Article: Performance evaluation of various deep convolutional neural network models through classification of malware Journal: International Journal of Information and Computer Security (IJICS) 2023 Vol.21 No.3/4 pp.414 - 435 Abstract: Malware, a collective name for malicious programs, is a piece of software, system or scripts, causing damage to the system. Lately, use of internet has favoured criminal activities like malware assaults. Hence, malware classification comes in the first line of defense. Machine learning (ML) techniques have drawn attention to malware classifiers over all other techniques in the last decade. Very little investigation highlights the results of the existing studies in malware classification using ML approach. The progress is slow due to difficulties of developing a deep learning system: dataset collection, labelling, feature extraction, model construction, training and testing the models, and evaluation. A systematic way of summarising the current knowledge also lacks in latest methods. This study utilises a systematic literature review and presents implementation of different CNN models for malware classification into their respective families. Its objective is to analyse the most popular architectures and evaluate their results. Inderscience Publishers - linking academia, business and industry through research

Title: Performance evaluation of various deep convolutional neural network models through classification of malware

Authors: Zareen Tasneem; Maria Afnan; Md. Arman Hossain; Md. Mahbubur Rahman; Samrat Kumar Dey

Addresses: Department of Computer Science and Engineering, Military Institute of Science and Technology, Mirpur, Dhaka-1216, Bangladesh ' Department of Computer Science and Engineering, Military Institute of Science and Technology, Mirpur, Dhaka-1216, Bangladesh ' Department of Computer Science and Engineering, Military Institute of Science and Technology, Mirpur, Dhaka-1216, Bangladesh ' Department of Computer Science and Engineering, Military Institute of Science and Technology, Mirpur, Dhaka-1216, Bangladesh ' School of Science and Technology, Bangladesh Open University, Gazipur-1705, Bangladesh

Abstract: Malware, a collective name for malicious programs, is a piece of software, system or scripts, causing damage to the system. Lately, use of internet has favoured criminal activities like malware assaults. Hence, malware classification comes in the first line of defense. Machine learning (ML) techniques have drawn attention to malware classifiers over all other techniques in the last decade. Very little investigation highlights the results of the existing studies in malware classification using ML approach. The progress is slow due to difficulties of developing a deep learning system: dataset collection, labelling, feature extraction, model construction, training and testing the models, and evaluation. A systematic way of summarising the current knowledge also lacks in latest methods. This study utilises a systematic literature review and presents implementation of different CNN models for malware classification into their respective families. Its objective is to analyse the most popular architectures and evaluate their results.

Keywords: malware; classification; image; deep learning; convolutional neural network.

DOI: 10.1504/IJICS.2023.132767

International Journal of Information and Computer Security, 2023 Vol.21 No.3/4, pp.414 - 435

Received: 25 May 2022
Accepted: 08 Oct 2022
Published online: 09 Aug 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Performance evaluation of various deep convolutional neural network models through classification of malware

Keep up-to-date