Determining similar phrases

• US 9,298,700 B1
• Filed: 07/28/2009
• Issued: 03/29/2016
• Est. Priority Date: 07/28/2009
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Jesensky, James, Huynh, Steve
• Primary Examiner(s): BAKER, MATTHEW H
• Application Number:
  US12/510,761
• Time in Patent Office: 2,436 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 40/247 Thesauruses; Synonyms
• CPC Class Codes: G06F 40/289 Phrasal analysis, e.g. finite state techniques or chunking
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

Techniques for determining phrases that are related or similar to a source phrase are described herein. These techniques may include analyzing a corpus of text to locate a first set of phrases that differ from a received source phrase by one word. The techniques may then combine these words of the first set of phrases that differ from the source phrase to create a second set of phrases. The techniques may then filter out a portion of the second set of phrases before outputting another portion for use as an identifier, for association with a user account or for any other use.

Claims

1. One or more non-transitory computer-readable media storing computer-executable instructions that, when executed, cause one or more processors to perform acts comprising:

identifying a first set of phrases that exist within a corpus of text, individual ones of the first set of phrases including a substitute word that replaces one word of a source phrase comprising multiple words;

generating a second set of phrases by combining substitute words from the first set of phrases with one another such that individual ones of the second set of phrases include a first substitute word from a first phrase of the first set of phrases and a second substitute word from a second phrase of the first set of phrases;

generating a personalized user interface including at least a portion of the second set of phrases and at least one rule selector associated with a corresponding rule for using the at least the portion of the second set of phrases; and

receiving a user selection of the at least the portion of the second set of phrases and the at least one rule selector associated with the corresponding rule for using the at least the portion of the second set of phrases.

2. One or more non-transitory computer-readable media as recited in claim 1, wherein the identifying of the first set of phrases comprises:

determining a part-of-speech combination of the source phrase; and

for individual words of the source phrase, identifying phrases that differ only by that word but that comprise the part-of-speech combination of the source phrase.

3. One or more non-transitory computer-readable media as recited in claim 1, further storing computer-executable instructions that, when executed, cause the one or more processors to perform acts comprising:

determining a part-of-speech combination of the source phrase; and

filtering out phrases of the second set of phrases that do not comprise the part-of-speech combination of the source phrase.

4. One or more non-transitory computer-readable media as recited in claim 1, further storing computer-executable instructions that, when executed, cause the one or more processors to perform acts comprising:

determining, for individual phrases of the second set of phrases, that a phrase appears in a published corpus of text at least a predetermined number of times; and

filtering out phrases of the second set of phrases that do not appear in the published corpus of text at least the predetermined number of times.

5. One or more non-transitory computer-readable media as recited in claim 1, further storing computer-executable instructions that, when executed, cause the one or more processors to perform an act comprising ranking the second set of phrases according to a mutual information score for individual phrases of the second set of phrases, and wherein the generating the personalized user interface including the at least the portion of the second set of phrases comprises selecting individual phrases of the second set of phrases to be presented at least in part based on the ranking.

6. One or more non-transitory computer-readable media as recited in claim 1, further storing computer-executable instructions that, when executed, cause the one or more processors to perform acts comprising associating the user selection of the at least the portion of the second set of phrases and the at least one rule selector associated with the corresponding rule for using the at least the portion of the second set of phrases with an aspect of a user account of a user.

7. One or more non-transitory computer-readable media as recited in claim 6, wherein the aspect of the user account comprises a payment instrument associated with the user account, a physical or digital location associated with the user account, or a delivery method associated with the user account.

8. One or more non-transitory computer-readable media as recited in claim 1, wherein individual phrases of the portion of the second set of phrases are grammatically correct and comprise two or more words.

9. A computer-implemented method comprising:

under control of one or more computer systems configured with executable instructions;

identifying a source phrase comprising multiple words, individual ones of the multiple words comprising a respective part-of-speech and a respective position within the source phrase;

determining a first set of phrases that;

(i) exist within a corpus of text;

(ii) individually comprise multiple words having a same respective part-of-speech in a same position as the source phrase, and (iii) individually differ from the source phrase by one word;

generating a second set of phrases that individually include multiple words from the first set of phrases that differ from the source phrase;

generating a personalized user interface including at least a portion of the second set of phrases and at least one rule selector associated with a corresponding rule for using the at least the portion of the second set of phrases; and

receiving a user selection of the at least the portion of the second set of phrases and the at least one rule selector associated with the corresponding rule for using the at least the portion of the second set of phrases in an electronic commerce transaction.

