Malware Detection and Classification with Machine Learning Algorithms

The Windows platform can be easily affected by abnormal growth of malicious software, so we need to find abnormal activities in the system and vary them effectively. Manually finding an abnormal growth of activities in the system is not possible at all times and takes time. We need to assign humans to monitor activities. To solve this problem, we have to develop an automated system for finding abnormal activities at all times without human interaction. We applied several kinds of techniques involving machine learning to quickly determine whether the algorithm models were effective and detect any abnormal system growth. Machine learning methods such as XGBoost, Naive Bayes, K-NN, Random Forest, decision tree, logistic regression, and Support Vector Machine have been utilized to detect abnormal growth in the system. We found that the XGBoost classification algorithm performs the most accurate classification among the algorithms tested. At the end of the testing, a 99.28% accurate prediction was obtained.