com.ibm.bluej.util.common

Class CollectionUtil

java.lang.Object

com.ibm.bluej.util.common.CollectionUtil

public class CollectionUtil
extends java.lang.Object

Constructor Summary

Constructors

Constructor and Description
CollectionUtil()

Method Summary

Methods

Modifier and Type	Method and Description
static <T> java.util.List<Pair<java.util.Collection<T>,java.util.Collection<T>>>	partitionToCrossvalidatedTrainTestDivisions(java.util.Collection<T> seq, int numDivisions) This helps you convert a collection of items into a series of train/test divisions for the purpose of running cross-validation or jack-knifing.
static <T> java.util.List<java.util.Collection<T>>	partitionToFolds(java.util.Collection<T> seq, int numFolds) Given a sequence, partitions it into roughly numFolds equal sized folds.

Methods inherited from class java.lang.Object

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

Constructor Detail

CollectionUtil

public CollectionUtil()

Method Detail

partitionToFolds

public static <T> java.util.List<java.util.Collection<T>> partitionToFolds(java.util.Collection<T> seq,
                                                           int numFolds)

Given a sequence, partitions it into roughly numFolds equal sized folds. The last fold will contain fewer items than the others if the number of items do not divide evenly.

partitionToCrossvalidatedTrainTestDivisions

public static <T> java.util.List<Pair<java.util.Collection<T>,java.util.Collection<T>>> partitionToCrossvalidatedTrainTestDivisions(java.util.Collection<T> seq,
                                                                                                                    int numDivisions)

This helps you convert a collection of items into a series of train/test divisions for the purpose of running cross-validation or jack-knifing. First, this splits the data into numDivisions folds. Given a sequence of items, produces a list of divisions. Each division contains a pair of collections which represent the training and testing set. Each fold of the data will be used as the test set in only one division.