10. A computer-implemented method as recited in claim 9, wherein the generating of the second set of phrases comprises combining the words of the first set of phrases that differ from the source phrase together in their respective positions.

11. A computer-implemented method as recited in claim 9, wherein the generating the personalized user interface comprises presenting phrases of the second set of phrases that comprise a same part-of-speech combination as the source phrase and refraining from presenting phrases of the second set of phrases that comprise a different part-of-speech combination than the source phrase.

12. A computer-implemented method as recited in claim 9, further comprising:

before the determining of the first set of phrases, adding an additional word having a corresponding part-of-speech to a particular position of the source phrase to create a derived source phrase; and

wherein individual phrases of the first set of phrases further comprise a word having a same part-of-speech in a same respective position as the additional word.

13. A computer-implemented method as recited in claim 9, wherein the determining of the first set of phrases comprises analyzing the corpus of text to locate the first set of phrases.

14. One or more non-transitory computer-readable media storing computer-executable instructions that, when executed, cause one or more processors to perform acts comprising:

receiving a grammatically correct source phrase as a received source phrase comprising multiple words;

identifying a derived source phrase that comprises at least a portion of the received source phrase and at least one additional word;

identifying a first set of phrases that exist within a corpus of text, individual ones of the first set of phrases differing from the derived source phrase by one word;

generating a second set of phrases from the first set of phrases, the second set of phrases individually including multiple words from the first set of phrases that differ from the derived source phrase;

removing at least one word from a portion of phrases of the second set of phrases;

generating a personalized user interface including at least the portion of the phrases and at least one rule selector associated with a corresponding rule for using the at least the portion of the phrases; and

receiving a user selection of the at least the portion of the phrases and the at least one rule selector associated with the corresponding rule for using the at least the portion of the phrases.

15. One or more non-transitory computer-readable media as recited in claim 14, wherein the identifying of the derived source phrase comprises:

analyzing the corpus of text to determine how often the derived source phrase appears in the corpus of text relative to other phrases that also comprise at least a portion of the received source phrase and a different at least one additional word; and

selecting the derived source phrase based at least in part on the determination.

16. One or more non-transitory computer-readable media as recited in claim 14, wherein the at least one additional word resides within the derived source phrase at a particular position, and wherein the removing of the at least one word from the portion of the phrases comprises removing a word at the particular position from individual phrases of the portion of the phrases.

17. One or more non-transitory computer-readable media as recited in claim 14, wherein the derived source phrase comprises a first derived source phrase, and the one or more non-transitory computer-readable media further storing computer-executable instructions that, when executed, cause the one or more processors to perform acts comprising:

identifying a second derived source phrase that comprises at least a portion of the received source phrase and a different at least one additional word;

identifying another first set of phrases individually found within the corpus of text and individually differing from the second derived source phrase by one word;

generating another second set of phrases from the another first set of phrases by combining the words from the another first set of phrases that differ from the second derived source phrase; and

merging the second set of phrases associated with the first derived source phrase with the another second set of phrases associated with the second derived source phrase before the generating the personalized user interface including the at least the portion of the phrases.

18. One or more non-transitory computer-readable media as recited in claim 17, wherein the merging comprises merging phrases based at least in part on a number of derived source phrases that produced individual respective phrases.

19. One or more non-transitory computer-readable media as recited in claim 14, wherein the derived source phrase excludes one or more words of the received source phrase.

20. One or more non-transitory computer-readable media as recited in claim 14, wherein at least a portion of phrases of the first set of phrases individually differ from the derived source phrase by two words.

21. One or more computing devices comprising:

one or more processors; and

the one or more non-transitory computer-readable media as recited in claim 14.

22. One or more computing devices comprising:

one or more processors; and

one or more computer-readable media storing computer-executable instructions that, when executed, cause the one or more processors to perform acts comprising;

identifying a first set of phrases that exist within a corpus of text, individual ones of the first set of phrases including a substitute word that replaces one word of a source phrase comprising multiple words;

generating a second set of phrases by combining substitute words from the first set of phrases with one another, the second set of phrases individually including multiple substitute words from the first set of phrases;

generating a personalized user interface including at least a portion of the second set of phrases and at least one rule selector associated with a corresponding rule for using the at least the portion of the second set of phrases; and

receiving a user selection of the at least the portion of the second set of phrases and the at least one rule selector associated with the corresponding rule for using the at least the portion of the second set of phrases.

