US5943670A
Expired / Lapsed
US
⏱️ 0.0 years remaining
System and method for categorizing objects in combined categories
Assignee: International Business Machines Corp
•
Filing: 1997-11-21
•
Granted: 1999-08-24
•
Expires: 2017-11-21
Abstract
Abstract The present invention is a system and method for determining whether the best category for an object under investigation is a mixture of preexisting categories, and how the mixture is constituted. This invention is useful both for suggesting the need for new categories, and for a fixed set of categories, determining whether a document should be assigned to multiple categories. The objects of the categorization system are typically, but need not be, documents. Categorization may be by subject-matter, language or other criteria. The invention causes extra information to be stored in a category index, so that the determination of mixed categories using the methods presented here is performed extremely efficiently.
Inventors
- 👤 John Martin Prager
Bibliographic Details
Application Number
—
Jurisdiction
US
Data Source
google_patents
Last Synchronized
2026-11-04 18:20
CPC Classifications
G; G06; G06F; G06F16/00; G06F16/30; G06F16/35; Y; Y10; Y10S; Y10S707/00; Y10S707/99931; Y10S707/99933; Y10S707/99935; Y10S707/99936
Claims Analysis
Total Claims: 5
Claim 1
Independent Claim
287 words
A system for combining categories for categorizing objects, comprising: a computer with a memory and a central processing unit (CPU), the memory having one or more documents with one or more document features, the memory further having two or more categories; a categorization process, executed by the CPU, that matches the document features against the categories to determine a score indicating a degree of match that is stored in the memory as a plurality of result entries in a result table with the result entries sorted by score and the categorization process determining a magnitude of a document feature vector; a pair-wise scalar product table, stored in the memory, that has a plurality of pair-wise comparisons of one or more pairs of a first and second category to determine a degree of similarity between the first and second category in the pair, the degree of similarity being stored in the memory, the categorization process generating the degree of similarity, the categorization process further generating a magnitude of a category feature vector for each category that is also stored in memory; a combining process, executed by the CPU, that selects candidate pairs entries in the result table where the entries meet a criterion; and a weight determination process, executed by the CPU, that determines a pair weight that indicates how the first and second category in each candidate pair are weighted with respect to each other to form a virtual mixed category using the degree of similarity between the first and second categories, the magnitude of the category feature vector for the first and second categories, the degree of match between the document and each of the first and second categories, and magnitude of the document feature vector.
Claim 2
Independent Claim
287 words
A system for combining categories for categorizing objects, comprising: a computer with a memory and a central processing unit (CPU), the memory having one or more documents with one or more document features, the memory further having two or more categories; a categorization process, executed by the CPU, that matches the document features against the categories to determine a score indicating a degree of match that is stored in the memory as a plurality of result entries in a result table with the result entries sorted by score and the categorization process determining a magnitude of a document feature vector; a pair-wise scalar product table, stored in the memory, that has a plurality of pair-wise comparisons of one or more pairs of a first and second category to determine a degree of similarity between the first and second category in the pair, the degree of similarity being stored in the memory, the categorization process generating the degree of similarity, the categorization process further generating a magnitude of a category feature vector for each category that is also stored in memory; a combining process, executed by the CPU, that selects candidate pairs entries in the result table where the entries meet a criterion; and a weight determination process, executed by the CPU, that determines a pair weight that indicates how the first and second category in each candidate pair are weighted with respect to each other to form a virtual mixed category using the degree of similarity between the first and second categories, the magnitude of the category feature vector for the first and second categories, the degree of match between the document and each of the first and second categories, and magnitude of the document feature vector.
Claim 3
Dependent Claim
71 words
A system, as in claim 1, where a virtual degree of match between the document and the virtual mixed category is determined using the pair weight, the degree of similarity between the first and second categories, the magnitude of the category feature vector for the first and second categories, the degree of match between the document and each of the first and second categories, and magnitude of the document feature vector.
Claim 4
Dependent Claim
23 words
A system, as in claim 2, where the virtual mixed category and the virtual degree of match are entered in the results table.
Claim 5
Dependent Claim
36 words
A system, as in claim 3, where the virtual mixed category is entered in the results table only if the virtual degree of match is higher than the best degree of match in the results table.
References Cited (Backward Citations)
US5537586A
Cited
USPTO Annuity & Maintenance Windows
| Annuity Stage | Window Opens | Payment Deadline | Surcharge Period Until | Estimated Fee | Status |
|---|---|---|---|---|---|
| 3.5 Years (Year 4) | 2002-08-23 | 2003-02-22 | 2003-08-24 | $2000.00 | Window Closed / Lapsed |
| 7.5 Years (Year 8) | 2006-08-23 | 2007-02-22 | 2007-08-24 | $3760.00 | Window Closed / Lapsed |
| 11.5 Years (Year 12) | 2010-08-23 | 2011-02-22 | 2011-08-24 | $7700.00 | Window Closed / Lapsed |