TUTORIAL SYSTEM IN LEARNING ACTIVITIES THROUGH MACHINE LEARNING-BASED CHATBOT APPLICATIONS IN THE FIELD OF PHARMACOLOGY EDUCATION

• Main Author: Rizki Fonna, Muhammad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/56247
• Institution: Institut Teknologi Bandung

Education is one of the most important human needs because education ha s an influence on the quality of human resources for the development of the nation and state. The Government of the Republic of Indonesia in this case has facilitated the education of its people through various ways, such as curriculum development, school construction, and so on. However, when viewed from the education index issued by the United Nations Development Program in 2017, Indonesia still occupies the seventh position in ASEAN (out of ten countries), with a score of 0.622 (United Nations Development Programme, 2018). In addition, based on the 2019 Global Talent Competitiveness Index (GTCI), Indonesia is ranked 67th out of 125 countries in the world in terms of the competitiveness of human resources (HR) as seen from several indicators, one of which is education. Referring to the needs and problems of education in Indonesia as mentioned above, policy makers and stakeholders need to make breakthroughs and innovations by utilizing various potentials, especially the use of information technology, so that the educational needs and problems mentioned above can be overcome or at least gradually can be solved. One of the uses of information technology is through the use of chatbots in the field of learning. In this work, a tutorial system is built in learning activities with a machine learningbased chatbot application. The general architecture of a chatbot consists of NLU components, message processor, dialogue management, action executor, and utterance generator. The NLU chatbot capabilities are built on the Rasa framework with intent classification and entity extraction capabilities. The chatbot knowledge using is data obtained from openfda.