Klasifikasi Deteksi Malware Menggunakan Metode Deep Neural Network (DNN)

Research on Malware Detection Classification Using Deep Neural Networks (DNN). With
the rising threat of cyberattacks—such as the 30% increase in global cyber incidents reported
by Check Point Research in 2024, and the average cost of a data breach reaching USD 4.88
million according to IBM—malware detection has become crucial. This study aims to
examine the effectiveness of DNN in detecting and classifying various types of malware,
including new variants and zero-day threats, as well as comparing its performance with
signature-based methods and other machine learning approaches. The dataset used consists
of 10,868 samples, including 9,339 malware and 1,529 benign files, covering features such
as section entropy, API calls, and DLL imports. The results show that the LSTM-based DNN
model achieved an accuracy of 98.84%, precision of 99.44%, recall of 99.27%, and F1-score
of 99.36%. This research demonstrates that DNN provides a more adaptive approach
compared to traditional methods in addressing the evolving challenges of malware threats.