A SURVEY ON PLAGIARISM DETECTION TECHNIQUES

Plagiarism is a major act of academic dishonesty; hence the detection of plagiarism is very essential. Therefore, Plagiarism Detection is a thriving area of research in Natural Language Processing that involves the identification of misappropriated segments of text and the retrieval of the source of the original text. This paper surveys the types of plagiarism and tasks involved in the detection of plagiarism, and analyses the existing algorithms and methods used in the Plagiarism Detection Framework. The techniques explored in this paper are: Word2vec, Monte Carlo ANN, Candidate Retrieval and Text Alignment, PV-DM and PV-DBOW, RabinKarp Algorithm, IR-based plagiarism detection, LSI, and Joint Word Embedding. This survey concludes that Deep Learning Based Plagiarism Detection methods show a higher accuracy than others. The survey also concludes that the existing methods (excluding LSI), lack the ability to effectively perform Cross-Language Plagiarism Detection.