Chatbot Implementation for Automating Services at University Library of Muhammadiyah Surakarta with RASA Framework

Improving the accreditation and the quality of services in higher education, the Library of Universitas Muhammadiyah Surakarta (UMS) faces challenges in providing fast, efficient, and flexible access to information for students, especially in this digital era that increasingly demands service automation. This study aims to develop a technology-based chatbot integrated into the library's website to automate information services, such as operational schedules, book collections, and borrowing procedures, thereby enhancing user experience. The chatbot was designed using the RASA framework and employed the CRISP-DM method, focusing on model development, data processing, and system performance evaluation. Testing was conducted using four main parameters: accuracy, recall, precision, and F1-score, with a confusion matrix approach through cross-validation testing to achieve the best model. The test results show that the chatbot not only provides accurate responses to inquiries related to library services but also represents an innovative step in leveraging technology to improve accessibility and efficiency of information services in an academic environment.