Contextual Plagiarism Detection Using Latent Semantic Analysis

Plagiarism means to commit literary theft, where in one steals ideas or words from another and claim it as their own without crediting the source. There are mainly two types of plagiarism possible; text based similarity and context based similarity. Text based similarity works on a word by word match algorithm. Context based similarity tries to find a match between two series of text with respect to implemented meaning or ideas implied. Latent Semantic Analysis(LSA) is a method which tries to find deeper correlation between words used in a document, here semantic implies that the documents try to find a root topic under which both the documents hold true. In this paper LSA is used for context based similarity between documents.