de.tudarmstadt.ukp.dkpro.core.norvig

Class NorvigSpellingAlgorithm

java.lang.Object

de.tudarmstadt.ukp.dkpro.core.norvig.NorvigSpellingAlgorithm

public class NorvigSpellingAlgorithm
extends Object

Spelling corrector based on Norvig's algorithm.

See Also:

Norvig's algorithm

Constructor Summary

Constructors

Constructor and Description
NorvigSpellingAlgorithm()

Method Summary

Modifier and Type	Method and Description
protected void	consider(String candidate) Consider the given candidate.
String	correct(String aWord) Try to find a correction for the given word.
protected List<String>	edits(String word) Get a list for all possible variants of the given word containing an insertion, deletion, replacement or transposition.
protected String	getBestCandidate(String word) Try to find a correction for the given word.
protected boolean	isNumber(String aWord)
void	reset()
protected void	resetScore()
void	train(Reader aReader) Read words from the given reader and count their occurrences.
void	train(URL aUrl, String aEncoding) Read words from the given URL and count their occurrences.

Methods inherited from class java.lang.Object

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

Constructor Detail

NorvigSpellingAlgorithm

public NorvigSpellingAlgorithm()

Method Detail

resetScore

protected void resetScore()

reset

public void reset()

train

public void train(Reader aReader)
           throws IOException

Read words from the given reader and count their occurrences.

Parameters:

aReader - the reader.

Throws:

IOException - if the words cannot be read.

train

public void train(URL aUrl,
                  String aEncoding)
           throws IOException

Read words from the given URL and count their occurrences.

Parameters:

aUrl - the URL to load the words from.

aEncoding - the encoding.

Throws:

IOException - if the words cannot be read.

edits

protected List<String> edits(String word)

Get a list for all possible variants of the given word containing an insertion, deletion, replacement or transposition.

Parameters:

word - the word.

Returns:

the list of variants.

correct

public String correct(String aWord)

Try to find a correction for the given word. The word may contain up to two edits. If no better alternative is found, the word is returned verbatim. For performance reasons corrections are cached.

Parameters:

aWord - the word to correct (has to be lower-case)

Returns:

the possible correction.

isNumber

protected boolean isNumber(String aWord)

getBestCandidate

protected String getBestCandidate(String word)

Try to find a correction for the given word. The word may contain up to two edits. If no better alternative is found, the word is returned verbatim.

Parameters:

word - the word to correct (has to be lower-case)

Returns:

the possible correction.

consider

protected void consider(String candidate)

Consider the given candidate. If it is better than a previously found candidate, then remember it, otherwise forget it.

Parameters:

candidate - the candidate to consider.