METHODOLOGIES AND ANALYTICS TOOLS FOR IDENTIFYING WHITE SPACE OPPORTUNITIES IN A GIVEN INDUSTRY

• US 20080235220A1
• Filed: 06/03/2008
• Published: 09/25/2008
• Est. Priority Date: 02/13/2007
• Status: Active Grant

• Current Assignee: GlobalFoundries, Inc.
• Original Assignee: International Business Machines Corporation
• Inventors: Chen, Ying, Spangler, William Scott, Rhodes, James J., Kreulen, Jeffrey Thomas
• Granted Patent: US 9,183,286 B2
• Time in Patent Office: Days
• US Class Current: 707/5
• CPC Class Codes: G06F 16/353 into predefined classes
• CPC Class Codes: G06F 2216/11 Patent retrieval
• CPC Class Codes: Y10S 707/923 Intellectual property

Abstract

A method for analyzing predefined subject matter in a patent database being for use with a set of target patents, each target patent related to the predefined subject matter, the method comprising: creating a feature space based on frequently occurring terms found in the set of target patents; creating a partition taxonomy based on a clustered configuration of the feature space; editing the partition taxonomy using domain expertise to produce an edited partition taxonomy; creating a classification taxonomy based on structured features present in the edited partition taxonomy; creating a contingency table by comparing the edited partition taxonomy and the classification taxonomy to provide entries in the contingency table; and identifying all significant relationships in the contingency table to help determine the presence of any white space.

Claims

1. A method for use with at least one keyword retrieved from a first set of documents related to a predefined subject matter, the method comprising:

constructing snippets from textual material in said first set of documents, each of said snippets including at least one word appearing within a specified text distance of said at least one keyword;

defining a plurality of categories wherein each of said snippets is assigned to one of said categories, each of said categories designated for receiving similar snippets;

creating a respective mathematical model for each of said categories;

analyzing a second set of documents to determine an assignment for each document in said second set of documents to one of said categories, said assignment based on matching each of said documents in said second set of documents to said mathematical model for said assigned category; and

identifying at least one white space in said second set of documents, said at least one white space including fewer than a specified number of documents.

2. The method of claim 1 wherein said second set of documents is directed to patents.

3. The method of claim 1 wherein each of said mathematical models comprises a numeric vector space.

4. The method of claim 1 wherein said analyzing step is performed using information from said predefined subject matter.

5. A method of analyzing predefined subject matter in a patent database, the method being for use with a set of target patents, each of said target patents related to the predefined subject matter, the method comprising:

creating a feature space based on frequently occurring terms found in said set of target patents;

creating a partition taxonomy based on a clustered configuration of said feature space;

editing said partition taxonomy using domain expertise to produce an edited partition taxonomy;

creating a classification taxonomy based on structured features present in said edited partition taxonomy;

creating a contingency table by comparing said edited partition taxonomy and said classification taxonomy to provide entries in said contingency table; and

identifying significant relationships in said contingency table which help determine the presence of a white space.

6. The method of claim 5 comprising assembling the set of target patents by:

retrieving a set of initial patents from the patent database, each initial patent containing at least one word representative of the predefined subject matter;

reviewing said initial patents to derive one or more terms of interest;

retrieving a set of secondary patents, each of said secondary patents containing at least one of said one or more terms of interest; and

merging said set of initial patents with said set of secondary patents to produce said set of target patents.

7. The method of claim 6 wherein said assembling step comprises at least one of a search operation and a query operation.

8. The method of claim 5 wherein said feature space comprises one or more of a structured feature, an unstructured feature, and an annotation.

9. The method of claim 5 wherein said structured feature comprises at least one term from one of said initial documents.

10. The method of claim 8 wherein said unstructured feature comprises a textual segment from one of said initial patents.

11. The method of claim 8 wherein said annotation comprises a structured feature derived from an unstructured feature.

12. The method of claim 5 wherein said step of creating a partition taxonomy comprises:

analyzing each of said target patents to derive a count of occurrences of feature space terms within each said target patent; and

partitioning said set of target patents into a plurality of patent clusters, each of said patent clusters including target patents having occurrences of feature space terms determined to be similar to one another.

13. The method of claim 5 wherein said editing step comprises at least one of the following:

deleting a taxonomy category selected for deletion;

merging two or more taxonomy categories selected for consolidation; and

creating a new taxonomy category.

14. The method of claim 5 further comprising storing new domain knowledge in a knowledge database as a serialized object.

15. The method of claim 5 further comprising performing a time dimension analysis on at least one of said entries in said contingency table.

16. A method for conducting an analysis to provide patent information to a customer, said method being for use with a set of customer patents, each of said customer patents related to business needs of the customer, the method comprising:

creating a first taxonomy for said set of customer patents, said first taxonomy related to technology underlying said customer patents;

creating a second taxonomy for said set of customer patents, said second taxonomy related to an application of said technology; and

creating a contingency table by comparing said first taxonomy to said second taxonomy, said contingency table providing an indication of one or more relationships of interest for the customer.

17. The method of claim 16 wherein said first taxonomy is based on a patent classification system and said second taxonomy is based on web page information.

18. The method of claim 16 further comprising providing a classification model for classifying a given text into a technology category in said second taxonomy.

19. A computer program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform a method comprising the steps of:

assembling a set of target documents using one or more keywords, each of said target documents related to a predefined subject matter;

analyzing each of said target documents to derive a count of occurrences of said keywords in each of said target documents;

creating a first taxonomy for said set of target documents, said first taxonomy related to technology underlying said target documents;

partitioning said set of target documents into a plurality of categories based on words or phrases appearing within a specified distance of one of said keywords; and

accepting input for applying domain expertise to selectively delete, merge, and create categories.

20. The computer program storage device of claim 19 wherein said method further comprises:

creating a second taxonomy for said set of target documents, said second taxonomy related to an application of said technology in said target documents;

creating a contingency table by comparing said first taxonomy to said second taxonomy; and

accepting input for applying domain expertise to recognize one or more relationships of interest in said contingency table.

21. A computer program product for use with at least one keyword retrieval from a set of initial documents related to a predefined subject matter, the program comprising a computer usable medium including a computer readable program, wherein when executed on a computer the computer readable program causes the computer to:

construct snippets from textual material in said first set of documents, each of said snippets including at least one word appearing within a specified text distance of said at least one keyword;

define a plurality of categories wherein each of said snippets is assigned to one of said categories, each of said categories designated for receiving similar snippets;

create a respective mathematical model for each of said categories;

analyze a second set of documents to determine an assignment for each document in said second set of documents to one of said categories, said assignment based on matching each of said documents in said second set of documents to said mathematical model for said assigned category; and

identify at least one white space in said second set of documents, said at least one white space including fewer than a specified number of documents.