org.deeplearning4j.nn.simple.multiclass

Class RankClassificationResult

java.lang.Object

org.deeplearning4j.nn.simple.multiclass.RankClassificationResult

All Implemented Interfaces:

Serializable

public class RankClassificationResult
extends Object
implements Serializable

A RankClassificationResult is an abstraction over an activation matrix for ranking classes.

Author:

Adam Gibson

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
RankClassificationResult(org.nd4j.linalg.api.ndarray.INDArray outcome) Takes in just a classification matrix and initializes the labels to just be indices
RankClassificationResult(org.nd4j.linalg.api.ndarray.INDArray outcome, List<String> labels) Takes in a classification matrix and the labels for each column

Method Summary

Modifier and Type	Method and Description
String	maxOutcomeForRow(int r) Get the max index for the given row
List<String>	maxOutcomes()

Methods inherited from class java.lang.Object

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

Constructor Detail

RankClassificationResult

public RankClassificationResult(org.nd4j.linalg.api.ndarray.INDArray outcome)

Takes in just a classification matrix and initializes the labels to just be indices

Parameters:

outcome - the outcome matrix (usually from a softmax or sigmoid output)

RankClassificationResult

public RankClassificationResult(org.nd4j.linalg.api.ndarray.INDArray outcome,
                                List<String> labels)

Takes in a classification matrix and the labels for each column

Parameters:

outcome - the outcome

labels - the labels for the outcomes

Method Detail

maxOutcomeForRow

public String maxOutcomeForRow(int r)

Get the max index for the given row

Parameters:

r - the row to get the max index for

Returns:

the label for the given element

maxOutcomes

public List<String> maxOutcomes()