Automatic relevance and variety checking for web and vertical search engines

1. A method comprising a computer system implementing a testing interface that allows testing of a loss in variety of a set of results returned by a specific search query, wherein implementing the testing interface comprises the steps of:

identifying a set of one or more terms related to the specific search query;

wherein the set of one or more terms are not part of the specific search query;

obtaining a plurality of search results for the specific search query;

determining a first number of results of the plurality of search results;

determining a second number of results for each term of the set of one or more terms related to the specific search query, wherein the second number is the number of results that contain said each term;

determining, for said each term, a third number of historical search queries, and adding said third number to a running total;

wherein said third number is based on a count of those historical search queries that include both the specific search query and said each term;

determining a first ratio of said third number to the running total;

determining a second ratio of the second number to the first number;

determining a relevance of the set of results;

wherein the set of results includes the plurality of search results, and

wherein determining the relevance of the set of results is based at least in part on:

(a) said second ratio, for one or more related terms, of the second number to the first number; and

(b) said first ratio, for one or more related terms, of said third number to the running total.

2. The method of claim 1 ,

wherein a query is a collection of one or more terms submitted by a user for identifying content for which a user wishes to search; and

wherein identifying the set of one or more terms includes examining a query log containing a plurality of queries, wherein examining the query log further comprises:

selecting one or more queries from the query log; and

determining that the one or more queries further contain the specific search query.

3. The method of claim 2 , wherein each query within the query log includes data specifying a time period at which said each query was newly submitted, and wherein selecting one or more queries from the query log includes selecting one or more queries submitted only within a specified time period.

4. The method of claim 1 , wherein identifying the set of one or more terms includes examining session based search data that includes only queries submitted by a single user during a specified period of time.

5. The method of claim 1 , wherein obtaining the plurality of search results for the specific search query further comprises:

performing operations against a data repository, wherein the data repository contains an index of content; and

in response to performing the operations, receiving data representing the plurality of search results.

6. The method of claim 1 , wherein each result of the plurality of search results contains a title and an abstract, and wherein determining said ratio of the second number to the first number includes examining the title and abstract of each result to determine whether said each term is present.

7. The method of claim 5 , wherein each result of the plurality of search results is associated with web content, and wherein determining said ratio of the second number to the first number includes examining the web content associated with each result to determine whether the particular term is contained within the web content.

8. The method of claim 1 , wherein determining the relevance of a set results includes:

determining the relevance of each search result in the plurality of search results based on each term in the set of one or more terms;

generating a total relevance for the set of results based on all the terms in the set of one or more terms; and

the steps further comprise:

generating a value representing the relevance of the set of results; and

reporting the total relevance for the set of results.

9. The method of claim 1 , further comprising:

determining that none of the results in the set of results contain a particular term from the set of one or more terms; and

determining, based at least in part on the ratio of said third number of the particular term to the running total, the loss in variety for the set of results.

10. A method comprising the steps of:

a computer identifying a set of one or more terms related to a specific search query;

wherein the set of one or more terms are not part of the specific search query;

the computer determining, for each term of the set of one or more terms, a third number of historical search queries, and adding said third number to a running total;

wherein said third number is based on a count of those historical search queries that include both the specific search query and said each term;

determining a ratio of said third number to the running total;

the computer obtaining a plurality of search results for the search query;

the computer determining that none of the results in the plurality of search results contain a particular term from the set of one or more terms;

the computer generating a value representing a loss in variety for a set of results;

wherein the set of results includes the plurality of search results, and

wherein generating the value representing the loss in variety for the set of results is based at least in part on the ratio of said third number, for each term of the set of one or more terms, to the running total; and

the computer reporting the value representing the loss in variety for the set of results.

11. The method of claim 10 , wherein identifying the set of one or more terms includes examining session based search data that includes only queries submitted by a single user during a specified period of time.

12. The method of claim 10 , wherein obtaining the plurality of search results for the specific search query further comprises:

performing operations against a data repository, wherein the data repository contains an index of content; and

in response to performing the operations, receiving data representing the plurality of search results.

13. The method of claim 10 , wherein each result of the plurality of search results contains a title and an abstract, and wherein determining that none of the results contain the particular term includes examining the title and abstract of each result to determine that the particular term is not present.

14. The method of claim 12 , wherein each result of the plurality of search results is associated with a content page, and wherein determining that none of the results contain the particular term includes examining the content page associated with each result to determine that the particular term is not contained within the content page.

15. The method of claim 10 , wherein

the query is a collection of one or more terms submitted by a user for identifying content for which a user wishes to search; and

wherein identifying the set of one or more terms related to the search query includes examining a query log, wherein the query log contains a plurality of search queries;

wherein examining the query log further comprises:

selecting one or more queries from the query log; and

determining that the one or more queries further contain the specific search query.

