| dc.description.abstract | First year evaluation point cannot show end year evaluation point that generate value, namely Grade Point Average (GPA), because GPA is a cumulative grade that take from first year evaluation point up to end year evaluation value. In this research, we want to know GPA without seeing GPA in every year, just look grade from the first year evaluation point and some courses feature that we can say GPA Prediction. In this research, we use VFI5 algorithm for GPA prediction. VFI5 algorithm have two processes, training process and prediction process. The training process output can describe feature character and relation between features that are courses feature and GPA relation.This research aims to analyze and looks for TPB courses that influence to GPA of Computer Science students, analyze with describing distributing point for each GPA classes, and predict GPA classes based on TPB courses. Training description is very important thing in this research. Training description can explain distributing points of courses grade. Distributing points is important thing for everyone, students and teachers. Students can see courses that influence to their GPA, otherwise teacher can look for the most succesful courses for their students. On General Sociology course, this research show that it is very difficult to get A grade. On File System and Mathematic Discret courses, with 2.00 GPA <2.50 class always have vote for A grade, and little possibility to get D grade. The highest accuration rate of 2001/2002 generation is found from combination between Fisika II, Calculus, and GPA of TPB features. Accuration rate that is resulted by VFI5 Algorithm on prediction process with the features is 70.61%. In otherwise that case is not same with 2002/2003 generation, because accuration rate that resulted is 60.03%. The similiarity between 2001/2002 and 2002/2003 generations is highest accuration rate for each features test showed from Fisika II feature. Accuration rate of fisika II feature on 2001/2002 generation is 65.35%, otherwise on 2002/2003 generation is 65.39%. | id |
