Using Knowledge From Wikipedia To Improve Document Classification

The popularity of Wikipedia has grown and currently ranks fifth in overallglobal web traffic. Web users looking for information on any topic willlikely come across a Wikipedia article fairly quickly. The intention of theresearch work in this paper is to understand how document classificationcan be improved using Wikipedia. A variety of literatures have beenreviewed to study details of the functions of Wikipedia and the effect it canhave on document classification. We proposed a way to enrich the bag-of-words representation of processed documents using the semanticknowledge from Wikipedia. Keeping this in mind we fail to reject1)Document classification can be improved using Wikipedia,2) Algorithmmodels are the factors that contribute to document classification, and3) thecurrent position document classification is good. Although the proposedmethods need deeper evaluation, they seem to suppress short-textdocuments limitations that BOW model has.