Image Counting Berbasis Citra Realtime untuk Mengendalikan Arus Lalu Lintas Menggunakan Phyton dan Matlab

Abstrak

Research with the title “Performace Comparison of Naive Bayes and Support Vector Machine for Herregistration Prediction” aims to determine the comparison of naive bayes and support vector machine in herregistration prediction with accuracy parameters and AUC using test scenarios with split validation, which will later be used as a reference for parties universities to carry out policies for students especially those who have the potential to experience non-registration. In this study, only herregistration prediction of prospective new students at the computer science faculty for the class of 2015 to 2017 using of the naive bayes algorithm and support vector machine. The accuracy obtained in the Naive Bayes method is 93.54% and AUC 0.946 while the support vector machine method is 92.67% and AUC 0.877 uses the RBF kernel with parameter cost (C) 1.0 and Epsilon 0.0. In addition, performance accuracy and AUC are very influential when deleting one of the variables used.