Penerapan Natural Language Processing Menggunakan TF-IDF dan N-Gram untuk Layanan Donor Darah

Blood donor information services require a system capable of understanding user questions and automatically providing relevant answers. Utilizing chatbot technology based on Natural Language Processing (NLP) is one solution to support this process. In this study, a chatbot system was developed to support the blood donor information service at Blood for Life (BFL) Makassar. The study aimed to implement the TF-IDF and N-Gram methods in the question-answer matching process and analyze the accuracy of the system's responses. A mixed methods approach was used, with a qualitative approach for system requirements analysis and a quantitative approach for chatbot performance evaluation. System development was conducted using Agile methodology. System testing was conducted using the match accuracy method on 50 test questions representing a variety of user questions related to blood needs and donor registration. The test results showed that the system was able to generate 47 appropriate responses, with an accuracy rate of 94%. These results demonstrate that the implementation of the TF-IDF and N-Gram methods is capable of producing relevant answers to user questions in the blood donor information service at BFL Makassar.