Question Classification for e-Learning Using Machine Learning Approach

Along with the use of e-learning, the collection of questions in the database is also increasing. The questions that have been uploaded to the e-learning system can certainly be used repeatedly. The use of repetitive questions will be easy to do with the process of reinvention if the questions have been grouped properly. Unfortunately, grouping questions based on subjects to details on grouping per topic is rarely done in e-learning. This makes the process of tracking the existing problems difficult. In addition, the state of the educational curriculum in Indonesia is often changing. Changes can be changes in whole or in part. This change can make the material content in the subject change order or be eliminated. When these changes occur, the previous problems that can still be used will be difficult to use when curriculum changes occur. Therefore, the large amount of question data and the difficulty in finding the questions again made the researcher to conduct this research. E-learning systems must have machine learning capabilities that are able to classify questions automatically based on topics to sub topics. Examined deeper, each question can have a classification in the form of subject categories, topics, sub topics to the level of difficulty of the questions. In this study, researchers focused on grouping questions based on subject categories and topics. The purpose of this study is to find the right method in building machine learning models in classifying questions according to topic categories and subtopics accurately and quickly. So, the questions in e-learning can be called back exactly as needed automatically.