COMPARISON OF ALGORITHM BETWEEN CLASSIFICATION & REGRESSION TREES AND SUPPORT VECTOR MACHINE IN DETERMINING STUDENT ACCEPTANCE IN STATE UNIVERSITIES

Abstract

Higher education entrance selection activities are intended to obtain superior student candidates. The opportunity to take part in the selection is given to all high school graduate students and equivalent. The student entrance test at PTN consists of three types of selection routes, namely the SNMPTN or invitation route, the SBMPTN, and the independent examination held by state universities. Starting from the dataset, data selection was carried out from 143 students' data and 7 attribute selections were carried out using preprocessing using data transformation first. The aim of using data transformation is to simplify the data training process for MAN 1 students in Cirebon. Preprocessing for prediction of classification results, accuracy of testing data for 143 students is implemented in the program and the resulting calculation process will be more efficient. After going through the preprocessing stage, the data is divided into training data and testing data using 10-fold cross validation. Next, for the classification process, a comparison of two methods will be used, namely for the first method using CART, the second method using SVM by adding Gain ratio weighting. The results of the research show that in the first experiment the researcher carried out a comparative trial of cross validation and classification performance and used the CART and SVM algorithms. The results comparison using the CART algorithm gets an accuracy of 86.10% and the SVM algorithm method for classifying students entering PTN was 86.71%.