jsat.classifiers.bayesian

Class ConditionalProbabilityTable

java.lang.Object

jsat.classifiers.bayesian.ConditionalProbabilityTable

All Implemented Interfaces:

Serializable, Cloneable, Classifier

public class ConditionalProbabilityTable
extends Object
implements Classifier

The conditional probability table (CPT) is a classifier for categorical attributes. It builds the whole conditional probability table for a data set. The size of the CPT grows exponentially with the number of dimensions and options, and requires exponentially more data to get a good fit. CPTs can be useful for small data sets, or as a building block for another algorithm

Author:

Edward Raff

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
ConditionalProbabilityTable()

Method Summary

Modifier and Type	Method and Description
CategoricalResults	classify(DataPoint data) Performs classification on the given data point.
Classifier	clone()
int	dataPointToCord(DataPointPair<Integer> dataPoint, int targetClass, int[] cord) Converts a data point pair into a coordinate.
int	getDimensionSize() Returns the number of dimensions in the CPT
double	query(int targetClass, DataPointPair<Integer> dataPoint) Queries the CPT for the probability that the class value of targetClas would occur with the given DataPointPair.
double	query(int targetClass, int targetValue, int[] cord) Queries the CPT for the probability of the target class occurring with the specified value given the class values of the other attributes
boolean	supportsWeightedData() Indicates whether the model knows how to train using weighted data points.
void	trainC(ClassificationDataSet dataSet) Trains the classifier and constructs a model for classification using the given data set.
void	trainC(ClassificationDataSet dataSet, ExecutorService threadPool) Trains the classifier and constructs a model for classification using the given data set.
void	trainC(ClassificationDataSet dataSet, Set<Integer> categoriesToUse) Creates a CPT using only a subset of the features specified by categoriesToUse.

Methods inherited from class java.lang.Object

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

Constructor Detail

ConditionalProbabilityTable

public ConditionalProbabilityTable()

Method Detail

classify

public CategoricalResults classify(DataPoint data)

Description copied from interface: Classifier

Performs classification on the given data point.

Specified by:

classify in interface Classifier

Parameters:

data - the data point to classify

Returns:

the results of the classification.

getDimensionSize

public int getDimensionSize()

Returns the number of dimensions in the CPT

Returns:

the number of dimensions in the CPT

dataPointToCord

public int dataPointToCord(DataPointPair<Integer> dataPoint,
                           int targetClass,
                           int[] cord)

Converts a data point pair into a coordinate. The paired value contains the value for the predicting index. Though this value will not be used if the predicting class of the original data set was not used to make the table.

Parameters:

dataPoint - the DataPointPair to convert

targetClass - the index in the original data set of the category that we would like to predict

cord - the array to store the coordinate in.

Returns:

the value of the target class for the given data point

Throws:

ArithmeticException - if the cord array does not match the dimension of the CPT

trainC

public void trainC(ClassificationDataSet dataSet,
                   ExecutorService threadPool)

Description copied from interface: Classifier

Trains the classifier and constructs a model for classification using the given data set. If the training method knows how, it will used the threadPool to conduct training in parallel. This method will block until the training has completed.

Specified by:

trainC in interface Classifier

Parameters:

dataSet - the data set to train on

threadPool - the source of threads to use.

trainC

public void trainC(ClassificationDataSet dataSet)

Description copied from interface: Classifier

Trains the classifier and constructs a model for classification using the given data set.

Specified by:

trainC in interface Classifier

Parameters:

dataSet - the data set to train on

trainC

public void trainC(ClassificationDataSet dataSet,
                   Set<Integer> categoriesToUse)

Creates a CPT using only a subset of the features specified by categoriesToUse.

Parameters:

dataSet - the data set to train from

categoriesToUse - the attributes to use in training. Each value corresponds to the categorical index in dataSet, and adding the value DataSet.getNumCategoricalVars(), which is not a valid index, indicates to used the predicting class of the data set in the CPT.

query

public double query(int targetClass,
                    DataPointPair<Integer> dataPoint)

Queries the CPT for the probability that the class value of targetClas would occur with the given DataPointPair.

Parameters:

targetClass - the index in the original data set of the class that we want the probability of

dataPoint - the data point of values paired with the value of the predicting attribute in the original training set

Returns:

the probability in [0,1] of the target class value occurring with the given DataPointPair

query

public double query(int targetClass,
                    int targetValue,
                    int[] cord)

Queries the CPT for the probability of the target class occurring with the specified value given the class values of the other attributes

Parameters:

targetClass - the index in the original data set of the class that we want to probability of

targetValue - the value of the targetClass that we want to probability of occurring

cord - the coordinate array that corresponds the the class values for the CPT, where the coordinate of the targetClass may contain any value.

Returns:

the probability in [0, 1] of the targetClass occurring with the targetValue given the information in cord

See Also:

dataPointToCord(jsat.classifiers.DataPointPair, int, int[])

supportsWeightedData

public boolean supportsWeightedData()

Description copied from interface: Classifier

Indicates whether the model knows how to train using weighted data points. If it does, the model will train assuming the weights. The values returned by this method may change depending on the parameters set for the model.

Specified by:

supportsWeightedData in interface Classifier

Returns:

true if the model supports weighted data, false otherwise

clone

public Classifier clone()

Specified by:

clone in interface Classifier

Overrides:

clone in class Object