A Malicious Domain Detection Method of Cryptomining Based on Deep Learning

• Main Authors: Zheng, Wei, Huang, Xuange, Xie, Renchao, Tang, Qinqin, Huang, Tao
• Format: Conference Proceeding
• Language: English
• Subjects: blacklist; Blocklists; Costs; cryptomining; Deep learning; Design methodology; detection; Electricity; malicious domain name; Malware; Privacy
• ISSN: 2837-7109
• DOI: 10.1109/ICCC59590.2023.10507257

With the skyrocketing prices and soaring trading volumes of Bitcoin and other cryptocurrencies, the harms caused by malicious cryptomining activities are also increasing. Hackers are increasingly utilizing malicious software to conduct network attacks for cryptocurrency mining, posing threats not only to user privacy but also leading to the consumption of computing resources and increased electricity costs. Despite these challenges, existing detection methods, such as using blacklists to protect users' browser antivirus programs, only offer partial solutions to this problem, as attackers can easily bypass their detection by frequently changing their domain names using domain generation algorithms. To address these issues, this paper employs deep learning technology and designs a method for detecting malicious cryptomining domains. This method combines blacklist detection with Long Short-Term Memory (LSTM) and is capable of identifying malicious domains from a large number of domain samples. Experimental results demonstrate that the proposed method produces excellent classification and detection outcomes.