jsat.classifiers

Class ClassificationDataSet

java.lang.Object

jsat.DataSet<ClassificationDataSet>

jsat.classifiers.ClassificationDataSet

public class ClassificationDataSet
extends DataSet<ClassificationDataSet>

ClassificationDataSet is a data set meant specifically for classification problems. The true class of each data point is stored separately from the data point, so that it can be feed into a learning algorithm and not interfere.
Additional functionality specific to classification problems is also available.

Constructor Summary

Constructors

Constructor and Description
ClassificationDataSet(DataSet dataSet, int predicting) Creates a new data set for classification problems.
ClassificationDataSet(int numerical, CategoricalData[] categories, CategoricalData predicting) Creates a new, empty, data set for classification problems.
ClassificationDataSet(java.util.List<DataPoint> data, int predicting) Creates a new data set for classification problems from the given list of data points.
ClassificationDataSet(java.util.List<DataPointPair<java.lang.Integer>> data, CategoricalData predicting) Creates a new data set for classification problems from the given list of data points.

Method Summary

Modifier and Type	Method and Description
void	addDataPoint(DataPoint dp, int classification) Creates a new data point and add it
void	addDataPoint(Vec v, int classification) Creates a new data point with no categorical variables and adds it to this data set.
void	addDataPoint(Vec v, int[] classes, int classification) Creates a new data point and adds it to this data set.
void	addDataPoint(Vec v, int[] classes, int classification, double weight) Creates a new data point and add its to this data set.
void	addDataPoint(Vec v, int classification, double weight) Creates a new data point with no categorical variables and adds it to this data set.
int	classSampleCount(int targetClass) Returns the number of data points that belong to the specified class, irrespective of the weights of the individual points.
static ClassificationDataSet	comineAllBut(java.util.List<ClassificationDataSet> list, int exception) A helper method meant to be used with DataSet.cvSet(int), this combines all classification data sets in a given list, but holding out the indicated list.
java.util.List<DataPointPair<java.lang.Integer>>	getAsDPPList() Returns the data set as a list of DataPointPair.
java.util.List<DataPointPair<java.lang.Double>>	getAsFloatDPPList() Returns the data set as a list of DataPointPair.
int	getClassSize() Returns the number of target classes in this classification data set.
DataPoint	getDataPoint(int i) Returns the i'th data point from the data set
int	getDataPointCategory(int i) Returns the integer value corresponding to the true category of the i'th data point.
DataPointPair<java.lang.Integer>	getDataPointPair(int i) Returns the i'th data point from the data set, paired with the integer indicating its true class
CategoricalData	getPredicting()
double[]	getPriors() Computes the prior probabilities of each class, and returns an array containing the values.
java.util.List<DataPoint>	getSamples(int category) Returns the list of all examples that belong to the given category.
int	getSampleSize() Returns the number of data points in this data set
Vec	getSampleVariableVector(int category, int n) This method is a counter part to DataSet.getNumericColumn(int).
ClassificationDataSet	getTwiceShallowClone() Returns a new version of this data set that is of the same type, and contains a different listing pointing to shallow data point copies.
void	setDataPoint(int i, DataPoint dp) Replaces an already existing data point with the one given.
ClassificationDataSet	shallowClone() Returns a new version of this data set that is of the same type, and contains a different list pointing to the same data points.
java.util.List<ClassificationDataSet>	stratSet(int folds, java.util.Random rnd)

Methods inherited from class jsat.DataSet

applyTransform, applyTransform, applyTransform, applyTransform, countMissingValues, cvSet, cvSet, getCategories, getCategoryName, getColumnMeanVariance, getDataMatrix, getDataMatrixView, getDataPointIterator, getDataPoints, getDataVectors, getDataWeights, getMissingDropped, getNumCategoricalVars, getNumericColumn, getNumericColumns, getNumericColumns, getNumericName, getNumFeatures, getNumNumericalVars, getOnlineColumnStats, getOnlineDenseStats, getSparsityStats, randomSplit, randomSplit, replaceNumericFeatures, setNumericName

Methods inherited from class java.lang.Object

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

Constructor Detail

ClassificationDataSet

public ClassificationDataSet(DataSet dataSet,
                             int predicting)

Creates a new data set for classification problems.

