Legal Text Document (i.e. Trademark Ordinance) Categorization using Ontology

Abstract

Text categorization has been done vastly in different domains such as sp011s, food, law, politics, fashion etc as they are enrich with a lot of useful information. Legal text categorization has been performed by many authors to support the law professional for their preparation before initiating or defending their cases. Text classification is that type of machine lea1nillg technique mostly applied for legal text document classifications as it reduces the review time and the expense of the corpus documents as it categorizes the retrieved document list. In the previous research, the authors ha·,1e pre,ented legal text categorization process by dividing the document text in different categories which include deontic models such as ol;ligation defini/tion, provision, prohibition and permission. Many classifiers were proposed with around 94% accuracy rate calculated. The literature explains that Machine learning is a preferable for text classification as it build the classifier by learning the characteristics of labelled examples automatically. But the drawback of this approach is that it requires large dataset of text documents for learning and training of model. In this thesis document, a semantic based approach is proposed for the legal text categorization in support of law practitioners for understanding the laws related to their cases before court ruling.