weka.tools.data

Class InstancesOperator

java.lang.Object

weka.tools.data.InstancesOperator

public class InstancesOperator
extends Object

Author:

pawel

Constructor Summary

Constructors

Constructor and Description
InstancesOperator()

Method Summary

Modifier and Type	Method and Description
static double[]	classFreq(weka.core.Instances inputSet) Calculates the frequency of class-specific instances.
static weka.core.Instances[]	classSpecSplit(weka.core.Instances input) Split a dataset according to class indices.
static weka.core.Instances	desparse(weka.core.Instances inputSet) Generates dense dataset from sparse dataset.
static weka.core.Instances	generateBaggingSample(weka.core.Instances baseSet) generates bootstrap sample from base set
static weka.core.Instances	generateBaggingSample(weka.core.Instances baseSet, int seed) generates bootstrap sample from base set
static weka.core.Instances	generateBaggingSample(weka.core.Instances baseSet, int seed, double percentage) generates bootstrap sample from base set
static void	main(String[] args)
static weka.core.Instances	normalizeData(weka.core.Instances inputSet) Normalizes data to interval [0,1] for each attribute
static int[]	objPerClass(weka.core.Instances inputSet) Counts the number of objects per class
static weka.core.Instances[]	splitSet(weka.core.Instances inputSet, double splitCoeff) Splits input set into two sets according to splitCoeff First set crdinality -> numOfInstances*splitCoeff Second set crdinality -> numOfInstances*(1-splitCoeff)
static weka.core.Instances[]	splitSet(weka.core.Instances inputSet, double splitCoeff, int seed) Splits input set into two sets according to splitCoeff First set crdinality -> numOfInstances*splitCoeff Second set crdinality -> numOfInstances*(1-splitCoeff)
static weka.core.Instances[]	splItSet(weka.core.Instances inputSet, int firstSetNum) Splits input set into two sets according to firstSetNum First set -> first firstSetNum instances Second set -> remaining instances Does not shuffle
static weka.core.Instances	standarizeData(weka.core.Instances inputSet) Standarize data for 0 mean and unit variance
static weka.core.Instances[]	stratifiedSplitSet(weka.core.Instances inputSet, double splitCoeff, int seed) Splits input set into two sets according to splitCoeff First set crdinality -> numOfInstances*splitCoeff Second set crdinality -> numOfInstances*(1-splitCoeff) Stratified splitt

Methods inherited from class java.lang.Object

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

Constructor Detail

InstancesOperator

public InstancesOperator()

Method Detail

generateBaggingSample

public static weka.core.Instances generateBaggingSample(weka.core.Instances baseSet,
                                                        int seed,
                                                        double percentage)
                                                 throws Exception

generates bootstrap sample from base set

Parameters:

baseSet - - input set

seed - - random generator seed

percentage - - percentage of instances to return

Returns:

Throws:

Exception

generateBaggingSample

public static weka.core.Instances generateBaggingSample(weka.core.Instances baseSet,
                                                        int seed)
                                                 throws Exception

generates bootstrap sample from base set

Parameters:

baseSet -

seed - - random generator seed

Returns:

resampled set

Throws:

Exception

generateBaggingSample

public static weka.core.Instances generateBaggingSample(weka.core.Instances baseSet)
                                                 throws Exception

generates bootstrap sample from base set

Parameters:

baseSet -

Returns:

resapmled set

Throws:

Exception

splitSet

public static weka.core.Instances[] splitSet(weka.core.Instances inputSet,
                                             double splitCoeff,
                                             int seed)

Splits input set into two sets according to splitCoeff First set crdinality -> numOfInstances*splitCoeff Second set crdinality -> numOfInstances*(1-splitCoeff)

Parameters:

inputSet -

splitCoeff -

Returns:

Instances table - two splitted set

stratifiedSplitSet

public static weka.core.Instances[] stratifiedSplitSet(weka.core.Instances inputSet,
                                                       double splitCoeff,
                                                       int seed)

Splits input set into two sets according to splitCoeff First set crdinality -> numOfInstances*splitCoeff Second set crdinality -> numOfInstances*(1-splitCoeff) Stratified splitt

Parameters:

inputSet -

splitCoeff -

seed -

Returns:

splitSet

public static weka.core.Instances[] splitSet(weka.core.Instances inputSet,
                                             double splitCoeff)

Splits input set into two sets according to splitCoeff First set crdinality -> numOfInstances*splitCoeff Second set crdinality -> numOfInstances*(1-splitCoeff)

Parameters:

inputSet -

splitCoeff -

Returns:

Instances table - two splitted set

splItSet

public static weka.core.Instances[] splItSet(weka.core.Instances inputSet,
                                             int firstSetNum)

Splits input set into two sets according to firstSetNum First set -> first firstSetNum instances Second set -> remaining instances Does not shuffle

Parameters:

inputSet -

firstSetNum -

Returns:

Instances[] - two sets

normalizeData

public static weka.core.Instances normalizeData(weka.core.Instances inputSet)
                                         throws Exception

Normalizes data to interval [0,1] for each attribute

Parameters:

inputSet -

Returns:

Throws:

Exception

standarizeData

public static weka.core.Instances standarizeData(weka.core.Instances inputSet)
                                          throws Exception

Standarize data for 0 mean and unit variance

Parameters:

inputSet -

Returns:

Throws:

Exception

desparse

public static weka.core.Instances desparse(weka.core.Instances inputSet)
                                    throws Exception

Generates dense dataset from sparse dataset.

Parameters:

inputSet -

Returns:

Throws:

Exception

objPerClass

public static int[] objPerClass(weka.core.Instances inputSet)
                         throws Exception

Counts the number of objects per class

Parameters:

inputSet - -- input data

Returns:

-- array of class specific instance counts

Throws:

Exception

classFreq

public static double[] classFreq(weka.core.Instances inputSet)
                          throws Exception

Calculates the frequency of class-specific instances. The resulting array sums up to one.

Parameters:

inputSet -

Returns:

Throws:

Exception

classSpecSplit

public static weka.core.Instances[] classSpecSplit(weka.core.Instances input)
                                            throws Exception

Split a dataset according to class indices. Accepts only nominal and string class attributes.

Parameters:

input - -- input dataset

Returns:

-- an array of splitted instances.

Throws:

Exception - when a dataset with a wrong attribute class has been passed.

main

public static void main(String[] args)