| As cyberthreats and attacks are increasingly pervasive,constantly,and rapidly evolving,the battle between pro-malware and anti-malware developers seems to be nowhere close to the end.Also,Covid-19 pandemic and social distancing drives the need for computer and electronic systems for remote and virtual communication which has exponentially influence and boosted the demand curve for computing resources and computer devices upwards,this created an opportunity for cybercriminals leading to a significant increase in malware attacks.Modern-day malware takes various sophisticated and unsuspecting novel means to gain access and attack computer systems without detection.Therefore,the classical ways of dealing with malware detection such as;signature-based,AV scanning,and other ways of preventing malware attacks have proven to be obsolete and no longer efficient.Due to this challenge,dealing with malware is one of the major challenges to the world of computer science and engineering,electronic communication,and technology development.Stakeholders such as law enforcement agents,business entities,and general users have suffered several severe losses.The significance of data science in cybersecurity can never be overrated when it comes to predicting possible future threats for the mitigation and prevention of malware attacks.Several applications of data sciences,machine learning,and Artificial Intelligence in the prevention and control of malware have been proposed in the academic world and implemented in the real world.Our work in this paper is focused on an exploratory implementation of Machine Learning for malware detection leveraging PE file DLL information,N-Grams,and Deep Learning Techniques.We used public malware data sets for our experiments and also detailed the overview and workflow of ML techniques and algorithms for malware detection.The short fall in the usage of data science in cybersecurity has also been explored. |
PDF Full Text Request