com.imsl.datamining.supportvectormachine

Class SVClassification

java.lang.Object

com.imsl.datamining.PredictiveModel

com.imsl.datamining.supportvectormachine.SupportVectorMachine

com.imsl.datamining.supportvectormachine.SVClassification

All Implemented Interfaces:

Serializable, Cloneable

public class SVClassification
extends SupportVectorMachine
implements Serializable, Cloneable

Specifies a support vector machine for classification (SVC).

The C-support vector classification (C-SVC) is the fundamental algorithm for the SVM optimization problem and its primal form is given as

$$\min_ {w, b, \xi} \frac{1}{2} w^Tw+C\sum_{i=1}^{l} \xi _i$$

$$\text{subject to} \, \, y_i(w^T \phi (x_i)+b) \geq 1- \xi _i,$$

$$ \xi_i\ge 0,i=1,\;\ldots,\;l $$

The set \(\{ (x_i,y_i): i=1,\ldots,l \}\) is the set of instance-label pairs in the training data, with \(x_i \in R^n \), \(y_i \in \{1,-1\}\), and \(l\) equal to the number of training examples. The \(\xi _i\) are the slack variables in the optimization and represent an upper bound on the number of errors. The regularization parameter \(C \, > \,0 \) acts as a tradeoff parameter between error and margin. This is the default algorithm for the classification problem.

See Also:

Example 1, Example 2, Serialized Form

Nested Class Summary

Nested classes/interfaces inherited from class com.imsl.datamining.supportvectormachine.SupportVectorMachine

SupportVectorMachine.ReflectiveOperationException

Nested classes/interfaces inherited from class com.imsl.datamining.PredictiveModel

PredictiveModel.CloneNotSupportedException, PredictiveModel.PredictiveModelException, PredictiveModel.StateChangeException, PredictiveModel.SumOfProbabilitiesNotOneException, PredictiveModel.VariableType

Constructor Summary

Constructors

Constructor and Description
SVClassification(double[][] xy, int responseColumnIndex, PredictiveModel.VariableType[] varType) Constructs a support vector machine for classification (SVC).
SVClassification(double[][] xy, int responseColumnIndex, PredictiveModel.VariableType[] varType, Kernel k) Constructs a support vector machine for classification (SVC).
SVClassification(SVClassification classModel) Copy constructor.

Method Summary

Modifier and Type	Method and Description
SVClassification	clone() Clones an SVClassification predictive model.
double[]	getClassPenaltyWeights() Returns the class weights.
int[]	getClassWeightLabels() Returns the weight labels array.
protected SVModel	optimize(DataNode[][] x, double[] y, double[] w, int len, Kernel kernel) Performs the classification support vector machine optimization problem.
protected double[]	predictValues(SVModel model, double[][] attributeData) Generates the predicted values on the attribute data using the given support vector machine model.
void	setClassPenaltyWeights(int[] weightLabels, double[] weights) Sets the class penalty weights.
protected void	setConfiguration(PredictiveModel pm) Sets the configuration to that of the input PredictiveModel.

Methods inherited from class com.imsl.datamining.supportvectormachine.SupportVectorMachine

fitModel, getConvergenceTolerance, getKernel, getKernelParameters, getModel, getNuParameter, getRegularizationParameter, getWorkingArraySize, isNuFormulation, isProbability, isShrinking, predict, predict, setConvergenceTolerance, setKernel, setKernelParameters, setNuFormulation, setNuParameter, setProbability, setRegularizationParameter, setShrinking, setWorkArraySize

Methods inherited from class com.imsl.datamining.PredictiveModel

getClassCounts, getClassErrors, getClassErrors, getClassLabels, getClassProbabilities, getCostMatrix, getMaxNumberOfCategories, getMaxNumberOfIterations, getNumberOfClasses, getNumberOfColumns, getNumberOfMissing, getNumberOfPredictors, getNumberOfRows, getNumberOfUniquePredictorValues, getPredictorIndexes, getPredictorTypes, getPrintLevel, getPriorProbabilities, getRandomObject, getResponseColumnIndex, getResponseVariableAverage, getResponseVariableMostFrequentClass, getResponseVariableType, getTotalWeight, getVariableType, getWeights, getXY, isConstantSeries, isMustFitModel, isUserFixedNClasses, predict, setClassCounts, setClassLabels, setClassProbabilities, setCostMatrix, setMaxNumberOfCategories, setMaxNumberOfIterations, setMustFitModel, setNumberOfClasses, setPredictorIndex, setPredictorTypes, setPrintLevel, setPriorProbabilities, setRandomObject, setResponseColumnIndex, setTrainingData, setVariableType, setWeights

Methods inherited from class java.lang.Object

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

Constructor Detail

SVClassification

public SVClassification(double[][] xy,
                        int responseColumnIndex,
                        PredictiveModel.VariableType[] varType)

Constructs a support vector machine for classification (SVC).