Parameters:

dataSet - the source data set

predicting - the categorical attribute to use as the target class

ClassificationDataSet

public ClassificationDataSet(java.util.List<DataPoint> data,
                             int predicting)

Creates a new data set for classification problems from the given list of data points. It is assume the data points are consistent.

Parameters:

data - the list of data points for the problem.

predicting - the categorical attribute to use as the target class

ClassificationDataSet

public ClassificationDataSet(java.util.List<DataPointPair<java.lang.Integer>> data,
                             CategoricalData predicting)

Creates a new data set for classification problems from the given list of data points. The class value is paired with each data point.

Parameters:

data - the list of data points, paired with their class values

predicting - the information about the target class

ClassificationDataSet

public ClassificationDataSet(int numerical,
                             CategoricalData[] categories,
                             CategoricalData predicting)

Creates a new, empty, data set for classification problems.

Parameters:

numerical - the number of numerical attributes for the problem

categories - the information about each categorical variable in the problem.

predicting - the information about the target class

Method Detail

getClassSize

public int getClassSize()

Returns the number of target classes in this classification data set. This value can also be obtained by calling getPredicting(). getNumOfCategories()

Returns:

the number of target classes for prediction

comineAllBut

public static ClassificationDataSet comineAllBut(java.util.List<ClassificationDataSet> list,
                                                 int exception)

A helper method meant to be used with DataSet.cvSet(int), this combines all classification data sets in a given list, but holding out the indicated list.

Parameters:

list - a list of data sets

exception - the one data set in the list NOT to combine into one file

Returns:

a combination of all the data sets in list except the one at index exception

getDataPoint

public DataPoint getDataPoint(int i)

Returns the i'th data point from the data set

Specified by:

getDataPoint in class DataSet<ClassificationDataSet>

Parameters:

i - the i'th data point in this set

Returns:

the ith data point in this set

getDataPointPair

public DataPointPair<java.lang.Integer> getDataPointPair(int i)

Returns the i'th data point from the data set, paired with the integer indicating its true class

Parameters:

i - the i'th data point in this set

Returns:

the i'th data point from the data set, paired with the integer indicating its true class

setDataPoint

public void setDataPoint(int i,
                         DataPoint dp)

Description copied from class: DataSet

Replaces an already existing data point with the one given. Any values associated with the data point, but not apart of it, will remain intact.

Specified by:

setDataPoint in class DataSet<ClassificationDataSet>

Parameters:

i - the i'th dataPoint to set.

dp - the data point to set at the specified index

getDataPointCategory

public int getDataPointCategory(int i)

Returns the integer value corresponding to the true category of the i'th data point.

Parameters:

i - the i'th data point.

Returns:

the integer value for the category of the i'th data point.

Throws:

java.lang.IndexOutOfBoundsException - if i is not a valid index into the data set.

stratSet

public java.util.List<ClassificationDataSet> stratSet(int folds,
                                                      java.util.Random rnd)

addDataPoint

public void addDataPoint(Vec v,
                         int[] classes,
                         int classification)

Creates a new data point and adds it to this data set.

Parameters:

v - the numerical values for the data point

classes - the categorical values for the data point

classification - the true class value for the data point

Throws:

java.lang.IllegalArgumentException - if the given values are inconsistent with the data this class stores.

addDataPoint

public void addDataPoint(Vec v,
                         int classification)

Creates a new data point with no categorical variables and adds it to this data set.

