#jsDisabledContent { display:none; } My Account |  Register |  Help

# W-shingling

Article Id: WHEBN0005548352
Reproduction Date:

 Title: W-shingling Author: World Heritage Encyclopedia Language: English Subject: Collection: Natural Language Processing Publisher: World Heritage Encyclopedia Publication Date:

### W-shingling

In natural language processing a w-shingling is a set of unique "shingles" (n-grams, contiguous subsequences of tokens in a document) that can be used to gauge the similarity of two documents. The w denotes the number of tokens in each shingle in the set.

The document, "a rose is a rose is a rose" can be tokenized as follows:

(a,rose,is,a,rose,is,a,rose)

The set of all contiguous sequences of 4 tokens (4-grams) is

{ (a,rose,is,a), (rose,is,a,rose), (is,a,rose,is), (a,rose,is,a), (rose,is,a,rose) } = { (a,rose,is,a), (rose,is,a,rose), (is,a,rose,is) }

## Contents

• Resemblance 1
• References 3

## Resemblance

For a given shingle size, the degree to which two documents A and B resemble each other can be expressed as the ratio of the magnitudes of their shinglings' intersection and union, or

r(A,B)=

where |A| is the size of set A. The resemblance is a number in the range [0,1], where 1 indicates that two documents are identical. This definition is identical with the Jaccard coefficient describing similarity and diversity of sample sets.

## References

• (Manber 1993) Finding Similar Files in a Large File System. Does not yet use the term "shingling". Available as PDF
• (Broder, Glassman, Manasse, and Zweig 1997) Syntactic Clustering of the Web. SRC Technical Note #1997-015. Available at HTML here