Accentuating search results

• US 9,623,119 B1
• Filed: 06/29/2010
• Issued: 04/18/2017
• Est. Priority Date: 06/29/2010
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Kim, Hyung-Jin, Wei, Leonard
• Primary Examiner(s): Reyes, Mariela
• Assistant Examiner(s): Davanlou, Soheila
• Application Number:
  US12/825,461
• Time in Patent Office: 2,485 Days
• Field of Search: 707706, 707709, 707723, 11706, 11709, 11723
• CPC Class Codes: A61K 48/00 Medicinal preparations containing genetic material which is inserted into cells of the living body to treat genetic diseases; Gene therapy
• CPC Class Codes: G06F 16/24578 using ranking

Abstract

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for improving search results. In one aspect, a method includes receiving documents responsive to a query, each document having an associated score indicative of the document'"'"'s relevance to the query. The method includes determining, for a plurality of the associated scores, a distribution of the scores along a dimension that measures trustworthiness of the scores. The method also includes reducing each of one or more of the plurality of associated scores by an amount determined based on the score'"'"'s respective trustworthiness and properties of the distribution, and ranking the documents to account for the reduced scores.

Claims

1. A computer-implemented method, comprising:

receiving documents responsive to a query, each document having a respective associated score indicative of the document'"'"'s relevance to the query, wherein;

each respective associated score is based on a respective value for each of a plurality of different score factors;

each score factor for a given associated score corresponds to a different criterion than each other score factor for the given associated score; and

each score factor has a respective trustworthiness score that indicates a respective reliability of values determined based on the score factor;

determining, based on the trustworthiness scores for the score factors of a plurality of the associated scores of a plurality of different documents, a value-based distribution of the trustworthiness scores for each of the different score factors of the plurality of associated scores along a dimension based on relative values of the trustworthiness scores, the distribution including trustworthiness scores for multiple score factors for each of the different documents;

adjusting the value of at least one score factor of a given associated score by an amount that is determined based on;

a relative position of the at least one score factor'"'"'s respective trustworthiness score in the distribution with respect to other trustworthiness scores in the distribution; and

a measure of how widely the trustworthiness scores for each of the different score factors of the plurality of associated scores are distributed in the distribution;

adjusting the given associated score based on the adjusted value of the at least one score factor; and

ranking the documents to account for the adjusted associated score.

2. The method of claim 1, wherein the distribution is a frequency distribution and wherein the amount is determined based on a distance between a lowest and a highest trustworthiness score in the distribution and the at least one score factor'"'"'s trustworthiness score.

3. The method of claim 1, wherein the amount is determined based on a number of intervals in the distribution and the at least one score factor'"'"'s trustworthiness score.

4. The method of claim 1, wherein the amount is determined based on a shape of the distribution and the score factor'"'"'s trustworthiness score.

5. The method of claim 1, wherein a first score factor of the plurality of score factors of a particular associated score represents an amount added to an information retrieval score for the document associated with the particular associated score.

6. The method of claim 1, further comprising providing the ranked documents to a client.

7. The method of claim 1, wherein each score factor'"'"'s trustworthiness score is determined by a metric that calculates a degree of trustworthiness for the score factor.

8. The method of claim 1, wherein the trustworthiness score of a particular score factor is based on a language of the score factor'"'"'s associated document.

9. The method of claim 1, wherein the trustworthiness score of a particular score factor is based on a similarity between the query and a previous query.

10. The method of claim 1, wherein the trustworthiness score of a particular score factor is based on a country from which the query was submitted and a country in which the documents were created.

31. The method of claim 1, further comprising:

determining that the measure of how widely the trustworthiness scores for each of the different score factors of the plurality of associated scores are distributed in the distribution satisfies a threshold; and

adjusting the value of the at least one score factor of a given associated score whenever the measure of how widely the trustworthiness scores for each of the different score factors of the plurality of associated scores are distributed in the distribution satisfies the threshold.

32. The method of claim 1, wherein the respective trustworthiness score for each score factor is a measure of confidence that the score factor is likely to improve search results provided to a user that submitted the query.

