Prediction Model For Students' On-Time Graduation Using Algorithm Support Vector Machine (SVM) Based  Particle Swarm Optimization (PSO)

One of the indicators in assessing the tridharma of higher education outcomes and achievements is students' timely graduation and one indicator of the success of the higher education system is timely graduation. However, some students cannot complete their studies on time. In the process of completing the study, several problems emerged, one of which was in completing studies on time, there are problems that arise, such as students who are still repeating because there are grades that have not passed in the course, the Grade Point Average (GPA) is still lacking, the Semester Achievement Index (SAI) is still below the minimum, the total number of semester credit units (total credits) which still have not reached the minimum limit, then the number of active lecture statuses still exceeds 8 semesters, so this problem will have an impact on the accuracy of student study graduate data, where the target performance indicator for graduates is on time The student's target of graduating on time has not been achieved. The factors that cause the students not to graduate on time are not known. In identifying the problem, it was found that the Study Program does not have adequate information regarding the potential for students to graduate on time and the limitations of the study program in assisting students in completing on-time graduation The method used to solve this problem is by creating a prediction model for students' on-time graduation, so that students can receive adequate information regarding the potential for graduating on time. This research aims to create a prediction model for graduating on time using the Support Vector Machine (SVM) method based on Particle Swarm Optimization (PSO) with feature selection. information gain, so that the attributes selected and used are Semester Achievement Index 1, Semester Achievement Index 2, Semester Achievement Index 3, Semester Achievement Index 4, Grade Point Average (GPA) 1, Grade Point Average (GPA) 2, Grade Point Average (GPA) 3, Grade Point Average (GPA) 4, Semester Credit Units 1, and Semester Credit Units 4. The results in this study obtained accuracy values ​​of 0.799, precision 0.851, recall 0.605 and AUC 0.86