Perpustakaan Digital ITB

According to increasing the use of information system, stored data in database also grows rapidly. New data is going up and the old data becomes rarely used. For instance, on Academic Infromation System of Engineering Faculty Mataram University, only active student data has high intensity use, whereas passed student data is inactive. These old data can not be deleted because it is important for validating passed student’s transcript in the future. This study gives an alternative way on the use of old academic data for predicting student successfulness study by using Bayesian Network. It is very valuable for student to determie their learning strategy, taking the next study subjects, and deciding expertise field. Student supervising process is one of the main tasks of lecturer as an academic supervisor that has important factor of student’s success in their study in a university. It is a responsibility of a lecturer to drive students for finishing their study in the best way that consider to previous academic achievement. It can be implemented by Bayesian Network and it may help a student to achieve the best result in the appropriate studying field. On the other hand, this study gives a convenient tool to the lectures for supervising students in their study planning. The idea of this research came from several papers about Intelligent Tutorial System, ITS. ITS has ability to know several pedagogic aspects of a student by their Student Model; therefore, it can give a suitable study method for students. A method that commonly used by ITS for constructing student model is Bayesian Network because it can be performed in unceratin domain. The ITS principle can be adopted for predicting the successfulness study by implementing Student Model of ITS and then students take next study subjects based on this prediction that are put in their study planning card (Kartu Rencana Studi, KRS). This predictor devides successfulness study into three categories, there are high GPA students, medium GPA students, and low GPA students. Overall, accuracy of this predictor is 74,7% and especially accuracy 0f high category is 93,75%, medium is 60%, and low is 83,3%. However, as commonly understood that the result of this predictor is if it is the same or higher then the prediction result; therefore, by this point of view, the accuracy of this predictor is 93,3%.