SVM-RF: A Hybrid Machine Learning Model for Detection of Malicious Network Traffic and Files

Since the inception of information technology, malwares have ravished cyberspace. Malware is a curse that is born from the boon of information technology. The term ‘malware’ is derived from the term ‘malicious softwares’. This term is also used to refer to any software that brings harm to computer resources. As technology advances malwares are becoming more and more of a nuisance. Since 2020, the world has not only been troubled by the COVID-19 pandemic but also from the cybercrime that was enabled by the pandemic. Malwares delivered via COVID-19 themed tactics came only second to scams that targeted the vulnerable society. Threat actors were able to infect individuals and organizations via zero-day vulnerabilities despite the presence of sophisticated defenses. An effective malware combating strategy has become the need of the hour. Effectively and efficiently identifying harmful files and network communications can become an impossible task if done manually. This same task may be greatly aided by application of artificial intelligence. In this research we propose a hybrid machine learning approach that can be used to identify harmful network traffic and malicious files. Several widely used datasets are used to train, validate, and test our model. When evaluated against a variety of classic machine learning methods, our suggested hybrid model outperforms the competition.