org.deeplearning4j.eval

Class ConfusionMatrix<T extends Comparable<? super T>>

java.lang.Object

org.deeplearning4j.eval.ConfusionMatrix<T>

Type Parameters:

T - The data type used to represent the class labels

public class ConfusionMatrix<T extends Comparable<? super T>>
extends Object

This data structure provides an easy way to build and output a confusion matrix. A confusion matrix is a two dimensional table with a row and table for each class. Each element in the matrix shows the number of test examples for which the actual class is the row and the predicted class is the column. Display of this matrix is useful for identifying when a system is confusing two classes
For more info @see The wikipedia page on Confusion Matrices
Copyright (c) 2011, Regents of the University of Colorado
All rights reserved.

Author:

Lee Becker

Constructor Summary

Constructors

Constructor and Description
ConfusionMatrix() Creates an empty confusion Matrix
ConfusionMatrix(ConfusionMatrix<T> other) Creates a new ConfusionMatrix initialized with the contents of another ConfusionMatrix.

Method Summary

Methods

Modifier and Type	Method and Description
void	add(ConfusionMatrix<T> other) Adds the entries from another confusion matrix to this one.
void	add(T actual, T predicted) Increments the entry specified by actual and predicted by one.
void	add(T actual, T predicted, int count) Increments the entry specified by actual and predicted by count.
int	getActualTotal(T actual) Computes the total number of times the class actually appeared in the data.
SortedSet<T>	getClasses() Gives the set of all classes in the confusion matrix.
int	getCount(T actual, T predicted) Gives the count of the number of times the "predicted" class was predicted for the "actual" class.
int	getPredictedTotal(T predicted) Computes the total number of times the class was predicted by the classifier.
static void	main(String[] args)
String	toCSV() Outputs the ConfusionMatrix as comma-separated values for easy import into spreadsheets
String	toHTML() Outputs Confusion Matrix in an HTML table.
String	toString()

Methods inherited from class java.lang.Object

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

Constructor Detail

ConfusionMatrix

public ConfusionMatrix()

Creates an empty confusion Matrix

ConfusionMatrix

public ConfusionMatrix(ConfusionMatrix<T> other)

Creates a new ConfusionMatrix initialized with the contents of another ConfusionMatrix.

Method Detail

add

public void add(T actual,
       T predicted)

Increments the entry specified by actual and predicted by one.

add

public void add(T actual,
       T predicted,
       int count)

Increments the entry specified by actual and predicted by count.

add

public void add(ConfusionMatrix<T> other)

Adds the entries from another confusion matrix to this one.

getClasses

public SortedSet<T> getClasses()

Gives the set of all classes in the confusion matrix.

getCount

public int getCount(T actual,
           T predicted)

Gives the count of the number of times the "predicted" class was predicted for the "actual" class.

getPredictedTotal

public int getPredictedTotal(T predicted)

Computes the total number of times the class was predicted by the classifier.

getActualTotal

public int getActualTotal(T actual)

Computes the total number of times the class actually appeared in the data.

toString

public String toString()

Overrides:

toString in class Object

toCSV

public String toCSV()

Outputs the ConfusionMatrix as comma-separated values for easy import into spreadsheets

toHTML

public String toHTML()

Outputs Confusion Matrix in an HTML table. Cascading Style Sheets (CSS) can control the table's appearance by defining the empty-space, actual-count-header, predicted-class-header, and count-element classes. For example

Returns:

html string

main

public static void main(String[] args)