Closed Domain Bangla Extrinsic Monolingual Plagiarism Detection and Corpus Creation Approach

Plagiarism is an act of presenting other's works or ideas as someone's own. In recent days, online publishing of text through internet media is rapidly increasing and plagiarism is one of the most dangerous side effects of that. Plagiarism detection in academic institutions has become a regular feature in recent years. There are several approaches available for detecting plagiarism in the English language and other western languages. However, there is no plagiarism detection tool available that supports plagiarism detection in Bangla language where Bangla is one of the most spoken languages in the world and widely used on the internet. Therefore, we have developed a first-ever plagiarism detection tool for Bangla language. To reduce the domain complexity at this stage, we have reduced our domain to textbooks only. We have also constructed a corpus which consists of all books of the National Curriculum and Textbook Board (NCTB) of Bangladesh from class I-XII. Our proposed approach shows an accuracy of 96.75% at this stage.