Enhancing Student Record Searches at Caraga State University using Mistral 7B Model

Abstract

This study evaluates the implementation of a generative AI system using the Large Language Model architecture with the Mistral 7B model to enhance student record searches at Caraga State University-Main Campus, particularly within the College of Computing and Information Sciences (CCIS). The current record retrieval process at Caraga State University, which relies on rigid system dropdown menus and predefined search conditions, is time-consuming, especially during student enrollment evaluations. Faculty members often need more time due to manual evaluations, specifically irregular student records, such as identifying subjects not yet enrolled and listing incomplete grades. This research assesses how generative AI with Mistral 7B can transform the process by allowing faculty to input natural language queries and receive accurate results without navigating multiple modules or systems. Data collection involved manual faculty evaluations and compiling of student profiles and grades from the university's MyWork website. Using the T-test statistical method, the AI-based system with Mistral 7B Model demonstrated a 97% reduction in time spent on record retrieval compared to traditional methods while maintaining 100% accuracy in generating relevant results. This significant improvement in operational efficiency alleviates the faculty workload and enables quicker decision-making during academic evaluations. The findings emphasized the broader applicability of generative AI in academic management, providing insights into how AI can improve the efficiency of administrative processes. This research contributes to the limited body of work on AI-LLM applications in student record management and highlights the potential of AI to streamline institutional operations in higher education. Future studies should explore system scalability and performance across academic programs and datasets.