<![CDATA[Abstrak. Fenomena malware yang terus bertumbuh pada sistem Android menjadikan peneliti berfokus untuk menganalisa malware dengan memanfaatkan teknologi kecerdasan buatan. Tujuan dari penelitian ini adalah menganalisa file-file APK (Application Package Kit) Android dengan mengklasifikasi keluarga malware. File malware akan dijadikan dataset untuk dilakukan training menggunakan algoritma pembelajaran mesin. Pembelajaran mesin yang digunakan adalah Naïve Bayes, Logistic Regression dan Support Vector Machine. Pengukuran performansi dan akurasi juga disajikan dalam perbandingan antara algoritma Naïve bayes, Logistic Regression dan Support Vector Machine yang merupakan algoritma Machine Learning dan bagian dari kecerdasan buatan. Hasil uji akurasi menunjukkan algoritma Naive Bayes mampu mengklasifikasi keluarga malware dengan tingkat akurasi 97.75%, sedangkan algoritma Logistic Regression akurasinya 88.75% dan akurasi Support Vector Machine mencapai 96,75%. Meskipun akurasi tidak setinggi penelitian sebelumnya, teknik analisa statis dengan fitur Permission dan fitur Intent cukup sederhana untuk mendeteksi file APK Android adalah malware atau bukan malware.Kata kunci: Malware Android; Naïve Bayes; Logistic Regression; Support Vector Machine Abstract. The phenomenon of malware that continues to grow on the Android system makes researchers focus on analyzing malware by utilizing artificial intelligence technology. The purpose of this research is to analyze Android APK (Application Package Kit) files by classifying malware families. The malware files will be used as a dataset for training using machine learning algorithms. The machine learning used is Naïve Bayes, Logistic Regression and Support Vector Machine. Performance and accuracy measurements are also presented in a comparison between the Naïve Bayes algorithm, Logistic Regression and Support Vector Machine which is a Machine Learning algorithm and part of artificial intelligence. The accuracy test results show that the Naive Bayes algorithm is able to classify malware families with an accuracy rate of 97.75%, while the Logistic Regression algorithm has an accuracy of 88.75% and an accuracy of Support Vector Machine reaches 96.75%. Although the accuracy is not as high as previous studies, the static analysis technique with the Permission feature and the Intent feature is quite simple to detect Android APK files are malware or not malware.Keyword: Malware Android; Naïve Bayes; Logistic Regression; Support Vector Machine]]>
Copyrights © 2022
