meka.classifiers.multilabel

Class Evaluation

java.lang.Object

meka.classifiers.multilabel.Evaluation

public class Evaluation
extends java.lang.Object

Evaluation.java - Evaluation functionality.

Version:

March 2014

Author:

Jesse Read

Constructor Summary

Constructors

Constructor and Description
Evaluation()

Method Summary

Methods

Modifier and Type	Method and Description
static Result[]	cvModel(MultilabelClassifier h, weka.core.Instances D, int numFolds, java.lang.String top) CVModel - Split D into train/test folds, and then train and evaluate on each one.
static Result[]	cvModel(MultilabelClassifier h, weka.core.Instances D, int numFolds, java.lang.String top, java.lang.String vop) CVModel - Split D into train/test folds, and then train and evaluate on each one.
static Result	evaluateModel(MultilabelClassifier h, weka.core.Instances D_train, weka.core.Instances D_test) EvaluateModel - Build model 'h' on 'D_train', test it on 'D_test'.
static Result	evaluateModel(MultilabelClassifier h, weka.core.Instances D_train, weka.core.Instances D_test, java.lang.String top) EvaluateModel - Build model 'h' on 'D_train', test it on 'D_test', threshold it according to 'top', using default verbosity option.
static Result	evaluateModel(MultilabelClassifier h, weka.core.Instances D_train, weka.core.Instances D_test, java.lang.String top, java.lang.String vop) EvaluateModel - Build model 'h' on 'D_train', test it on 'D_test', threshold it according to 'top', verbosity 'vop'.
static Result	evaluateModel(MultilabelClassifier h, weka.core.Instances D_test, java.lang.String tal, java.lang.String vop) EvaluateModel - Assume 'h' is already built, test it on 'D_test', threshold it according to 'top', verbosity 'vop'.
static boolean	isMT(weka.core.Instances D) IsMT - see if dataset D is multi-target (else only multi-label)
static weka.core.Instances	loadDataset(java.lang.String[] options) loadDataset - load a dataset, given command line option '-t' specifying an arff file.
static weka.core.Instances	loadDataset(java.lang.String[] options, char T) loadDataset - load a dataset, given command line options specifying an arff file.
static void	printOptions(java.util.Enumeration e)
static void	runExperiment(MultilabelClassifier h, java.lang.String[] options) RunExperiment - Build and evaluate a model with command-line options.
static Result	testClassifier(MultilabelClassifier h, weka.core.Instances D_test) TestClassifier - test classifier h on D_test

Methods inherited from class java.lang.Object

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

Constructor Detail

Evaluation

public Evaluation()

Method Detail

runExperiment

public static void runExperiment(MultilabelClassifier h,
                 java.lang.String[] options)
                          throws java.lang.Exception

RunExperiment - Build and evaluate a model with command-line options.

Parameters:

h - multi-label classifier

options - command line options

Throws:

java.lang.Exception

isMT

public static boolean isMT(weka.core.Instances D)

IsMT - see if dataset D is multi-target (else only multi-label)

Parameters:

D - data

Returns:

true iff D is multi-target only (else false)

evaluateModel

public static Result evaluateModel(MultilabelClassifier h,
                   weka.core.Instances D_train,
                   weka.core.Instances D_test,
                   java.lang.String top)
                            throws java.lang.Exception

EvaluateModel - Build model 'h' on 'D_train', test it on 'D_test', threshold it according to 'top', using default verbosity option.

Parameters:

h - a multi-dim. classifier

D_train - training data

D_test - test data

top - Threshold OPtion (pertains to multi-label data only)

Returns:

Result raw prediction data with evaluation statistics included.

Throws:

java.lang.Exception

evaluateModel

public static Result evaluateModel(MultilabelClassifier h,
                   weka.core.Instances D_train,
                   weka.core.Instances D_test,
                   java.lang.String top,
                   java.lang.String vop)
                            throws java.lang.Exception

EvaluateModel - Build model 'h' on 'D_train', test it on 'D_test', threshold it according to 'top', verbosity 'vop'.

Parameters:

h - a multi-dim. classifier

D_train - training data

D_test - test data

top - Threshold OPtion (pertains to multi-label data only)

vop - Verbosity OPtion (which measures do we want to calculate/output)

Returns:

Result raw prediction data with evaluation statistics included.

Throws:

java.lang.Exception

evaluateModel

public static Result evaluateModel(MultilabelClassifier h,
                   weka.core.Instances D_test,
                   java.lang.String tal,
                   java.lang.String vop)
                            throws java.lang.Exception

EvaluateModel - Assume 'h' is already built, test it on 'D_test', threshold it according to 'top', verbosity 'vop'.

Parameters:

h - a multi-dim. classifier

D_test - test data

tal - Threshold VALUES (not option)

vop - Verbosity OPtion (which measures do we want to calculate/output)

Returns:

Result raw prediction data with evaluation statistics included.

Throws:

java.lang.Exception

cvModel

public static Result[] cvModel(MultilabelClassifier h,
               weka.core.Instances D,
               int numFolds,
               java.lang.String top)
                        throws java.lang.Exception

CVModel - Split D into train/test folds, and then train and evaluate on each one.

Parameters:

h - a multi-dim. classifier

D - data

numFolds - test data

top - Threshold OPtion (pertains to multi-label data only)

Returns:

an array of 'numFolds' Results

Throws:

java.lang.Exception

cvModel

public static Result[] cvModel(MultilabelClassifier h,
               weka.core.Instances D,
               int numFolds,
               java.lang.String top,
               java.lang.String vop)
                        throws java.lang.Exception

CVModel - Split D into train/test folds, and then train and evaluate on each one.

Parameters:

h - a multi-dim. classifier

D - data

numFolds - test data

top - Threshold OPtion (pertains to multi-label data only)

vop - Verbosity OPtion (which measures do we want to calculate/output)

Returns:

an array of 'numFolds' Results

Throws:

java.lang.Exception

evaluateModel

public static Result evaluateModel(MultilabelClassifier h,
                   weka.core.Instances D_train,
                   weka.core.Instances D_test)
                            throws java.lang.Exception

EvaluateModel - Build model 'h' on 'D_train', test it on 'D_test'. Note that raw multi-label predictions returned in Result may not have been thresholded yet. However, data statistics, classifier info, and running times are inpregnated into the Result here.

Parameters:

h - a multi-dim. classifier

D_train - training data

D_test - test data

Returns:

raw prediction data (no evaluation yet)

Throws:

java.lang.Exception

testClassifier

public static Result testClassifier(MultilabelClassifier h,
                    weka.core.Instances D_test)
                             throws java.lang.Exception

TestClassifier - test classifier h on D_test

Parameters:

h - a multi-dim. classifier, ALREADY BUILT

D_test - test data

Returns:

Result with raw prediction data ONLY

Throws:

java.lang.Exception

loadDataset

public static weka.core.Instances loadDataset(java.lang.String[] options)
                                       throws java.lang.Exception

loadDataset - load a dataset, given command line option '-t' specifying an arff file.

Parameters:

options - command line options, specifying dataset filename

Returns:

the dataset

Throws:

java.lang.Exception

loadDataset

public static weka.core.Instances loadDataset(java.lang.String[] options,
                              char T)
                                       throws java.lang.Exception

loadDataset - load a dataset, given command line options specifying an arff file.

Parameters:

options - command line options, specifying dataset filename

T - set to 'T' if we want to load a test file (default 't': load train or train-test file)

Returns:

the dataset

Throws:

java.lang.Exception

printOptions

public static void printOptions(java.util.Enumeration e)