@thesis{thesis,
		author={Aini  Nadiatul},
		title ={Prediksi kelulusan mahasiswa dengan menggunakan metode SMOTE dan Support Vector Manchine (SVM)},
		year={2021},
		url={http://repository.universitasbumigora.ac.id/1061/},
		abstract={Graduation rate is one indicator of the success of higher education in the implementation of the teaching and learning process. Included in one of the assessment elements for college accreditation is the graduation of students who graduate on time. The problem that has occurred so far is the imbalance in the graduation data for Bumigora University students because students who graduate on time are in the minority compared to students who do not graduate on time as the majority.Thus the purpose of this study is how to classify the graduation of students majoring in computer science between students graduating on time and not graduating on time.
In this study, the data used is data from students who have graduated both on time and not on time in the 2009-2011 graduate year with a total of 265 data. For the completion process of this study using the SMOTE and SVM methods.
The SMOTE method is used to balance the data so as to produce a good classification. While the Support Vector Manchine (SVM) algorithm is for the classification process. The purpose of knowing the performance of the two methods used in analyzing class imbalance, because if using 1 method the results of accuracy, sensitivity, and specificity are not good.
The results of this study using the SVM (Support Vector Manchine) method with a total data of 265 student graduation data, of which 171 data represents the number of students who did not graduate on time, 94 the number of student data who graduated on time. then calculated using the confusion matrix table. Then get 74% accuracy, 52% sensitivity, and 87% specificity. And classification using SMOTE combined with SVM (Support Vector Manchine) using a total of 342 student graduation data, where 171 data is student data who did not graduate on time, 171 is the number of student data who graduated on time, the amount of data has been balanced using SMOTE. get 77% accuracy, 80% sensitivity, and 74% specificity.
Based on the grouping using SVM by balancing the data, the results of the data classification are not balanced where 171 student graduation data as data do not graduate on time while 94 student graduation data as data pass on time, resulting in poor accuracy, sensitivity, and specificity.}
		}