POTENTIAL OF AUTHORSHIP VERIFICATION THROUGH AUTOMATIC SYNTACTIC ANALYSIS

This paper discusses the potential for using Automatic Syntactic Analysis to verify authorship, particularly in educational settings. The focus is on the creation of a tool with the project name AVASA to detect an individual’s unique syntactic signature – a "linguistic fingerprint" shaped by their sentence structure, word choice, and punctuation use. The goal is to ensure that students submit original work, independent of AI-generated content or plagiarism, while also tracking their evolving writing style. The paper outlines existing authorship attribution tools, which analyze features like sentence length, clause structure, and function word usage to distinguish between authors. However, these tools typically focus on static profiles for forensic or cultural purposes, while the suggested tool is intended to handle the dynamic nature of student writing as it evolves throughout their education. Key steps in the authorship verification tool include data collection, feature extraction, and the creation of a measurable syntactic profile. The system begins by gathering a diverse set of text samples from the student, which are analyzed to extract features such as sentence length, complexity, verb tense, and punctuation patterns. These features are quantified and compared with future submissions to update the student’s profile dynamically. If a new submission significantly deviates from the profile, it is flagged as suspicious, potentially indicating plagiarism, AI use, or ghostwriting. The system’s dynamic nature ensures that it adapts to students’ developing writing skills over time, while providing teachers with objective data on linguistic progress. AVASA also is intended to support multi-author projects by segmenting documents and attributing sections to individual contributors based on their syntactic profiles. Challenges include ensuring the first submission is accurate, defining thresholds for syntactic deviations, and handling diverse writing styles. Despite these challenges, AVASA holds promise for improving authorship verification in educational and non-educational settings.