Chatbot for Digital Library Information System (Digilib) Using Retrieval-Augmented Generation (RAG)

Abstract

Digital libraries provide broad access to academic literature, yet finding relevant information remains a challenge for users. This study aims to develop and implement a chatbot based on Retrieval-Augmented Generation (RAG) technology to enhance the efficiency of information retrieval and academic reference searches within the Digital Library Information System (Digilib). The methodology includes designing a chatbot with an RAG architecture, integrating it with the digital library database, and evaluating its performance using ROUGE metrics to measure the relevance and accuracy of the chatbot's responses. The results indicate that the developed chatbot provides more relevant responses compared to conventional search methods, with ROUGE scores demonstrating improvements in accuracy and contextual relevance. For future development, it is recommended to enhance the model through fine-tuning, integrate external reference sources, and optimize contextual understanding to improve response quality.