An NLP-Based Plagiarism Detection Approach for Moderate-Length Sentences

Abstract

In the realm of plagiarism detection, a key challenge lies in evaluating semantic similarity between obfuscated sentences, particularly within moderate length sentences comprising only 10-20 words. To address this issue, a novel technique called Typed Dependencies Relationship (TDR), rooted in Natural Language Processing, has been introduced for identifying plagiarized content within moderate sentences. This method was tested on existing datasets of sentences and was compared against three leading plagiarism detection methods. The findings indicate that this method showing good results with keep the meaning of sentences with intricate linguistic structures.