smile.nlp.relevance

Class TFIDF

java.lang.Object

smile.nlp.relevance.TFIDF

All Implemented Interfaces:

RelevanceRanker

public class TFIDF
extends java.lang.Object
implements RelevanceRanker

The tf-idf weight (term frequency-inverse document frequency) is a weight often used in information retrieval and text mining. This weight is a statistical measure used to evaluate how important a word is to a document in a collection or corpus. The importance increases proportionally to the number of times a word appears in the document but is offset by the frequency of the word in the corpus. Variations of the tf-idf weighting scheme are often used by search engines as a central tool in scoring and ranking a document's relevance given a user query. One of the simplest ranking functions is computed by summing the tf-idf for each query term; many more sophisticated ranking functions are variants of this simple model.

One well-studied technique is to normalize the tf weights of all terms occurring in a document by the maximum tf in that document. For each document d, let tfmax(d) be the maximum tf over all terms in d. Then, we compute a normalized term frequency for each term t in document d by

tf = a + (1? a) tf_t,d / tfmax(d)

where a is a value between 0 and 1 and is generally set to 0.4, although some early work used the value 0.5. The term a is a smoothing term whose role is to damp the contribution of the second term - which may be viewed as a scaling down of tf by the largest tf value in d. The main idea of maximum tf normalization is to mitigate the following anomaly: we observe higher term frequencies in longer documents, merely because longer documents tend to repeat the same words over and over again. Maximum tf normalization does suffer from the following issues:

1. The method is unstable in the following sense: a change in the stop word list can dramatically alter term weightings (and therefore ranking). Thus, it is hard to tune.
2. A document may contain an outlier term with an unusually large number of occurrences of that term, not representative of the content of that document.
3. More generally, a document in which the most frequent term appears roughly as often as many other terms should be treated differently from one with a more skewed distribution.

See Also:

BM25

Constructor Summary

Constructors

Constructor and Description
TFIDF() Constructor.
TFIDF(double smoothing) Constructor.

Method Summary

Modifier and Type	Method and Description
double	rank(Corpus corpus, TextTerms doc, java.lang.String[] terms, int[] tf, int n) Returns a relevance score between a set of terms and a document based on a corpus.
double	rank(Corpus corpus, TextTerms doc, java.lang.String term, int tf, int n) Returns a relevance score between a term and a document based on a corpus.
double	rank(int tf, int maxtf, long N, long n) Returns a relevance score between a term and a document based on a corpus.

Methods inherited from class java.lang.Object

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

Constructor Detail

TFIDF

public TFIDF()

Constructor.

TFIDF

public TFIDF(double smoothing)

Constructor.

Parameters:

smoothing - the smoothing parameter in maximum tf normalization.

Method Detail

rank

public double rank(int tf,
                   int maxtf,
                   long N,
                   long n)

Returns a relevance score between a term and a document based on a corpus.

Parameters:

tf - the frequency of searching term in the document to rank.

maxtf - the maximum frequency over all terms in the document.

N - the number of documents in the corpus.

n - the number of documents containing the given term in the corpus;

rank

public double rank(Corpus corpus,
                   TextTerms doc,
                   java.lang.String term,
                   int tf,
                   int n)

Description copied from interface: RelevanceRanker

Returns a relevance score between a term and a document based on a corpus.

Specified by:

rank in interface RelevanceRanker

Parameters:

corpus - the corpus.

doc - the document to rank.

term - the searching term.

tf - the term frequency in the document.

n - the number of documents containing the given term in the corpus;

rank

public double rank(Corpus corpus,
                   TextTerms doc,
                   java.lang.String[] terms,
                   int[] tf,
                   int n)

Description copied from interface: RelevanceRanker

Returns a relevance score between a set of terms and a document based on a corpus.

Specified by:

rank in interface RelevanceRanker

Parameters:

corpus - the corpus.

doc - the document to rank.

terms - the searching terms.

tf - the term frequencies in the document.

n - the number of documents containing the given term in the corpus;