Authors:
Yoshimi Suzuki
and
Fumiyo Fukumoto
Affiliation:
University of Yamanashi, Japan
Keyword(s):
Thesaurus, Patent, Document classification.
Related
Ontology
Subjects/Areas/Topics:
Applications
;
Artificial Intelligence
;
Knowledge Engineering and Ontology Development
;
Knowledge-Based Systems
;
Natural Language Processing
;
Pattern Recognition
;
Symbolic Systems
Abstract:
This paper presents amethod for patent document classification by using an expanded technical term thesaurus. For classifying structural documents such as patent documents, structural information is very useful. However, if we use documents divided into several applicant tags, the number of words are limited. For example, ‘Title of invention’ tag is very important for patent document classification. However, the number of words in the tag is very few. Therefore, in order to deal with this problem, we employ two methods. One is to classify applicant tags into semantic tags, the other is word expansion using an expanded technical term thesaurus.
For thesaurus expansion, our system integrates technical terms into a thesaurus using patent documents. The classification results showed the method using the expanded thesaurus was better than that without thesaurus. Although our method is very simple, it is comparable to other methods. These results suggest that thesaurus and our method to ex
pand thesaurus can be useful for patent document classification.
(More)