Using Case Based Reasoning, in the Student Graduation Prediction System (CBR)

Abstract

The performance of an educational institution is considered to have a negative impact due to the high percentage of students who do not graduate on time. Therefore, the number of students must be calculated and anticipated, so this paper needs to propose a model for predicting student graduation rates and predicting the time of graduation for each student. The K-Nearest Neighbor (KNN) method and algorithm were developed using Case-Based Reasoning (CBR). The research was conducted with training data taken from data from students who had graduated, and as a trial, the sixth-semester student data was used. The research attributes required to consist of the grade 1 to semester 5 of the achievement index (IPS) and the students' age. Model testing is carried out with four scenarios where the distribution of training data and testing data is also needed as support. The experimental results show that the stability of the data accuracy value on CBS using the KNN algorithm is found in the 90% training data scenario and 10% test data, namely 90.24% with the closest neighbors K = 5, K = 7, and K = 9. It can be concluded that the model. The resulting prediction system can predict whether students will graduate on time or not. This research will be helpful if implemented in higher education performance to increase student achievement to graduate on time.