Acquiring ontological knowledge from query logs

• US 8,051,056 B2
• Filed: 05/29/2007
• Issued: 11/01/2011
• Est. Priority Date: 05/29/2007
• Status: Expired due to Fees

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Sekine, Satoshi, Suzuki, Hisami
• Primary Examiner(s): Hwa; Shyue Jiunn
• Application Number:
  US11/807,410
• Publication Number:
  US 20080301121A1
• Time in Patent Office: 1,617 Days
• Field of Search: None
• US Class Current: 707/705
• CPC Class Codes: G06F 16/367 Ontology

Abstract

Methods are disclosed for acquiring ontological knowledge using query logs. In one embodiment, query logs are first utilized as a basis for identifying important contexts associated with terms belonging to a semantic category. Then, those contexts are as a basis for identifying new terms belonging to the same category or, in another embodiment, as a basis for removing extraneous or obsolete terms identified as being in the same category.

Claims

1. A method of using a computer to modify a list of terms assigned to a semantic category, the method comprising:

receiving an indication of a seed term;

using a computer to identify from a log of queries, a query having a set of query terms that includes at least one instance of the seed term in combination with a context word;

utilizing the computer to automatically identify, from the log of queries, a second query having a set of query terms that does not include the seed termbut does include the context word in combination with a different term, the different term being different than the seed term and the context term; and

utilizing the computer to remove the different term from the list of terms assigned to the semantic category, the different term being removed based on a determination that a significant enough context does not exist between the different term and the seed term, the significance of the context being determined utilizing the formula,

Score(c) =F_type{c}* log(g(c)/C),

where g(c) =F_type{c}/F_inst{c},

where C =F_type{ctopx}/F_inst{ctopx},

where F_type is a frequency of context c in the semantic category,

where F_inst is a frequency of context in an entire data, and

where ctopx is x number of the most frequent contexts.

2. The method of claim 1, wherein receiving the indication of the seed term comprises receiving a user-initiated input.

3. The method of claim 1, wherein a bias is implemented utilizing a mutual information test.

4. The method of claim 1, wherein identifying a query comprises identifying a plurality of queries in the log of queries.

5. The method of claim 1, wherein the log of queries is a log that tracks queries from multiple predefined users.

6. The method of claim 1, wherein the log of queries is a combination of multiple collections of query log information.

7. The method of claim 6, wherein said combination is the combination of multiple collections of query log information obtained from multiple search processors.

8. The method of claim 1, wherein the semantic category is indicative of a person.

9. The method of claim 1, wherein the log of queries is a log of queries from a single predefined user.

10. The method of claim 1, wherein the semantic category is indicative of a location.

11. The method of claim 1, wherein the log of queries is a log of remotely initiated queries.

12. The method of claim 1, wherein the semantic category is indicative of a physical object.

13. The method of claim 1, wherein the log of queries is a log of locally initiated queries.

14. The method of claim 1, wherein the semantic category is customizable.

15. The method of claim 1, wherein the semantic category is subdivided into subcategories.

16. The method of claim 1, wherein the semantic category is subdivided through application of a clustering function.