Parameters:

v - the numerical values for the data point

classification - the true class value for the data point

Throws:

java.lang.IllegalArgumentException - if the given values are inconsistent with the data this class stores.

addDataPoint

public void addDataPoint(Vec v,
                         int classification,
                         double weight)

Creates a new data point with no categorical variables and adds it to this data set.

Parameters:

v - the numerical values for the data point

classification - the true class value for the data point

weight - the weight value to give to the data point

Throws:

java.lang.IllegalArgumentException - if the given values are inconsistent with the data this class stores.

addDataPoint

public void addDataPoint(Vec v,
                         int[] classes,
                         int classification,
                         double weight)

Creates a new data point and add its to this data set.

Parameters:

v - the numerical values for the data point

classes - the categorical values for the data point

classification - the true class value for the data point

weight - the weight value to give to the data point

Throws:

java.lang.IllegalArgumentException - if the given values are inconsistent with the data this class stores.

addDataPoint

public void addDataPoint(DataPoint dp,
                         int classification)

Creates a new data point and add it

Parameters:

dp - the data point to add to this set

classification - the label for this data point

getSamples

public java.util.List<DataPoint> getSamples(int category)

Returns the list of all examples that belong to the given category.

Parameters:

category - the category desired

Returns:

all given examples that belong to the given category

getSampleVariableVector

public Vec getSampleVariableVector(int category,
                                   int n)

This method is a counter part to DataSet.getNumericColumn(int). Instead of returning all values for a given attribute, all values for the attribute that are members of a specific class are returned.

Parameters:

category - the category desired

n - the n'th numerical variable

Returns:

a vector of all the values for the n'th numerical variable for the given category

getPredicting

public CategoricalData getPredicting()

Returns:

the CategoricalData object for the variable that is to be predicted

getAsDPPList

public java.util.List<DataPointPair<java.lang.Integer>> getAsDPPList()

Returns the data set as a list of DataPointPair. Each data point is paired with it's true class value. Altering the data points will effect the data set. Altering the list will not.
The list of data points will come in the same order they would be retrieved in using getDataPoint(int)

Returns:

a list of each data point paired with its class value

getAsFloatDPPList

public java.util.List<DataPointPair<java.lang.Double>> getAsFloatDPPList()

Returns the data set as a list of DataPointPair. Each data point is paired with it's true class value, which is stored in a double. Altering the data points will effect the data set. Altering the list will not.
The list of data points will come in the same order they would be retrieved in using getDataPoint(int)

Returns:

a list of each data point paired with its class value stored in a double

getPriors

public double[] getPriors()

Computes the prior probabilities of each class, and returns an array containing the values.

Returns:

the array of prior probabilities

classSampleCount

public int classSampleCount(int targetClass)

Returns the number of data points that belong to the specified class, irrespective of the weights of the individual points.

Parameters:

targetClass - the target class

Returns:

how many data points belong to the given class

getSampleSize

public int getSampleSize()

Description copied from class: DataSet

Returns the number of data points in this data set

Specified by:

getSampleSize in class DataSet<ClassificationDataSet>

Returns:

the number of data points in this data set

shallowClone

public ClassificationDataSet shallowClone()

Description copied from class: DataSet

Returns a new version of this data set that is of the same type, and contains a different list pointing to the same data points.

Specified by:

shallowClone in class DataSet<ClassificationDataSet>

Returns:

a shallow copy of this data set

getTwiceShallowClone

public ClassificationDataSet getTwiceShallowClone()

Description copied from class: DataSet

Returns a new version of this data set that is of the same type, and contains a different listing pointing to shallow data point copies. Because the data point object contains the weight itself, the weight is not shared - while the vector and array information is. This allows altering the weights of the data points while preserving the original weights.
Altering the list or weights of the returned data set will not be reflected in the original. Altering the feature values will.

Overrides:

getTwiceShallowClone in class DataSet<ClassificationDataSet>

Returns:

a shallow copy of shallow data point copies for this data set.