de.tudarmstadt.ukp.dkpro.core.api.frequency.util

Class FrequencyDistribution<T>

java.lang.Object

de.tudarmstadt.ukp.dkpro.core.api.frequency.util.FrequencyDistribution<T>

Type Parameters:

T - the type of the samples

All Implemented Interfaces:

Serializable

public class FrequencyDistribution<T>
extends Object
implements Serializable

It is basically a mapping from samples (keys) to long values (counts).

Suppose we want to record the number of occurrences of each word in a sentence, then this class can be used as follows:

 FrequencyDistribution<String> fd = new FrequencyDistribution<String>();
 for (String word : "foo bar baz foo".split(" ")) {
     fd.inc(word);
 }
 System.out.println(fd.getCount("foo"));
 

The last call to getCount(T) will yield 2, because the word "foo" has appeared twice in the given sequence of words.

This class was inspired by NLTK's FreqDist.

See Also:

ConditionalFrequencyDistribution, Serialized Form

Constructor Summary

Constructors

Constructor and Description
FrequencyDistribution() Creates a new empty FrequencyDistribution.
FrequencyDistribution(Iterable<T> iterable) Creates a new FrequencyDistribution prefilled with samples from an Iterable.

Method Summary

Modifier and Type	Method and Description
void	addSample(T sample, long number) Increases the count for a given sample.
void	clear()
boolean	contains(T sample) Indicates whether this distribution contains outcomes for a given sample.
long	getB() Returns the total number of sample values (or bins) that have counts greater than zero.
long	getCount(T sample) Returns the count for a given sample.
Set<T>	getKeys() Returns the Set of sample values (or bins) for which counts have been recorded.
long	getMaxFreq() Returns the highest frequency that is currently stored.
List<T>	getMostFrequentSamples(int n) Returns the n most frequent samples in the distribution.
long	getN() Returns the total number of sample outcomes that have been recorded by this frequency distribution.
T	getSampleWithMaxFreq() Returns the sample which has currently the highest frequency.
void	inc(T sample) Increments the count for a given sample.
void	incAll(Iterable<T> iterable) Increments the count for each sample in a given Iterable.
void	load(File file)
void	save(File file)

Methods inherited from class java.lang.Object

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

Constructor Detail

FrequencyDistribution

public FrequencyDistribution()

Creates a new empty FrequencyDistribution.

FrequencyDistribution

public FrequencyDistribution(Iterable<T> iterable)

Creates a new FrequencyDistribution prefilled with samples from an Iterable. The count for each sample in the iterable is cumulatively increased by 1.

Parameters:

iterable - the Iterable used to fill the FrequencyDistribution

Method Detail

contains

public boolean contains(T sample)

Indicates whether this distribution contains outcomes for a given sample.

Parameters:

sample - the sample to look up

Returns:

true if samples exist

inc

public void inc(T sample)

Increments the count for a given sample.

Parameters:

sample - the sample to increment the count for

incAll

public void incAll(Iterable<T> iterable)

Increments the count for each sample in a given Iterable.

Parameters:

iterable - the samples used to increment the counts

getN

public long getN()

Returns the total number of sample outcomes that have been recorded by this frequency distribution. This is equal to the accumulated count of all samples (duplicates included).

Returns:

the total number of sample outcomes

getB

public long getB()

Returns the total number of sample values (or bins) that have counts greater than zero. This is equal to the accumulated counts of all distinct samples (duplicates excluded).

Returns:

the total number of bins

getCount

public long getCount(T sample)

Returns the count for a given sample. If no such samples have been recorded yet, 0 will be returned.

Parameters:

sample - the sample to get the count for

Returns:

the count for a given sample

getKeys

public Set<T> getKeys()

Returns the Set of sample values (or bins) for which counts have been recorded.

Returns:

the set of bins

addSample

public void addSample(T sample,
                      long number)

Increases the count for a given sample.

Parameters:

sample - the sample to increase the count for

number - the number to increase by

getMaxFreq

public long getMaxFreq()

Returns the highest frequency that is currently stored.

Returns:

highest frequency that is currently stored.

getSampleWithMaxFreq

public T getSampleWithMaxFreq()

Returns the sample which has currently the highest frequency. If there is more than one sample which share the highest frequency, returns the one that was added first.

Returns:

the sample which has currently the highest frequency

save

public void save(File file)
          throws IOException

Throws:

IOException

load

public void load(File file)
          throws IOException,
                 ClassNotFoundException

Throws:

IOException

ClassNotFoundException

clear

public void clear()

getMostFrequentSamples

public List<T> getMostFrequentSamples(int n)

Returns the n most frequent samples in the distribution. The ordering within in a group of samples with the same frequency is undefined.

Parameters:

n - the numer of most frequent samples to return.

Returns:

the n most frequent samples in the distribution.