CONTENT BASED RECOMMENDATION SYSTEM USING NAIVE BAYES ALGORITHM (Case study: Vidyanusa Forum)

• Main Author: GALANG PERMANA PUTRA - NIM : 23215391 , ADRIANUS
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/25059
• Institution: Institut Teknologi Bandung

Discussion forum is one of features in e-learning vidyanusa. The forum will function as a discussion platform between students and teachers. Students will be able to raise their questions on certain subjects on this forum and other students or teachers will be able to provide answers. An issue may arise when a question is asked or answered in the discussion forum so frequently, the more information in that subject is made available in the forum. As a consequence, students will find it difficult to find the right discussion they need. <br /> <br /> Information overload brings forth the need of recommendation system. Recommendation system is a software that can be used to produce recommended forums for students. Content based filtering is a method that will be used in this recommendation system. Content based filtering is chosen because this method will only recommend forums to students based on forums that a particular student liked before and not depending on other students. Naive bayes is the algorithm that will used for implemented content based filtering method. Naive bayes is chosen because this algorithm is highly accurate, easy to implement, has a short computation process and can be used for text classification. Variables to be used in naive bayes calculation are subject of forum, sub-subject of forum, title of forum and content of forum. <br /> <br /> From the result of the test conducted, show that the final score of the recommendation calculation is a representation of forum is like or dislike by user. If the final score is greater than 1 is mean that forum is liked by user, however if the final score is lower than 1, is mean that forum is disliked by user. The test result show that top 9 forum provided to the user is a forum that has subject, sub subject, title and content that the users like with a final score is more than 3.9 and the bottom 3, is show forum that have subject, sub subject, title and content that the user dislike with a final score is less than 0.06.