Patent · US Expired

Method of automatically classifying a text appearing in a document when said text has been converted into digital data

US6295543A · kind A · utility

22Cited by

9References

7Claims

0Family size

Assignee

Siemens Aktiengesellschaft · DE

Inventors

Hans-Ulrich Block · Birkenfeld, DE
Thomas Bruckner · Günzburg, DE

Key dates

Filing date	Oct 5, 1998
Grant date	Sep 25, 2001
Priority date	—
Expiry date	Oct 5, 2018

Classification

Technology area (CPC Y)Emerging Cross-Sectional Technologies
CPC primaryY10S707/99935
WIPO fieldComputer technology
WIPO sectorElectrical engineering

Abstract

The text to be classified is compared with the contents of a relevance lexicon in which the significant words of the texts to be classified per text class and their relevance to the text classes is stored. The fuzzy set is calculated which specifies for the significant words of the text to be classified, their occurrence per text class and their relevance to the text class. A probability calculation is used for determining the probability with which the fuzzy set occurs per class for the corresponding class. The class having the highest probability is selected and the text is allocated to this class.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.