org.deeplearning4j.eval

Class RegressionEvaluation

java.lang.Object

org.deeplearning4j.eval.RegressionEvaluation

public class RegressionEvaluation
extends Object

Evaluation method for the evaluation of regression algorithms.
Provides the following metrics, for each column:
- MSE: mean squared error
- MAE: mean absolute error
- RMSE: root mean squared error
- RSE: relative squared error
- correlation coefficient
See for example: http://www.saedsayad.com/model_evaluation_r.htm For classification, see Evaluation

Author:

Alex Black

Field Summary

Fields

Modifier and Type	Field and Description
static int	DEFAULT_PRECISION

Constructor Summary

Constructors

Constructor and Description
RegressionEvaluation(int nColumns) Create a regression evaluation object with the specified number of columns, and default precision for the stats() method.
RegressionEvaluation(int nColumns, int precision) Create a regression evaluation object with the specified number of columns, and specified precision for the stats() method.
RegressionEvaluation(List<String> columnNames) Create a regression evaluation object with default precision for the stats() method
RegressionEvaluation(List<String> columnNames, int precision) Create a regression evaluation object with specified precision for the stats() method
RegressionEvaluation(String... columnNames) Create a regression evaluation object with default precision for the stats() method

Method Summary

Modifier and Type	Method and Description
double	correlationR2(int column)
void	eval(org.nd4j.linalg.api.ndarray.INDArray labels, org.nd4j.linalg.api.ndarray.INDArray predictions)
void	evalTimeSeries(org.nd4j.linalg.api.ndarray.INDArray labels, org.nd4j.linalg.api.ndarray.INDArray predictions) Convenience method for evaluation of time series.
void	evalTimeSeries(org.nd4j.linalg.api.ndarray.INDArray labels, org.nd4j.linalg.api.ndarray.INDArray predictions, org.nd4j.linalg.api.ndarray.INDArray outputMask) Evaluate a time series, whether the output is masked usind a masking array.
double	meanAbsoluteError(int column)
double	meanSquaredError(int column)
int	numColumns()
double	relativeSquaredError(int column)
double	rootMeanSquaredError(int column)
String	stats()

Methods inherited from class java.lang.Object

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

Field Detail

DEFAULT_PRECISION

public static final int DEFAULT_PRECISION

See Also:

Constant Field Values

Constructor Detail

RegressionEvaluation

public RegressionEvaluation(int nColumns)

Create a regression evaluation object with the specified number of columns, and default precision for the stats() method.

Parameters:

nColumns - Number of columns

RegressionEvaluation

public RegressionEvaluation(int nColumns,
                            int precision)

Create a regression evaluation object with the specified number of columns, and specified precision for the stats() method.

Parameters:

nColumns - Number of columns

RegressionEvaluation

public RegressionEvaluation(String... columnNames)

Create a regression evaluation object with default precision for the stats() method

Parameters:

columnNames - Names of the columns

RegressionEvaluation

public RegressionEvaluation(List<String> columnNames)

Create a regression evaluation object with default precision for the stats() method

Parameters:

columnNames - Names of the columns

RegressionEvaluation

public RegressionEvaluation(List<String> columnNames,
                            int precision)

Create a regression evaluation object with specified precision for the stats() method

Parameters:

columnNames - Names of the columns

Method Detail

eval

public void eval(org.nd4j.linalg.api.ndarray.INDArray labels,
                 org.nd4j.linalg.api.ndarray.INDArray predictions)

evalTimeSeries

public void evalTimeSeries(org.nd4j.linalg.api.ndarray.INDArray labels,
                           org.nd4j.linalg.api.ndarray.INDArray predictions)

Convenience method for evaluation of time series. Reshapes time series (3d) to 2d, then calls eval

See Also:

eval(INDArray, INDArray)

evalTimeSeries

public void evalTimeSeries(org.nd4j.linalg.api.ndarray.INDArray labels,
                           org.nd4j.linalg.api.ndarray.INDArray predictions,
                           org.nd4j.linalg.api.ndarray.INDArray outputMask)

Evaluate a time series, whether the output is masked usind a masking array. That is, the mask array specified whether the output at a given time step is actually present, or whether it is just padding.
For example, for N examples, nOut output size, and T time series length: labels and predicted will have shape [N,nOut,T], and outputMask will have shape [N,T].

See Also:

evalTimeSeries(INDArray, INDArray)

stats

public String stats()

numColumns

public int numColumns()

meanSquaredError

public double meanSquaredError(int column)

meanAbsoluteError

public double meanAbsoluteError(int column)

rootMeanSquaredError

public double rootMeanSquaredError(int column)

correlationR2

public double correlationR2(int column)

relativeSquaredError

public double relativeSquaredError(int column)