Pemodelan Rekomendasi Topik Skripsi Berdasarkan Performa Akademik Mahasiswa

Abstract

About 68% of students experience delays in completing the thesis, meaning that students in determining research topics are not in accordance with their interests and expertise. This phenomenon will affect student performance in completing educational studies on time. There are various factors that cause this to happen, one of which is that many students do not know their abili-ties. This problem can be overcome by building a classification model that can help students to determine thesis topics based on their abilities. The indicators used in determining the ability of this student use academic data, namely transcripts of course grades taken by students from semester 1 to semester 6. This research uses Feature Selection method and smote before modeling with the SVM and Naive Bayes algorithms to create an optimal model. Based on the analysis results obtained from the application of this algorithm, the best model is made with Support Vector Machine kernel RBF and Naive Bayes type Categorical able to produce the highest accuracy of 96.81% and 83.75%. The courses related to each thesis topic are different. Similar research can be done using other methods that have the same pur-pose as the feature selection method or the SMOTE method.