Predicting Future Malware Attacks on Cloud Systems using Machine Learning

Machine learning is one of the fastest-growing fields nowadays and its application to cybersecurity is gaining much attention. With the development and increased adoption of cloud computing, numerous malware threaten both service providers and consumers. Many machine learning algorithms were used to predict the future behavior of cloud systems to protect them from malicious insiders and external attacks. However, conventional machine learning algorithms have the limitation that they show weak performance when the dataset is large and sparse. In this paper, we explore a gradient boosting decision tree, especially LightGBM, which is a relatively new and powerful method, to predict future malware attacks on cloud computing systems. We use a large and sparse dataset provided by Microsoft and show that our approach is suitable for predicting malware attacks using large datasets with 73.89% accuracy compared to conventional machine learning methods.