Implementation of the Support Vector Machine Method in Predicting Student Graduation

Student graduation is an important indicator of the quality of education at a higher education institution. A high graduation rate not only indicates success in the learning process but also has a positive impact on the reputation of the institution. Conversely, a low graduation rate can be a signal of problems that require special attention. Various factors, both academic and non-academic, influence higher education institutions in ensuring timely student graduation. Therefore, we need a method that can accurately predict student graduation to carry out early intervention. This study aims to apply the support vector machine method in predicting student graduation. We chose this method due to its capacity to classify complex data. We use historical student data, such as Semester Achievement Index scores, as input variables to build a prediction model. We evaluate the model using precision, recall, and f1-score metrics. According to the study's findings, the support vector machine model's accuracy level is 71.20%. This method is good at predicting students who graduate with a precision of 95%, recall of 72%, and f1-score of 82%. However, the model's performance in predicting students who failed was less than satisfactory, with a precision of only 17%, a recall of 62%, and an f1-score of 26%. The imbalance in data between passed and failed students contributed to this result. The Support Vector Machine method effectively predicts student graduation for the majority class (passed), but requires special handling of the data imbalance to enhance the accuracy of predictions for the minority class (failed). Universities expect to use the results of this study to carry out early intervention and increase student graduation rates.