jsat.classifiers

Class CategoricalResults

java.lang.Object

jsat.classifiers.CategoricalResults

All Implemented Interfaces:

Serializable, Cloneable

public class CategoricalResults
extends Object
implements Cloneable, Serializable

This class represents the probabilities for each possible result classification.

Author:

Edward Raff

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
CategoricalResults(double[] probabilities) Creates a new Categorical Result using the given array.
CategoricalResults(int numCategories) Create a new Categorical Results, values will default to all zero.

Method Summary

Modifier and Type	Method and Description
CategoricalResults	clone() Creates a deep clone of this
void	divideConst(double c) Divides all the probabilities by a constant value in order to scale them
double	getProb(int cat) Returns the stored probability for the given category
Vec	getVecView()
void	incProb(int cat, double prob) Increments the stored probability that a sample belongs to a given category
int	mostLikely() Returns the category that is the most likely according to the current probability values
void	normalize() Adjusts the probabilities by dividing each value by the total sum, so that all values are in the range [0, 1]
void	setProb(int cat, double prob) Sets the probability that a sample belongs to a given category.
int	size() Returns the number of classes that are in the result.
String	toString()

Methods inherited from class java.lang.Object

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

Constructor Detail

CategoricalResults

public CategoricalResults(int numCategories)

Create a new Categorical Results, values will default to all zero.

Parameters:

numCategories - the number of options to support.

CategoricalResults

public CategoricalResults(double[] probabilities)

Creates a new Categorical Result using the given array. It will use only a reference to the given array, and will assume the values are already normalized and sum to one.

Parameters:

probabilities - the array of probabilities for each outcome

Method Detail

size

public int size()

Returns the number of classes that are in the result.

Returns:

the class count

setProb

public void setProb(int cat,
                    double prob)

Sets the probability that a sample belongs to a given category.

Parameters:

cat - the category

prob - the value to set, may be greater then one.

Throws:

IndexOutOfBoundsException - if a non existent category is specified

ArithmeticException - if the value set is negative or not a number

incProb

public void incProb(int cat,
                    double prob)

Increments the stored probability that a sample belongs to a given category

Parameters:

cat - the category

prob - the value to increment by, may be greater then one.

Throws:

IndexOutOfBoundsException - if a non existent category is specified

ArithmeticException - if the value set is negative or not a number

mostLikely

public int mostLikely()

Returns the category that is the most likely according to the current probability values

Returns:

the the most likely category

divideConst

public void divideConst(double c)

Divides all the probabilities by a constant value in order to scale them

Parameters:

c - the constant to divide all probabilities by

normalize

public void normalize()

Adjusts the probabilities by dividing each value by the total sum, so that all values are in the range [0, 1]

getVecView

public Vec getVecView()

getProb

public double getProb(int cat)

Returns the stored probability for the given category

Parameters:

cat - the category

Returns:

the associated probability

clone

public CategoricalResults clone()

Creates a deep clone of this

Overrides:

clone in class Object

Returns:

a deep clone

toString

public String toString()

Overrides:

toString in class Object