pt.tumba.ngram
Class NGramConstants

java.lang.Object
  pt.tumba.ngram.NGramConstants

public class NGramConstants
extends java.lang.Object

Contant values used in the TCatNG package.

Author:

Bruno Martins

Field Summary

static int	SIMILARITYJIANG Use the similarity metric proposed by Jiand & Conranth in "Semantic Similarity Based on Corpus Statistics and Lexical Taxonomy".
static int	SIMILARITYLIN Use the similarity metric proposed by Lin in "An information-theoretic definition of similarity".
static int	SIMILARITYOUTOFPLACE Use the similarity metric proposed by Cavnar & Trenkle.
static byte[]	SKIPABLE Bytes skipable while building the proviles.
static boolean	SMOOTHING Use Good-Turing smoothing on the NGram occurence frequency.
static int	USEDNGRAMSMAX The lowest ranking position for storage in the N-gram profiles.
static int	USEDNGRAMSMIN The highest ranking position for storage in the N-gram profile.

Constructor Summary

NGramConstants()

Method Summary

Methods inherited from class java.lang.Object

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

Field Detail

SKIPABLE

public static final byte[] SKIPABLE

Bytes skipable while building the proviles. These correspond to number characters.

USEDNGRAMSMAX

public static final int USEDNGRAMSMAX

The lowest ranking position for storage in the N-gram profiles. For instance with USEDNGRAMSMAX=400 only the top 400 highest occurring N-grams will be stored.

See Also:

Constant Field Values

USEDNGRAMSMIN

public static final int USEDNGRAMSMIN

The highest ranking position for storage in the N-gram profile. For instance with USEDNGRAMSMIN=200 the top 200 highest occurring N-grams will be skipped.

See Also:

Constant Field Values

SIMILARITYLIN

public static final int SIMILARITYLIN

Use the similarity metric proposed by Lin in "An information-theoretic definition of similarity".

See Also:

Constant Field Values

SIMILARITYJIANG

public static final int SIMILARITYJIANG

Use the similarity metric proposed by Jiand & Conranth in "Semantic Similarity Based on Corpus Statistics and Lexical Taxonomy".

See Also:

Constant Field Values

SIMILARITYOUTOFPLACE

public static final int SIMILARITYOUTOFPLACE

Use the similarity metric proposed by Cavnar & Trenkle.

See Also:

Constant Field Values

SMOOTHING

public static final boolean SMOOTHING

Use Good-Turing smoothing on the NGram occurence frequency.

See Also:

Constant Field Values

Constructor Detail

NGramConstants

public NGramConstants()