Results returned for list-seeking queries

• US 8,832,057 B2
• Filed: 12/02/2011
• Issued: 09/09/2014
• Est. Priority Date: 12/02/2011
• Status: Expired due to Fees

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Gupta, Anjani, Jain, Ankur
• Primary Examiner(s): Thai, Hanh
• Application Number:
  US13/310,658
• Publication Number:
  US 20130144870A1
• Time in Patent Office: 1,012 Days
• Field of Search: 707/706, 707/710, 707/722, 707/765, 707/769, 707/708, 707/723, 707/728, 707748-749
• US Class Current: 707/706
• CPC Class Codes: G06F 16/334 Query execution G06F16/335 takes precedence
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

List-based search results are generated. According to one technique, items are extracted from multiple resources deemed relevant to a user-submitted search query, and a comprehensive master list of those items is compiled and returned to the query-submitted user in response to his submission. According to another technique, lists of items are identified within such query-relevant resources. For each list-containing resource deemed to be relevant to the query terms, a list is extract from that resource and included within that resource'"'"'s abstract on the search results page returned to the user in response to his submission. Additionally, the resources may be re-ranked for generation of the search results page based on the lists contained within those resources in addition to (or regardless of) occurrences of query terms within those resources.

Claims

1. A computer-implemented method comprising:

in response to receiving the query, selecting, from a search corpus, a set of documents that are relevant to the query;

wherein the set of documents includes at least a first document and a second document;

for each particular document within at least a subset of the set of documents, extracting one or more lists contained within that particular document and adding the one or more lists into a set of lists;

wherein extracting one or more lists includes, extracting a first list from the first document and extracting a second list from the second document;

for each particular list in the set of lists, generating a list score for the particular list based at least in part on a number of occurrences of items in the particular list within other lists in the set of lists;

wherein generating a list score includes generating a first list score for the first list extracted from the first document, and generating a second list score for the second list extracted from the second document;

for each particular document within at least the subset of documents, generating a document score for the particular document based at least in part on the list score generated for at least one list contained within the particular document;

wherein generating a document score includes

generating a first document score for the first document based, at least in part, on the first list score generated for the first list that was extracted from the first document;

generating a second document score for the second document based, at least in part, on the second list score generated for the second list that was extracted from the second document; and

generating a search results page that presents references to selected documents from at least the subset of documents in an order that is based at least in part on the document scores generated for the selected documents;

wherein the method is performed by one or more computing devices.

2. The method of claim 1, wherein generating the list score for the particular list comprises generating the list score based at least in part on a number of the other lists that contain at least one item that is contained in the particular list.

3. The method of claim 1, wherein generating the list score for the particular list comprises generating the list score based at least in part on a number of the items in the particular list that are also contained in at least one of the other lists.

4. The method of claim 1, wherein generating the list score for the particular list comprises:

for each particular item in the particular list, generating an item score for that particular item by counting a number of times that the particular item occurs within the other lists; and

generating the list score for the particular list at least in part by adding together the item scores generated for all items in the particular list.

5. The method of claim 1, further comprising:

prior to extracting the one or more lists, removing, from the set of documents, all documents on which no list is centrally located.

6. The method of claim 1, further comprising:

prior to extracting the one or more lists, removing, from the set of documents, all documents for which no list occupies at least a specified threshold proportion of a total document area.

7. The method of claim 1, further comprising:

prior to extracting the one or more lists, removing, from the set of documents, all documents that do not contain at least one list in which list items are numbered or ranked relative to other list items.

8. The method of claim 1, further comprising:

prior to extracting the one or more lists, removing, from the set of documents, all documents that do not contain at least one list in which at least a specified threshold proportion of list items are also main topics in a specified encyclopedia.

9. The method of claim 1, wherein the step of generating the document score for the result document based at least in part on the list score generated for at least one list contained within the result document comprises setting the document score equal to a highest list score among list scores of multiple lists contained within the result document.

10. The method of claim 1, wherein the step of generating the search results page that presents references to the selected result documents comprises placing, with each reference to each particular selected result document within the search results page, at least a subset of list items contained within a highest-scoring list extracted from that particular selected result document.

15. One or more non-transitory storage media storing instructions which, when executed by one or more computing devices, cause performance of the method recited in claim 1.

16. One or more non-transitory storage media storing instructions which, when executed by one or more computing devices, cause performance of the method recited in claim 4.

17. One or more non-transitory storage media storing instructions which, when executed by one or more computing devices, cause performance of the method recited in claim 10.

11. A computer-implemented method comprising:

selecting, from a search corpus, a set of documents that are relevant to a specified topic;

wherein the set of documents includes at least a first document and a second document;

for each particular document within at least a subset of the set of documents, extracting one or more lists contained within that particular document and adding the one or more lists into a set of lists;

wherein extracting one or more lists includes, extracting a first list from the first document and extracting a second list from the second document;

for each particular list in the set of lists, and for each particular item within that particular list, generating an item score for that particular item based at least in part on a number of occurrences of that particular item within other lists in the set of lists;

selecting a subset of list items from among all list items in lists in the set of lists based at least in part on the item scores generated for those list items;

wherein selecting the subset of list items includes selecting a first subset of items from the first list and selecting a second subset of items from the second list;

ranking the selected list items in the subset of list items based at least in part on the item scores generated for the selected list items, thereby generating a ranked master list for the specified topic; and

storing the ranked master list in association with the specified topic;

wherein the method is performed by one or more computing devices.

12. The method of claim 11, wherein the step of selecting the set of documents that are relevant to the specified topic comprises:

locating, within the search corpus, documents that contain words that are associated with the specified topic.

13. The method of claim 11, wherein the step of generating the item score for the particular item comprises:

adjusting an item score for the particular item based on a rank of an occurrence of that particular item within another list of the set of lists.

14. The method of claim 11, wherein the step of generating the item score for the particular item comprises:

adjusting both (a) an item score for the particular item and (b) an item score for a second item from the same particular list as the particular item in response to a determination that an occurrence of the particular item is either more highly ranked or more lowly ranked than an occurrence of the second item within another list of the set of lists.

18. One or more non-transitory storage media storing instructions which, when executed by one or more computing devices, cause performance of the method recited in claim 11.

19. One or more non-transitory storage media storing instructions which, when executed by one or more computing devices, cause performance of the method recited in claim 13.

20. One or more non-transitory storage media storing instructions which, when executed by one or more computing devices, cause performance of the method recited in claim 14.