Indexed by:期刊论文

Date of Publication:2012-05-15

Journal:Journal of Computational Information Systems

Included Journals:EI、Scopus

Volume:8

Issue:10

Page Number:4339-4346

ISSN No.:15539105

Abstract:Currently, patent data have gained increasing attention in the data mining area. However, the traditional method was set by many factors in patent classification, such as professional terms and high dimensions. In this article, we propose a word sense disambiguation model, which would effectively reduce the structure of the traditional machine word. The experimental result shows that the semantic disambiguation and feature reduction strategy can effectively improve the classification accuracy. ? 2011 by Binary Information Press.