Today, malware appears in large numbers in online social networks and causes all kinds of damage. Typically, social networking services consist of connections between different user systems. As a result, malware can easily be transferred between systems of different users. Therefore, malware detection has become a popular topic in social network analysis.
Jia et al. (2018) proposed a novel model HSID to investigate the effects of heterogeneous infection rate. They concluded that in networks with lower and higher percentages of bidirectional edges, infection rates tended to follow a power-law and normal distribution, respectively. In addition, heterogeneous relationships lead to higher heterogeneity of infection rates in directed networks and lower heterogeneity in undirected networks compared to constant models. Besides, heterogeneous relations can improve the propagation range in directed networks, while heterogeneous security awareness can improve the range in both directed and undirected networks, and higher heterogeneity means greater improvement.
Reddy, Kolli, and Balakrishnan (2021) propose an approach that leverages community detection and social network analysis to apply malware detection and classification. They found out that the use of community detection leads to a high degree of accuracy in all kinds of machine learning algorithms they chose. The analysis revealed that the approach of exploiting social networking and community features has significantly improved the classification performance by achieving high accuracy with reduced classification time.
Sources:
Jia, P., Liu, J., Fang, Y., Liu, L., & Liu, L. (2018). Modeling and analyzing malware propagation in social networks with heterogeneous infection rates. Physica A: Statistical Mechanics and its Applications, 507, 240-254.
Reddy, V., Kolli, N., & Balakrishnan, N. (2021). Malware detection and classification using community detection and social network analysis. Journal of Computer Virology and Hacking Techniques, 17(4), 333-346.