Hub for strategic intelligence

• US 6,609,124 B2
• Filed: 08/13/2001
• Issued: 08/19/2003
• Est. Priority Date: 08/13/2001
• Status: Expired due to Term

• Current Assignee: Kyndryl Incorporated
• Original Assignee: International Business Machines Corporation
• Inventors: Danke, Michael J., Smierciak, Edward L., Pietrzak, Julie J., Tullis, Terry K., Proctor, Larry L., Chow, Amy W.
• Primary Examiner(s): CHOULES, JACK M
• Application Number:
  US09/928,743
• Publication Number:
  US 20030033274A1
• Time in Patent Office: 736 Days
• Field of Search: 707/3, 707/5, 707/6
• US Class Current: 1/1
• CPC Class Codes: G06F 16/355 Class or cluster creation or modification
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

A method and structure for searching a computerized network of databases containing documents uses a web crawler. The web crawler is provided with conceptual guidelines before the searching. The invention summarizes and performs text clustering on the summaries to produce classifications. The text clustering is performed using seeds based on the conceptual guidelines. The invention then provides, through a user interface, the classifications and a query entry to search the classifications and directs (in response to the query entry) the user to one or more of the classifications, such that the user is directed to the classifications (and hyperlinks to the documents) and the user is not provided the documents themselves.

Claims

1. A method of searching databases comprising:

searching a computerized network of databases containing documents using a web crawler, wherein said web crawler is provided with conceptual guidelines before said searching, and said searching produces search results;

summarizing said search results to produce summaries;

performing text clustering on said summaries to produce classifications, wherein said text clustering is performed using seeds based on said conceptual guidelines; and

providing, through a user interface, said classifications and a query entry to search said classifications.

2. The method in claim 1, further comprising providing hyperlinks to said documents in place of providing said documents.

3. The method in claim 1, wherein said summaries are based upon extensible markup language tags associated with said documents.

4. The method in claim 1, wherein links to each of said documents may appear in at least two classes of said classifications.

5. The method in claim 4, further comprising identifying intersections of multiple classes that respond to a user search in said query entry.

6. The method in claim 5, wherein said intersections represent occurrences of different classes which separately return links to a single document in response to said user search.

7. The method in claim 1, wherein said conceptual guidelines refine said searching and said text clustering to direct said classifications to a specific result.

8. A method of searching databases comprising:

searching a computerized network of databases containing documents using a web crawler, wherein said web crawler is provided with conceptual guidelines before said searching, and said searching produces search results;

summarizing said search results to produce summaries;

performing text clustering on said summaries to produce classifications, wherein said text clustering is performed using seeds based on said conceptual guidelines;

providing, through a user interface, said classifications and a query entry to search said classifications; and

directing, in response to said query entry, said user to one or more of said classifications, such that said user is directed to said classifications and not to said documents.

9. The method in claim 8, further comprising providing hyperlinks to said documents in place of providing said documents.

10. The method in claim 8, wherein said summaries are based upon extensible markup language tags associated with said documents.

11. The method in claim 8, wherein links to each of said documents may appear in at least two classes of said classifications.

12. The method in claim 11, further comprising identifying intersections of said classes that both respond to a user search in said query entry.

13. The method in claim 12, wherein said intersections represent occurrences of different classes which separately return links to a single document in response to said user search.

14. The method in claim 8, wherein said conceptual guidelines refine said searching and said text clustering to direct said classifications to a specific result.

15. A program storage device readable by machine tangibly embodying a program of instructions executable by said machine for performing a method of searching databases, said method comprising:

searching a computerized network of databases containing documents using a web crawler, wherein said web crawler is provided with conceptual guidelines before said searching, and said searching produces search results;

summarizing said search results to produce summaries;

performing text clustering on said summaries to produce classifications, wherein said text clustering is performed using seeds based on said conceptual guidelines; and

providing, through a user interface, said classifications and a query entry to search said classifications.

16. The program storage device in claim 15, wherein said method further comprises providing hyperlinks to said documents in place of providing said documents.

17. The program storage device in claim 15, wherein said summaries are based upon extensible markup language tags associated with said documents.

18. The program storage device in claim 15, wherein links to each of said documents may appear in at least two classes of said classifications.

19. The program storage device in claim 18, further comprising identifying intersections of said classes that both respond to a user search in said query entry.

20. The program storage device in claim 19, wherein said intersections represent occurrences of different classes which separately return links to a single document in response to said user search.