A Text Mining Approach for Automatic Selection of Academic Course Topics based on Course Specifications

Topics selection for an educational material can take a lot of manual work. The manual operations can be exhaustive, especially in case of large volume of materials. In order to overcome this problem, we have proposed an automated topic selection approach, which is able to select topics automatically for any educational material with a consideration of achieving course specifications. Our research focused on text mining and n-gram analysis. In addition, filtering criteria was applied to improve efficiency and to eliminate as many irrelevant or non-critical keyphrases as possible. The proposed method was applied on educational materials in Institute of Statistical Studies and Research (ISSR), information technology and Computer Sciences department, Cairo University and in the American University of Beirut (AUB), electrical and computer engineering department, Beirut, Lebanon. The results show that the automatic selection technique is more reliable than the manual selection and reduced a lot of time and effort for course coordinators and teachers to choose the topics that will be taught and discover them automatically.