.mlrg.olcut.util.MutableDouble>	meanMap
protected Map<String,com.oracle.labs.mlrg.olcut.util.MutableDouble>	minMap
protected long	overallCount
protected Map<String,com.oracle.labs.mlrg.olcut.util.MutableDouble>	sumSquaresMap
protected int	unknownCount

Method Summary

Modifier and Type	Method and Description
abstract RegressionInfo	copy() Generates a copy of this OutputInfo, including it's mutability.
ImmutableOutputInfo<Regressor>	generateImmutableOutputInfo() Generates an ImmutableOutputInfo which has a copy of the data in this OutputInfo, but also has id values and is immutable.
MutableOutputInfo<Regressor>	generateMutableOutputInfo() Generates a mutable copy of this OutputInfo.
Set<Regressor>	getDomain() Returns a set containing a Regressor for each dimension with the minimum value observed.
double	getMax(String name) Gets the maximum value this RegressionInfo has seen, or NaN if it's not seen that dimension.
double	getMean(String name) Gets the mean value this RegressionInfo has seen, or NaN if it's not seen that dimension.
double	getMin(String name) Gets the minimum value this RegressionInfo has seen, or NaN if it's not seen anything.
int	getUnknownCount() Returns the number of unknown Output instances (generated by OutputFactory.getUnknownOutput()) that this OutputInfo has seen.
double	getVariance(String name) Gets the variance this RegressionInfo has seen, or NaN if it's not seen that dimension.
Iterable<com.oracle.labs.mlrg.olcut.util.Pair<String,Long>>	outputCountsIterable() An Iterable over the possible outputs and the number of times they were observed.
int	size() The number of dimensions this OutputInfo has seen.

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.tribuo.OutputInfo

toReadableString

Field Detail

maxMap

protected Map<String,com.oracle.labs.mlrg.olcut.util.MutableDouble> maxMap

minMap

protected Map<String,com.oracle.labs.mlrg.olcut.util.MutableDouble> minMap

meanMap

protected Map<String,com.oracle.labs.mlrg.olcut.util.MutableDouble> meanMap

sumSquaresMap

protected Map<String,com.oracle.labs.mlrg.olcut.util.MutableDouble> sumSquaresMap

countMap

protected Map<String,com.oracle.labs.mlrg.olcut.util.MutableLong> countMap

overallCount

protected long overallCount

unknownCount

protected int unknownCount

Method Detail

getUnknownCount

public int getUnknownCount()

Description copied from interface: OutputInfo

Returns the number of unknown Output instances (generated by OutputFactory.getUnknownOutput()) that this OutputInfo has seen.

Specified by:

getUnknownCount in interface OutputInfo<Regressor>

Returns:

The number of unknown outputs observed.

getDomain

public Set<Regressor> getDomain()

Returns a set containing a Regressor for each dimension with the minimum value observed.

Specified by:

getDomain in interface OutputInfo<Regressor>

Returns:

A set of Regressors, each with one active dimension.

getMin

public double getMin(String name)

Gets the minimum value this RegressionInfo has seen, or NaN if it's not seen anything.

Parameters:

name - The dimension to check.

Returns:

The minimum value for that dimension.

getMax

public double getMax(String name)

Gets the maximum value this RegressionInfo has seen, or NaN if it's not seen that dimension.

Parameters:

name - The dimension to check.

Returns:

The maximum value for that dimension.

getMean

public double getMean(String name)

Gets the mean value this RegressionInfo has seen, or NaN if it's not seen that dimension.

Parameters:

name - The dimension to check.

Returns:

The mean value for that dimension.

getVariance

public double getVariance(String name)

Gets the variance this RegressionInfo has seen, or NaN if it's not seen that dimension.

Parameters:

name - The dimension to check.

Returns:

The variance for that dimension.

size

public int size()

The number of dimensions this OutputInfo has seen.

Specified by:

size in interface OutputInfo<Regressor>

Returns:

The number of dimensions this OutputInfo has seen.

generateImmutableOutputInfo

public ImmutableOutputInfo<Regressor> generateImmutableOutputInfo()

Description copied from interface: OutputInfo

Generates an ImmutableOutputInfo which has a copy of the data in this OutputInfo, but also has id values and is immutable.

Specified by:

generateImmutableOutputInfo in interface OutputInfo<Regressor>

Returns:

An immutable copy of this OutputInfo.

generateMutableOutputInfo

public MutableOutputInfo<Regressor> generateMutableOutputInfo()

Description copied from interface: OutputInfo

Generates a mutable copy of this OutputInfo.

Specified by:

generateMutableOutputInfo in interface OutputInfo<Regressor>

Returns:

A mutable copy of this OutputInfo.

copy

public abstract RegressionInfo copy()

Description copied from interface: OutputInfo

Generates a copy of this OutputInfo, including it's mutability.

Specified by:

copy in interface OutputInfo<Regressor>

Returns:

A copy of the OutputInfo.

outputCountsIterable

public Iterable<com.oracle.labs.mlrg.olcut.util.Pair<String,Long>> outputCountsIterable()

Description copied from interface: OutputInfo

An Iterable over the possible outputs and the number of times they were observed.

Provides similar functionality to an iterator over the entrySet of a Map, though this is a copy of the data and immutable.

Specified by:

outputCountsIterable in interface OutputInfo<Regressor>

Returns:

Lambda wrapping an iterator over Pairs.