edu.cmu.minorthird.classify
Class SampleDatasets

java.lang.Object
  edu.cmu.minorthird.classify.SampleDatasets

public class SampleDatasets
extends java.lang.Object

Some sample inputs for learners.

Author:

William Cohen

Field Summary

static java.lang.String[]	negTest
static java.lang.String[]	negTrain
static java.lang.String[]	posTest
static java.lang.String[]	posTrain

Constructor Summary

SampleDatasets()

Method Summary

static void	main(java.lang.String[] args)
static Dataset	makeLogisticRegressionData(java.util.Random rand, int m, double a, double b) Data useful for testing univariate logistic regression.
static Dataset	makeNumericData(java.util.Random r, int dim, int m) Random data, defined by a simple boolean combination of thresholds over two dimensions, with up to 5 irrelevant dimensions, and m examples.
static Dataset	makeSparseNumericData(java.util.Random r, int m) Sparse numeric data - some values are 1.0, and some are zero.
static Dataset	makeToy3ClassData(java.util.Random random, int numInstances) Makes a sample 3 class dataset
static SequenceDataset	makeToySequenceData()
static SequenceDataset	makeToySequenceData(java.lang.String[] lines)
static SequenceDataset	makeToySequenceTestData()
static Dataset	sampleData(java.lang.String name, boolean isTest)
static Dataset	toyBayesExtremeTest()
static Dataset	toyBayesExtremeTrain()
static Dataset	toyBayesExtremeUnlabeledTrain()
static Dataset	toyBayesTest() Test data for a trivial classification problem.
static Dataset	toyBayesTrain() Training data for a trivial classification problem.
static Dataset	toyTest() Test data for a trivial classification problem.
static Dataset	toyTrain() Training data for a trivial classification problem.

Methods inherited from class java.lang.Object

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

Field Detail

posTrain

public static final java.lang.String[] posTrain

negTrain

public static final java.lang.String[] negTrain

posTest

public static final java.lang.String[] posTest

negTest

public static final java.lang.String[] negTest

Constructor Detail

SampleDatasets

public SampleDatasets()

Method Detail

toyTrain

public static Dataset toyTrain()

Training data for a trivial classification problem.

toyTest

public static Dataset toyTest()

Test data for a trivial classification problem.

toyBayesExtremeTrain

public static Dataset toyBayesExtremeTrain()

toyBayesExtremeTest

public static Dataset toyBayesExtremeTest()

toyBayesExtremeUnlabeledTrain

public static Dataset toyBayesExtremeUnlabeledTrain()

toyBayesTrain

public static Dataset toyBayesTrain()

Training data for a trivial classification problem.

toyBayesTest

public static Dataset toyBayesTest()

Test data for a trivial classification problem.

makeSparseNumericData

public static Dataset makeSparseNumericData(java.util.Random r,
                                            int m)

Sparse numeric data - some values are 1.0, and some are zero.

makeNumericData

public static Dataset makeNumericData(java.util.Random r,
                                      int dim,
                                      int m)

Random data, defined by a simple boolean combination of thresholds over two dimensions, with up to 5 irrelevant dimensions, and m examples.

makeLogisticRegressionData

public static Dataset makeLogisticRegressionData(java.util.Random rand,
                                                 int m,
                                                 double a,
                                                 double b)

Data useful for testing univariate logistic regression. The dataset will contain m examples, each with a single uniformly-distributed numeric feature x. The probability of the positive class will be chosen according to logistic(a*x + b).

makeToySequenceData

public static SequenceDataset makeToySequenceData()

makeToySequenceTestData

public static SequenceDataset makeToySequenceTestData()

makeToySequenceData

public static SequenceDataset makeToySequenceData(java.lang.String[] lines)

makeToy3ClassData

public static Dataset makeToy3ClassData(java.util.Random random,
                                        int numInstances)

Makes a sample 3 class dataset

Parameters:

random - A random number generator for building the dataset.

numInstances - The number of instances to be created.

sampleData

public static Dataset sampleData(java.lang.String name,
                                 boolean isTest)

main

public static void main(java.lang.String[] args)