Feature-Level Comparison of Static and Dynamic Analysis for Malware Detection and Family Classification Zararli Yazilim Tespiti ve Aile Siniflandirmasi için Statik ve Dinamik Analizin Özellik Düzeyinde Karşilaştirilmasi
34th Signal Processing and Communications Applications Conference, SIU 2026, İstanbul, Türkiye, 7 - 10 Temmuz 2026, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Doi Numarası: 10.1109/siu71813.2026.11636628
- Basıldığı Şehir: İstanbul
- Basıldığı Ülke: Türkiye
- Anahtar Kelimeler: dynamic analysis, feature comparison, malware classification, malware detection, static analysis
- Yıldız Teknik Üniversitesi Adresli: Evet
Özet
The number and diversity of malware targeting information systems are steadily increasing, making the effective detection and classification of such threats critical. In this context, understanding the impact of different feature types on malware detection and family classification is of great importance. In this study, static features based on file structure and opcode sequences, as well as dynamic features reflecting runtime behaviors, were evaluated separately, and their performance on both problems was compared using various machine learning algorithms. The results demonstrate that features provide varying levels of discriminative power and that proper feature selection is crucial for model performance. Furthermore, it was shown that static analysis-based features can achieve high performance compared to dynamic features when combined with appropriate algorithms.