System for determining probable meanings of inputted words

• US 7,681,147 B2
• Filed: 12/13/2005
• Issued: 03/16/2010
• Est. Priority Date: 12/13/2005
• Status: Expired due to Fees

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Riise, Soren, Grealish, Paul J., Richardson-Bunbury, David, Stipp, Eugene H., Patel, Devesh
• Primary Examiner(s): Bashore; William L
• Assistant Examiner(s): KEATON, SHERROD L
• Application Number:
  US11/314,956
• Publication Number:
  US 20070136689A1
• Time in Patent Office: 1,554 Days
• Field of Search: 715/816, 715/532, 715/817, 715/780, 707/7, 707/100, 707 3- 5, 707/10, 707/102, 704/9
• US Class Current: 715/816
• CPC Class Codes: G06F 40/295 Named entity recognition
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A system is disclosed for determining probable meanings of words. An input of a word is obtained. Probable meanings of the word may be determined in accordance with a prior probability of probable meanings of the word and a context frequency probability of probable meanings of the word.

Claims

1. A computer-implemented method for determining probable meanings of inputted search query terms, the method comprising:

receiving an input of at least one search query word by a networked server;

determining a probable meaning of the at least one word by a processor of the networked server in accordance with a prior probability and a context frequency probability of probable meanings of the word, wherein the context frequency probability consists of the meaning of the word in a context of a plurality of terms immediately preceding or immediately following the word, wherein the prior probability comprises a probability that the word refers to a predetermined meaning irregardless of a query context in which the word is used, which prior probability is derived from previous analysis of documents containing the word, wherein determining the probable meaning comprises;

estimating an expected final probability for the at least one word given the prior probability of the at least one word;

deriving an inverse combine function that uses the prior probability and the expected final probability to determine the context frequency probability of the at least one word;

using a combine mathematical function to combine the prior probability and the context frequency probability to produce a final probability of the probable meaning of the at least one search query word; and

using the final probability of the probable meaning of the at least one word by the networked server to influence search results returned in response to the inputted at least one word.

2. The method of claim 1, further comprising displaying a result in accordance with the probable meanings of at least one inputted word.

3. The method of claim 2, wherein the meaning with the highest probability is displayed first.

4. The method of claim 1, further comprising displaying an advertisement in accordance with the probable meanings of at least one inputted word.

5. The method of claim 4, wherein the advertisement is geographically relevant to the user in accordance with the probable meaning of the word.

6. The method of claim 1, further comprising indexing web pages for a search in accordance with the probable meanings of the inputted word.

7. The method of claim 1, further comprising plotting locations on a map in accordance with the probable meanings of the inputted word.

8. The method of claim 1, further comprising calculating enhanced statistics for use in a query analysis in accordance with the probable meaning of the word.

9. The method of claim 1, wherein the prior probability further comprises a location related probability which includes the probability that the word refers to a location.

10. The method of claim 1, further comprising establishing disambiguators in accordance with a text in which the word is used, wherein the disambiguators comprise specific disambiguators which include the meaning of the word as determined in a context of a text that includes the word, other than the text immediately preceding or following the word.

11. The method of claim 10, wherein the disambiguators comprise at least one of:

specific disambiguators which includes the meaning of the word as determined in a context of a text that includes the word, other than the text immediately preceding or following the word; and

comprise general disambiguators which include a term in a text that uses the word, wherein the term shares an attribute with the word being disambiguated.

19. The method of claim 1, wherein the combine mathematical function comprises

$\frac{a·}{}$

b a ·

b + ( 1 - a ) ⁢

( 1 - b ) ,

where a comprises the prior probability and b comprises the context frequency probability.

29. The method of claim 1, further comprising:

ignoring the context frequency probability if it is within a predetermined range of a neutral context frequency.

30. The method of claim 1, further comprising:

moderating values for the context frequency probability.

