Browse Prior Art Database

# Filter for Spelling Correction Function

IP.com Disclosure Number: IPCOM000042095D
Original Publication Date: 1984-Mar-01
Included in the Prior Art Database: 2005-Feb-03
Document File: 1 page(s) / 12K

IBM

## Related People

Urquhart, RJ: AUTHOR

## Abstract

This article describes a method for increasing throughput in a spelling correction function by providing a cross-match filter between a limited number of characters of the input word and dictionary words. The filter examines the second, third, and fourth characters of both the input (I) word, presented for spelling correction, and candidate dictionary (D) words, looking for particular mismatch patterns. The reject conditions are (1) no match between the second and third characters and (2) exactly one match between the second and third characters and for each of those matches one particular additional mismatch. The following table defines the five reject patterns and gives an example of each. Ia and Db denote the a-th position of the input word and b-th position of dictionary word.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 72% of the total text.

Page 1 of 1

Filter for Spelling Correction Function

This article describes a method for increasing throughput in a spelling correction function by providing a cross-match filter between a limited number of characters of the input word and dictionary words. The filter examines the second, third, and fourth characters of both the input (I) word, presented for spelling correction, and candidate dictionary (D) words, looking for particular mismatch patterns. The reject conditions are (1) no match between the second and third characters and (2) exactly one match between the second and third characters and for each of those matches one particular additional mismatch. The following table defines the five reject patterns and gives an example of each. Ia and Db denote the a-th position of the input word and b-th position of dictionary word. Number Pattern Example 1 I2 = D2 & I2 = D3 & Input = after I3 = D2 & I3 = D3 Dict = accept 2a I2 = D2 & Input = delicious I3 = D3 & I3 = D4 & Dict = depreciate I4 = D3 & I4 = D4 2b I2 = D3 & I3 = D4 Input = stems Dict = situation 2c I3 = D2 & I4 = D3 Input = situation 2d I3 = D3 & I4 = D4 Input = ladder Dict = ledge This filter is particularly powerful in reducing processing time. When a word root is rejected by the filter, both the root and all its derived forms (variants) can be bypassed. This is due to the relationship of roots to variants in the dictionary structure. During the encoding or build phase the number of leading characters that are...