Cyber Security dan Forensik Digital (CSFD)
Vol. 5 No. 1 (2022): Edisi Mei 2022

ANALISIS STATIS DETEKSI MALWARE ANDROID MENGGUNAKAN ALGORITMA SUPERVISED MACHINE LEARNING

Raden Budiarto Hadiprakoso (Poltek Siber dan Sandi Negara)
Wahyu Rendra Aditya (Unknown)
Febriora Nevia Pramitha (Unknown)



Article Info

Publish Date
29 Nov 2022

Abstract

Saat ini, pertumbuhan sistem operasi Android di perangkat smartphone sedang populer. Bagaimana pun, dibalik popularitas tersebut platform Android juga menjadi target peluang kejahatan dunia maya terhadap ancaman keamanan siber seperti malware. Mengidentifikasi malware ini sangat penting untuk menjaga keamanan dan privasi pengguna. Karena proses identifikasi malware yang semakin rumit, maka perlu digunakan machine learning untuk klasifikasi malware. Penelitian ini mengumpulkan fitur analisis statis dari aplikasi aman dan berbahaya. (malware). Dataset yang digunakan pada penelitian adalah dataset malware DREBIN yang merupakan dataset malware yang tersedia secara publik. Dataset tersebut terdiri dari fitur API CALL, system command, manifest permission, dan Intent. Data tersebut kemudian diproses menggunakan berbagai algoritma supervised machine learning di antaranya Support Vector Machine (SVM), Naive Bayes, Decision Tree dan K-Nearest Neighbors. Kami juga berkonsentrasi pada memaksimalkan pencapaian dengan mengevaluasi berbagai algoritma dan menyesuaikan beberapa konfigurasi untuk mendapatkan kombinasi terbaik dari hyper-parameter. Hasil eksperimen menunjukkan bahwa klasifikasi model SVM mendapatkan hasil terbaik dengan mencapai akurasi 96,94% dan nilai AUC (Area Under Curve) 95%. Kata kunci: android, malware, machine learning, deteksi malware, analisis statis ------ Currently, the growth of the Android operating system on smartphone devices is popular. However, behind this popularity, the Android platform is also a potential target for cybercrimes against cybersecurity threats such as malware. Identifying this malware is critical to maintaining user security and privacy. Because the malware identification process is getting more complicated, it is necessary to use machine learning for malware classification. This study collects the static analysis features of safe and malicious applications. (malware). The dataset used in this study is a DREBIN malware dataset which is a publicly available malware dataset. The dataset consists of the CALL API features, system commands, manifest permissions, and Intents. The data is then processed using various supervised machine learning algorithms including Support Vector Machine (SVM), Naive Bayes, Decision Tree and K-Nearest Neighbors. We also concentrate on maximizing performance by evaluating various algorithms and adjusting some configurations to get the best combination of hyper-parameters. The experimental results show that the SVM model classification gets the best results by achieving an accuracy of 96.94% and an AUC (Area Under Curve) value of 95%. Keywords: android, malware, machine learning, malware detection, static analysis

Copyrights © 2022






Journal Info

Abbrev

cybersecurity

Publisher

Subject

Computer Science & IT

Description

Cyber Security dan Forensik Digital (CSFD), published by Center of Cyber Security Sunan Kalijaga, Faculty of Science and Technology - UIN Sunan Kalijaga Yogyakarta. This journal published twice a year, May and November, in the fields of Cyber Security and Digital Forensics. ...