12. A system to determine probable meanings of words input as search query terms, comprising:

a networked server to receive an input of at least one search query word;

a processor coupled with the networked server to analyze the inputted word and to determine a probable meaning of the word in accordance with a prior probability of probable meanings of the word and a context frequency probability of probable meanings of the word, wherein the prior probability comprises a probability that the word refers to a predetermined meaning irregardless of a query context in which the word is used, which prior probability is derived from previous analysis of documents containing the word, wherein the processor;

estimates an expected final probability for the at least one word given the prior probability of the at least one word;

derives an inverse combine function that uses the prior probability and the expected final probability to determine the context frequency probability of the at least one word; and

uses a combine mathematical function to combine the prior probability and the context frequency probability to produce a final probability of the probable meaning of the at least one search query word; and

a memory coupled with the processor to store the prior probabilities and the context frequency probabilities of probable meanings of previously-inputted words;

wherein the processor generates search results listings in accordance with the final probability of the probable meaning of the word for presentation to a user.

13. The system of claim 12, wherein the context frequency probability comprises the meaning of the word in a context of a plurality of terms immediately preceding or immediately following the word.

14. The system of claim 12, further comprising:

a user device to display an advertisement in accordance with the probable meanings of at least one inputted word.

15. The system of claim 12, wherein a meaning of the word comprises a geographic location.

20. The system of claim 12, wherein the processor establishes disambiguators in accordance with a text in which the word is used, wherein the disambiguators comprise at least one of:

specific disambiguators which includes the meaning of the word as determined in a context of a text that includes the word, other than the text immediately preceding or following the word; and

general disambiguators which include a term in a text that uses the word, wherein the term shares an attribute with the word being disambiguated.

21. The system of claim 12, wherein the processor delivers an advertisement to the client computer in accordance with the probable meaning of the word.

22. The system of claim 21, wherein the advertisement is geographically relevant to the user in accordance with the probable meaning of the word.

23. The system of claim 12, wherein the processor indexes web pages for a search in accordance with the probable meanings of the word.

24. The system of claim 12, wherein the processor plots locations on a map in accordance with the probable meanings of the word.

25. The system of claim 12, wherein the processor calculates enhanced statistics for use in a query analysis in accordance with the probable meaning of the word.

16. A client/server network for determining probable meanings of search query words, comprising:

a client computer to enter at least one word on the network;

a server to receive an input of the at least one word from the network;

a processor to determine a probable meaning of the at least one word in accordance with a prior probability and a context frequency probability of probable meanings of the word, wherein the prior probability comprises a probability that the word refers to a predetermined meaning irregardless of the query context in which the word is used, which prior probability is derived from previous analysis of documents containing the word;

wherein the processor;

estimates an expected final probability for the at least one word given the prior probability of the at least one word;

derives an inverse combine function that uses the prior probability and the expected final probability to determine the context frequency probability of the at least one word;

uses a combine mathematical function to combine the prior probability and the context frequency probability to produce a final probability of the probable meaning of the at least one search query word; and

a memory coupled with the processor to store the prior probabilities and the context frequency probabilities of probable meanings of previously-inputted words;

wherein the processor generates search results listings in accordance with the final probability of the probable meaning of the word for presentation to the client computer.

17. The client/server network of claim 16, wherein the context frequency probability comprises the meaning of the word in a context of terms immediately preceding or immediately following the word.

18. The client/server network of claim 16, wherein a meaning of the word comprises a geographic location.

26. The client/server network of claim 16, wherein the processor establishes disambiguators in accordance with a text in which the word is used, wherein the disambiguators comprise specific disambiguators which include the meaning of the word as determined in a context of a text that includes the word, other than the text immediately preceding or following the word.

27. The client/server network of claim 26, wherein the disambiguators comprise at least one of:

specific disambiguators which includes the meaning of the word as determined in a context of a text that includes the word, other than the text immediately preceding or following the word; and

general disambiguators which include a term in a text that uses the word, wherein the term shares an attribute with the word being disambiguated.

28. The client/server network of claim 16, wherein the combine mathematical function comprises

$\frac{a·}{}$

b a ·

b + ( 1 - a ) ⁢

( 1 - b ) ,

where a comprises the prior probability and b comprises the context frequency probability.

31. The client/server network of claim 16, wherein the processor:

ignores the context frequency probability if it is within a predetermined range of a neutral context frequency.