Grammatically correct contraction spelling suggestions for french

• US 20060200340A1
• Filed: 03/01/2005
• Published: 09/07/2006
• Est. Priority Date: 03/01/2005
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Caldwell, Nicholas Van, Fontenelle, Thierry
• Granted Patent: US 7,599,828 B2
• Time in Patent Office: Days
• US Class Current: 704/5
• CPC Class Codes: G06F 40/232 Orthographic correction, e.g. spell checking or vowelisation

Abstract

A method and product are provided that identify properly spelled words from an input string. Separate lexical entries are identified for different parts of an input word. Grammatical information associated with one of the lexical entries is examined to determine if it satisfies a grammatical constraint associated with another of the lexical entries. A properly spelled word is formed by combining the separate lexical entries if the grammatical information satisfies the constraint. Under some embodiments of the invention, the separate lexical entries are identified by performing a first search of a lexicon using a first part of the input string and by performing a second search of the lexicon using a second part of the input string.

Claims

1. A method of identifying a properly spelled word from an input string, the method comprising:

identifying a first part of an input string as representing a part of a possible contraction of two words;

identifying a second part of the input string as representing a complete word in a lexicon;

retrieving grammatical information for the complete word from the lexicon;

using the grammatical information to determine if the complete word satisfies a grammatical constraint; and

identifying a contraction formed by combining the first part of the input string and the complete word in the lexicon into a properly spelled word.

2. The method of claim 1 wherein identifying a first part of an input string comprises searching a lexicon storage structure.

3. The method of claim 2 wherein identifying a first part of an input string further comprises finding the first part of the input string followed by an apostrophe while searching the lexicon storage structure.

4. The method of claim 2 wherein identifying a second part of the input string comprises performing a second search of the lexicon storage structure.

5. The method of claim 4 wherein identifying a second part of the input string further comprises locating a complete word during the second search of the lexicon storage structure that is within a threshold distance of the second part of the input string.

6. The method of claim 5 wherein during the search of the lexicon storage structure used to identify the first part of the input string, a first threshold distance is used, and wherein during the second search of the lexicon storage structure a second threshold distance that is lower than the first threshold distance is used.

7. The method of claim 6 wherein the second threshold distance only allows changes in capitalization and accent marks.

8. The method of claim 1 wherein determining if the complete word satisfies a grammatical constraint comprises determining if the part of speech of the complete word satisfies a part of speech constraint.

9. The method of claim 1 wherein determining if the complete word satisfies a grammatical constraint comprises determining if the grammatical person of the complete word satisfies a grammatical person constraint.

10. The method of claim 1 wherein identifying a contraction further comprises placing an apostrophe between the first part of the input string and the complete word to form the properly spelled word.

11. A computer-readable medium having computer-executable instructions for performing steps comprising:

initiating a first traversal of a lexicon search structure;

finding an elided word in the lexicon search structure that corresponds to a portion of an input text string during the first traversal of the lexicon search structure;

initiating a second traversal of the lexicon search structure;

finding a complete word in the lexicon search structure that corresponds to a second portion of the input text string during the second traversal of the lexicon search structure; and

combining the elided word and the complete word to form a properly spelled word.

12. The computer-readable medium of claim 11 wherein initiating a first traversal and initiating a second traversal comprises initiating the first and second traversals at the same point in the lexicon search structure.

13. The computer-readable medium of claim 11 further comprising after finding the complete word, continuing the second traversal of the lexicon search structure.

14. The computer-readable medium of claim 13 further comprising after continuing the second traversal of the lexicon search structure:

finding a second complete word in the lexicon search structure that corresponds to a second portion of the input text string during the second traversal of the lexicon search structure; and

combining the elided word and the second complete word to form a second properly spelled word.

15. The computer-readable medium of claim 11 wherein initiating a second traversal of the lexicon search structure comprises pausing the first traversal of the lexicon search structure until the second traversal of the lexicon search structure is complete.

16. The computer-readable medium of claim 15 further comprising continuing the first traversal of the lexicon search structure after the second traversal of the lexicon search structure is complete.

17. The computer-readable medium of claim 16 further comprising finding a complete word in the lexicon search structure that corresponds to the entire input text string during the first traversal of the lexicon search structure.

18. The computer-readable medium of claim 11 wherein combining the elided word and the complete word comprises determining if the complete word satisfies a grammatical constraint associated with the elided word before combining the elided word and the complete word.

19. A computer-readable medium having computer-executable instructions for performing steps comprising:

identifying separate lexical entries for different parts of an input word;

determining if the grammatical information associated with one of the lexical entries satisfies a constraint associated with another of the lexical entries; and

forming a properly spelled word by combining the separate lexical entries if the grammatical information satisfies the constraint.

20. The computer-readable medium of claim 19 wherein identifying separate lexical entries comprises performing a first search of a lexicon using a first part of the input word and a second search of the lexicon using a second part of the input word.