11. A computer program product, encoded on a machine-readable storage device, operable to cause a data processing apparatus to perform operations comprising:

receiving documents responsive to a query, each document having a respective associated score indicative of the document'"'"'s relevance to the query, wherein;

each respective associated score is based on a respective value for each of a plurality of different score factors;

each score factor for a given associated score corresponds to a different criterion than each other score factor for the given associated score; and

each score factor has a respective trustworthiness score that indicates a respective reliability of values determined based on the score factor;

determining, based on the trustworthiness scores for the score factors of a plurality of the associated scores of a plurality of different documents, a value-based distribution of the trustworthiness scores for each of the different score factors of the plurality of associated scores along a dimension based on relative values of the trustworthiness scores, the distribution including trustworthiness scores for multiple score factors for each of the different documents;

adjusting the value of at least one score factor of a given associated score by an amount that is determined based on;

a relative position of the at least one score factor'"'"'s respective trustworthiness score in the distribution with respect to other trustworthiness scores in the distribution; and

a measure of how widely the trustworthiness scores for each of the different score factors of the plurality of associated scores are distributed in the distribution;

adjusting the given associated score based on the adjusted value of the at least one score factor; and

ranking the documents to account for the adjusted associated score.

12. The computer program product of claim 11, wherein the distribution is a frequency distribution and wherein the amount is determined based on a distance between a lowest and a highest trustworthiness score in the distribution and the at least one score factor'"'"'s trustworthiness score.

13. The computer program product of claim 11, wherein the amount is determined based on a number of intervals in the distribution and the at least one score factor'"'"'s trustworthiness score.

14. The computer program product of claim 11, wherein the amount is determined based on a shape of the distribution and the score factor'"'"'s trustworthiness score.

15. The computer program product of claim 11, wherein a first score factor of the plurality of score factors of a particular associated score represents an amount added to an information retrieval score for the document associated with the particular associated score.

16. The computer program product of claim 11, wherein the operations further comprise providing the ranked documents to a client.

17. The computer program product of claim 11, wherein each score factor'"'"'s trustworthiness score is determined by a metric that calculates a degree of trustworthiness for the score factor.

18. The computer program product of claim 11, wherein the trustworthiness score of a particular score factor is based on a language of the score factor'"'"'s associated document.

19. The computer program product of claim 11, wherein the trustworthiness score of a particular score factor is based on a similarity between the query and a previous query.

20. The computer program product of claim 11, wherein the trustworthiness score of a particular score factor is based on a country from which the query was submitted and a country in which the documents were created.

21. A system for processing search results, comprising:

one or more computers; and

a computer-readable medium coupled to the one or more computers having instructions stored thereon which, when executed by the one or more computers, cause the one or more computers to perform operations comprising;

receiving documents responsive to a query, each document having a respective associated score indicative of the document'"'"'s relevance to the query, wherein;

each respective associated score is based on a plurality of different score factors;

each score factor for a given associated score corresponds to a different criterion than each other score factor for the given associated score; and

each score factor has a respective trustworthiness score that indicates a respective reliability of values determined based on the score factor;

determining, based on the trustworthiness scores for the score factors of a plurality of the associated scores of a plurality of different documents, a value-based distribution of the trustworthiness scores for each of the different score factors of the plurality of associated scores along a dimension based on relative values of the trustworthiness scores, the distribution including trustworthiness scores for multiple score factors for each of the different documents;

adjusting the value of at least one score factor of a given associated score by an amount that is determined based on;

a relative position of the at least one score factor'"'"'s respective trustworthiness score in the distribution with respect to other trustworthiness scores in the distribution; and

a measure of how widely the trustworthiness scores for each of the different score factors of the plurality of associated scores are distributed in the distribution;

adjusting the given associated score based on the adjusted value of the at least one score factor; and

ranking the documents to account for the adjusted score.

22. The system of claim 21, wherein the distribution is a frequency distribution and wherein the amount is determined based on a distance between a lowest and a highest trustworthiness score in the distribution and the at least one score factor'"'"'s trustworthiness score.

23. The system of claim 21, wherein the amount is determined based on a number of intervals in the distribution and the at least one score factor'"'"'s trustworthiness score.

24. The system of claim 21, wherein the amount is determined based on a shape of the distribution and the score factor'"'"'s trustworthiness score.

25. The system of claim 21, wherein a first score factor of the plurality of score factors of a particular associated score represents an amount added to an information retrieval score for the document associated with the particular associated score.

26. The system of claim 21, wherein the operations further comprise providing the ranked documents to a client.

27. The system of claim 21, wherein each score factor'"'"'s trustworthiness score is determined by a metric that calculates a degree of trustworthiness for the score factor.

28. The system of claim 21, wherein the trustworthiness score of a particular score factor is based on a language of the score factor'"'"'s associated document.

29. The system of claim 21, wherein the trustworthiness score of a particular score factor is based on a similarity between the query and a previous query.

30. The system of claim 21, wherein the trustworthiness score of a particular score factor is based on a country from which the query was submitted and a country in which the documents were created.