Relevancy-based database retrieval and display techniques

• US 7,376,649 B2
• Filed: 02/02/2006
• Issued: 05/20/2008
• Est. Priority Date: 05/28/2002
• Status: Expired due to Fees

• Current Assignee: IAC Search & Media Incorporated (IAC/InterActiveCorp)
• Original Assignee: IAC Search & Media Incorporated (IAC/InterActiveCorp)
• Inventors: Wang, Wei, Yang, Tao, Gerasoulis, Apostolos
• Primary Examiner(s): Wu; Yicun
• Application Number:
  US11/347,193
• Publication Number:
  US 20060129552A1
• Time in Patent Office: 838 Days
• Field of Search: 707 3- 5
• US Class Current: 707/738
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9538 Presentation of query results
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99934 Query formulation, input preparation, or translation
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

Techniques to retrieve, rank and display data objects retrieved form a database are described. In particular, methods to assign a global ranking value to a data object based on a combination of that object'"'"'s link-based (e.g., vector-space cluster analysis) and text-based (e.g., word frequency) ranks are described. Additional techniques to determine a set of concepts, topics or key words associated with each retrieved data objects are described.

Claims

1. A method to assign topic terms to a plurality of data objects, comprising:

identifying a plurality of data objects;

identifying a plurality of level-1 groups from the plurality of identified data objects, each level-1 group including one or more of the identified plurality of data objects;

assigning one or more topic terms to each level-1 group, each of the one or more topic terms associated with at least one of the identified data objects comprising a level-1 group;

partitioning each level-1 group into one or more level-2 groups, each level-2 group associated with an identified topic term;

merging those level-2 groups having topic terms that overlap by at least a specified amount; and

processing the merged level-2 groups in accordance with their assigned topic terms.

2. The method of claim 1, wherein the act of identifying a plurality of level-1 groups comprises:

identifying one or more clusters within the identified plurality of data objects based on a vector-space clustering technique, each cluster representing a level-1 group; and

identifying an additional level-1 group comprising all those identified data objects not a member of the identified one or more clusters.

3. The method of claim 1, wherein the act of assigning one or more topic terms with each level-1 group comprises identifying significant single word terms associated with the identified plurality of data objects.

4. The method of claim 1, further comprising identifying significant multiple word terms associated with the identified plurality of data objects.

5. The method of claim 1, wherein the act of assigning one or more topic terms to a level-1 group comprises:

identifying terms based on their frequency of occurrence within each data object comprising the level-1 group;

selecting those terms that occur a significant number of times; and

assigning those terms identified as significant as topic terms for that level-1 group.

6. The method of claim 5, wherein the act of selecting those terms that occur a significant number of times comprises selecting those terms that appear more than a specified number of times.

7. The method of claim 5, wherein the act of identifying terms comprises identifying single word terms.

8. The method of claim 7, wherein the act of identifying terms comprises identifying multiple word terms.

9. The method of claim 1, wherein the act of partitioning each level-1 group into one or more level-2 groups further comprises assigning zero or more additional topic terms to each level-2 group.

10. The method of claim 1, wherein the act of merging two level-2 groups comprises merging those level-2 groups into a single level-2 group if the words comprising a topic term common to the two level-2 groups overlap by at least a specified number of words.

11. The method of claim 10, wherein the specified number of words is two-thirds of the total number of words comprising the topic terms.

12. The method of claim 1, wherein the act of processing the merged level-2 groups in accordance with their assigned topic terms comprises displaying at least some of the topic terms to a user such that when selected by the user, identifiers for the data objects associated with the topic term'"'"'s level-2 group are displayed.

13. The method of claim 1, wherein the acts of identifying are performed at run-time.

14. The method of claim 1, wherein the act of assigning, partitioning and merging are performed at run-time.

15. The method of claim 1, wherein the act of identifying a plurality of data objects comprises identifying a plurality of documents.

16. The method of claim 15, wherein the act of assigning one or more topic terms to each level-1 group, comprises associated terms found in a title section of one or more of the plurality of documents.

17. The method of claim 16, wherein the act of identifying a plurality of documents comprises identifying a plurality of web pages.