Assistant Professor, Golestan University, Gorgan, Iran
Abstract
Malware detection has become one of the important research areas in recent years. Despite the efforts made in this field, it is still possible to improve the accuracy of the presented models in detecting different types of malware. In this paper, an attempt has been made to adjust the parameters of deep learning networks based on the use of four meta-heuristic algorithms including bee colony, firefly, imperialist competitive and bat algorithm. Two deep learning networks including Convolutional Neural Network (CNN) and Long-Short Term Memory (LSTM) with different number of layers have been used to achieve maximum detection accuracy. The experimental results show that in the best case, using the firefly algorithm and LSTM learning network with 3 layers, batch size 40, filter size 5*5 and number of periods 300 leads to an accuracy of 99.997% in malware detection, which is significantly superior compared to other presented works. Other performance evaluation parameters also show very good values, which indicates the superiority of the proposed method over other similar works.
Tajari Siahmarzkooh, A. (2025). Malware Detection Based on Firefly Meta-Heuristic Algorithm and Long-Short-Term Memory Learning Network. Electronic and Cyber Defense, 13(3), -.
MLA
Aliakbar Tajari Siahmarzkooh. "Malware Detection Based on Firefly Meta-Heuristic Algorithm and Long-Short-Term Memory Learning Network", Electronic and Cyber Defense, 13, 3, 2025, -.
HARVARD
Tajari Siahmarzkooh, A. (2025). 'Malware Detection Based on Firefly Meta-Heuristic Algorithm and Long-Short-Term Memory Learning Network', Electronic and Cyber Defense, 13(3), pp. -.
VANCOUVER
Tajari Siahmarzkooh, A. Malware Detection Based on Firefly Meta-Heuristic Algorithm and Long-Short-Term Memory Learning Network. Electronic and Cyber Defense, 2025; 13(3): -.
)
