Determining Thesis Title using Winnowing Algorithm

Abstract

This study presents the development and application of a system utilizing the winnowing algorithm to evaluate the suitability of thesis titles based on their relevance to existing research. By generating fingerprint values from input titles and calculating similarity percentages using Jaccard’s Similarity Coefficient, the system provides a quantitative measure of title uniqueness and originality. In this study, the similarity between the proposed title and existing titles was determined to be 38.4%, indicating a low level of overlap and confirming the title’s appropriateness for academic research. The winnowing algorithm enhances the efficiency and accuracy of the title selection process, reducing the risk of redundancy and plagiarism while ensuring alignment with academic standards. This automated approach streamlines decision-making, enabling students to choose focused and high-quality thesis topics. The system’s ability to deliver precise similarity metrics supports academic integrity and promotes originality in research. Future research could explore the integration of advanced algorithms or machine learning to further improve the system’s performance and adaptability across diverse academic domains.