16. The method of claim 15 , wherein:

determining the value representing the loss in variety of the set of results for each term in the set of one or more terms, wherein the loss in variety for said each term of the set of one or more terms is determined by the steps comprising:

assigning a zero value if said each term is found in the set of results;

if said each term is not found in the results, assigning a non-zero value that is a

F

termx

F

total

wherein F total is the total number of historical queries in the query log containing the search query and any term of the set of one or more terms and F termx is the number of historical queries in the query log containing both said each term and the search query; and

generating a total loss in variety of the set of results further comprises adding together the value representing the loss in variety of the set of results for each term in the set of one or more terms.

17. The method of claim 10 , wherein the loss in variety measures a lack of sub-topics found within a set of search results;

wherein the sub-topics are associated with a general topic; and

wherein the sub-topics are comprised of the set of one or more terms, and the general topic is comprised of the terms in the search query.

18. The method of claim 10 , further comprising:

in response to determining the loss in variety for a set of results;

changing the search engine search algorithm;

after changing the search algorithm, performing the specific search query;

computing a loss in variety based on said set of one or more terms; and

determining no loss in variety for the set of results.

19. A computer-readable storage medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the steps of:

identifying a set of one or more terms related to a specific search query;

wherein the set of one or more terms are not part of the specific search query;

obtaining a plurality of search results for the specific search query;

determining a first number of results of the plurality of search results;

determining a second number of results for each term of the set of one or more terms, wherein the second number is the number of results that contain said each term;

determining, for said each term, a third number of historical search queries, and adding said third number to a running total;

wherein said third number is based on a count of those historical search queries that include both the specific search query and said each term;

determining a first ratio of said third number to the running total;

determining a second ratio of the second number to the first number;

determining a relevance of a set of results;

wherein the set of results includes the plurality of search results, and

wherein determining the relevance of the set of results is based at least in part on:

(a) said second ratio, for one or more related terms, of the second number to the first number; and

(b) said first ratio, for one or more related terms, of said third number to the running total;

determining that none of the results in the set of results contain a particular term from the set of one or more terms related to the specific search query; and

determining, based at least in part on the ratio of said third number of the particular term to the running total, a loss in variety for the set of results.

20. The computer-readable medium of claim 19 , wherein identifying the set of one or more terms includes examining session based search data that includes only queries submitted by a single user during a specified period of time.

21. The computer-readable medium of claim 19 , wherein

determining the relevance of a set results includes:

determining the relevance of each search result in the plurality of search results based on each term in the set of one or more terms;

generating a total relevance for the set of results based on all the terms in the set of one or more terms; and

the one or more sequences of instructions causes the one or more processors to further perform the steps of:

generating a value representing the relevance of the set of results; and

reporting the total relevance for the set of results.

22. The computer-readable medium of claim 19 , wherein the one or more sequences of instructions causes the one or more processors to provide:

an implementation of a testing interface that allows testing of at least one of:

(a) the relevance of the set of results returned by said specific search query; and

(b) the loss in variety of the set of results returned by said specific search query.

23. A computer-readable storage medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the steps of:

a computer identifying a set of one or more terms related to a specific search query;

wherein the set of one or more terms are not part of the specific search query;

determining, for each term of the set of one or more terms, a third number of historical search queries, and adding said third number to a running total;

wherein said third number is based on a count of those historical search queries that include both the specific search query and said each term;

determining a ratio of said third number to the running total;

obtaining a plurality of search results for the search query;

determining that none of the results in the plurality of search results contain a particular term from the set of one or more terms; and

generating a value representing a loss in variety for a set of results;

wherein the set of results includes the plurality of search results, and

wherein generating the value representing the loss in variety for the set of results is based at least in part on the ratio of said third number, for said each term of the set of one or more terms, to the running total; and

reporting the value representing the loss in variety for the set of results.

24. The computer-readable medium of claim 23 , wherein identifying the set of one or more terms includes examining session based search data that includes only queries submitted by a single user during a specified period of time.

25. The computer-readable medium of claim 23 , wherein the loss in variety measures a lack of sub-topics found within a set of search results;

wherein the sub-topics are associated with a general topic; and

wherein the sub-topics are comprised of the set of one or more terms, and the general topic is comprised of the terms in the search query.

26. The computer-readable volatile or non-volatile medium of claim 23 , wherein generating the value representing the loss in variety of the set of results further comprises adding together the value representing the loss of variety for said each term in the set of one or more terms that is determined by:

assigning a zero value if said each term is found in the set of results; or

assigning a non-zero value, if the term is not found in the results, wherein the non-zero value is a function of the ratio of the third number of said each term to the running total.