23. One or more computing devices as recited in claim 22, wherein the identifying of the first set of phrases comprises:

determining a part-of-speech combination of the source phrase; and

for individual words of the source phrase, identifying phrases that differ only by that word but that comprise the part-of-speech combination of the source phrase.

24. One or more computing devices as recited in claim 22, the one or more computer-readable media further storing computer-executable instructions that, when executed, cause the one or more processors to perform acts comprising:

determining a part-of-speech combination of the source phrase; and

filtering out phrases of the second set of phrases that do not comprise the part-of-speech combination of the source phrase.

25. One or more computing devices as recited in claim 22, the one or more computer-readable media further storing computer-executable instructions that, when executed, cause the one or more processors to perform acts comprising:

determining, for individual phrases of the second set of phrases, if a phrase appears in a published corpus of text at least a predetermined number of times; and

filtering out phrases of the second set of phrases that do not appear in the published corpus of text at least the predetermined number of times.

26. One or more computing devices as recited in claim 22, the one or more computer-readable media further storing computer-executable instructions that, when executed, cause the one or more processors to perform acts comprising ranking the second set of phrases according to a mutual information score for individual phrases of the second set of phrases, and wherein the generating the personalized user interface including the at least the portion of the second set of phrases comprises selecting individual phrases of the second set of phrases to be presented at least in part based on the ranking.

27. One or more computing devices as recited in claim 22, the one or more computer-readable media further storing computer-executable instructions that, when executed, cause the one or more processors to perform acts comprising associating the user selection of the at least the portion of the second set of phrases and the at least one rule selector associated with the corresponding rule for using the at least the portion of the second set of phrases with an aspect of a user account of a user.

28. One or more computing devices as recited in claim 27, wherein the aspect of the user account comprises a payment instrument associated with the user account, a physical or digital location associated with the user account, or a delivery method associated with the user account.

29. One or more computing devices as recited in claim 22, wherein individual phrases of the portion of the second set of phrases are grammatically correct and comprise two or more words.

30. One or more non-transitory computer-readable media storing computer-executable instructions that, when executed, cause one or more processors to perform a method comprising:

identifying a source phrase comprising multiple words, individual ones of the multiple words comprising a respective part-of-speech and a respective position within the source phrase;

determining a first set of phrases that;

(i) exist within a corpus of text;

(ii) individually comprise multiple words having a same respective part-of-speech in a same position as the source phrase, and (iii) individually differ from the source phrase by one word;

generating a second set of phrases that individually include multiple words from the first set of phrases that differ from the source phrase; and

generating a personalized user interface including at least a portion of the second set of phrases and at least one rule selector associated with a corresponding rule for using the at least the portion of the second set of phrases; and

receiving a user selection of the at least the portion of the second set of phrases and the at least one rule selector associated with the corresponding rule for using the at least the portion of the second set of phrases in an electronic commerce transaction.

31. One or more non-transitory computer-readable media as recited in claim 30, wherein the generating of the second set of phrases comprises combining the words of the first set of phrases that differ from the source phrase together in their respective positions.

32. One or more non-transitory computer-readable media as recited in claim 30, wherein individual phrases of the at least the portion of the second set of phrases comprise a same part-of-speech combination as the source phrase.

33. One or more non-transitory computer-readable media as recited in claim 30, further storing computer-executable instructions that, when executed, cause the one or more processors to perform acts comprising:

before the determining of the first set of phrases, adding an additional word having a corresponding part-of-speech to a particular position of the source phrase to create a derived source phrase; and

wherein individual phrases of the first set of phrases further comprise a word having a same part-of-speech in a same position as the additional word.

34. One or more non-transitory computer-readable media as recited in claim 30, wherein the determining of the first set of phrases comprises analyzing the corpus of text to locate the first set of phrases.

35. One or more non-transitory computer-readable media as recited in claim 30, wherein the at least one rule selector associated with the corresponding rule for using the at least the portion of the second set of phrases comprises at least an indication of how another user may use the at least the portion of the second set of phrases.