Efficacy of Topological Data Analysis for Contextual Plagiarism Detection: A Literature Analysis

Contextual plagiarism occurs when ideas, thoughts, or concepts are taken from a source and presented by altering or rephrasing the content while maintaining the original structure without giving proper acknowledgement. The detection of contextual plagiarism is crucial in multiple domains such as academic writing, journalism, and content creation, to uphold intellectual integrity and prevent misappropriation of ideas. Automated plagiarism detection tools are used to identify contextual plagiarism, which examines the similarities in text by considering both verbatim matching and contextual-semantic similarities. This article reviews contextual plagiarism and emphasizes the challenges associated with its detection in various academic and professional settings. The article evaluates traditional plagiarism detection methods and explores the possibility of the emerging research area known as topological data analysis. Traditional plagiarism detection tools focus on extracting verbatim matches and capturing only the local context window rather than the global context structure. By applying topological data analysis techniques, researchers can map textual data to a topological space by capturing the global structure to reveal conceptual flow and argumentation patterns, which are crucial for contextual plagiarism detection. The study hypothesizes that the use of topological data analysis techniques can improve the detection of contextual plagiarism. Despite its potential, to date, no work has explored the efficacy of using topological data analysis techniques in its detection. The findings explore the possibilities of topological data analysis techniques along with traditional methods to enhance contextual plagiarism detection systems.