Parameters:

xy - a double matrix containing the training data and associated response values

responseColumnIndex - an int, the column index of the response variable

varType - a PredictiveModel.VariableType array of length equal to xy[0].length containing the type of each variable

SVClassification

public SVClassification(double[][] xy,
                        int responseColumnIndex,
                        PredictiveModel.VariableType[] varType,
                        Kernel k)

Constructs a support vector machine for classification (SVC).

Parameters:

xy - a double matrix containing the training data and associated response values

responseColumnIndex - an int, the column index of the response variable

varType - a PredictiveModel.VariableType array of length equal to xy[0].length containing the type of each variable

k - a Kernel, the kernel function

SVClassification

public SVClassification(SVClassification classModel)

Copy constructor.

Parameters:

classModel -

Method Detail

clone

public SVClassification clone()

Clones an SVClassification predictive model.

Specified by:

clone in class PredictiveModel

Returns:

a clone of the SVClassification predictive model

setClassPenaltyWeights

public void setClassPenaltyWeights(int[] weightLabels,
                                   double[] weights)

Sets the class penalty weights.

Parameters:

weightLabels - an int array containing the class labels to which the weights should be applied

weights - a double array of length weightLabels.length containing the corresponding weights. The weight values must be non-negative.

Default: By default, the weight value is 1.0, leading to a penalty value of C for each class.

getClassPenaltyWeights

public double[] getClassPenaltyWeights()

Returns the class weights.

Returns:

a double array containing the class weights

setConfiguration

protected final void setConfiguration(PredictiveModel pm)

Sets the configuration to that of the input PredictiveModel.

Note that the input PredictiveModel object must be the same subclass of PredictiveModel as is this instance.

Overrides:

setConfiguration in class SupportVectorMachine

Parameters:

pm - an SVClassification instance

Throws:

CloneNotSupportedException - a java.lang.CloneNotSupportedException has occurred. The original exception has been added to the SupportVectorMachine.CloneNotSupportedException as a suppressed exception.

Default: The class uses its default configuration as described in the different methods.

getClassWeightLabels

public int[] getClassWeightLabels()

Returns the weight labels array.

Returns:

an int array containing the class labels corresponding to the penalty weights

optimize

protected SVModel optimize(DataNode[][] x,
                           double[] y,
                           double[] w,
                           int len,
                           Kernel kernel)
                    throws NoSuchMethodException,
                           InstantiationException,
                           IllegalAccessException,
                           InvocationTargetException

Performs the classification support vector machine optimization problem.

Specified by:

optimize in class SupportVectorMachine

Parameters:

x - a DataNode matrix containing the attribute data

y - a double array containing the response variable

w - a double array containing the observation weights

len - an int, the total possible number of support vectors

kernel - a Kernel object

Returns:

an SVModel structure containing the fitted model

Throws:

NoSuchMethodException - thrown when a particular method cannot be found

InstantiationException - thrown when an application tries to create an instance of a class using the newInstance method in class Class, but the specified class object cannot be instantiated

IllegalAccessException - thrown when an application tries to reflectively create an instance (other than an array), set or get a field, or invoke a method, but the currently executing method does not have access to the definition of the specified class, field, method or constructor

InvocationTargetException - a checked exception that wraps an exception thrown by an invoked method or constructor

predictValues

protected double[] predictValues(SVModel model,
                                 double[][] attributeData)
                          throws PredictiveModel.SumOfProbabilitiesNotOneException

Generates the predicted values on the attribute data using the given support vector machine model.

Specified by:

predictValues in class SupportVectorMachine

Parameters:

model - a fitted SVModel object

attributeData - a double matrix containing the attribute (or predictor) data

Returns:

a double array containing the predictions for each row in the input attribute data

Throws:

PredictiveModel.SumOfProbabilitiesNotOneException - the sum of probabilities is not approximately one