HCCI eScholar: An Online Research Database Archival System with TF-IDF and Cosine Similarity Algorithms

Abstract

The sudden increase in academic research outputs in higher education institutions presents challenges in file management, including redundancy of research topics, limited cross-department visibility, and difficulty in detecting similarities. The study developed HCCI eScholar: An Online Research Database Archival System with TFIDF and Cosine Similarity Algorithms, a centralized, searchable, and secure digital repository designed to archive undergraduate theses, capstone projects, faculty research, and institutional publications at a higher education institution in the Philippines. The system uses cross-department validation to prevent duplicate research submissions, utilizing an internal similarity-checking system that uses TF-IDF vectorization and the Cosine Similarity algorithms to check the originality of submitted abstracts without needing any third-party plagiarism detection services. To make sure that data is safe and can be stored for a long time, the system uses database normalization, encrypted PDF storage, structured directory management, and backup methods in a cPanel hosting environment. ISO/IEC 25010 Software Quality Model was used to test the system, and produced positive results. The system received excellent remarks across key software quality criteria, indicating that it satisfies essential functional and quality requirements. The results indicate that HCCI eScholar effectively enhances research management, reduces redundancy, and supports academic integrity by means of secure centralized archiving and similarity detection.