Perbandingan Kinerja Kernel RBF dan Linear pada Algoritma Support Vector Machine (SVM) untuk Prediksi Serangan Ransomware Locker
Nurul Afifah, ali bardadi
Abstract
Ransomware is one type of Malware that is very dangerous. The way Ransomware works is to infiltrate and then be able to duplicate files that run on the Windows operating system. Ransomware encrypts and locks onto a file system very quickly. Ransomware will give a notice by telling how to open the file system by making a payment via cryptocurrency. Ransomware is very detrimental to users. One type is Ransomware Locker. Ransomware locker is very dangerous, therefore it is necessary to make a prediction model about Ransomware Locker attacks. Several Machine Learning methods are able to solve prediction system problems, one of which is the Support Vector Machine (SVM) method. To get the best model in prediction, it is necessary to do a comparison between two SVM kernels namely RBF and Linear. As a result, the SVM RBF method was able to produce excellent performance, namely 93.27% for AUC Train and 93.41 for AUC Test