A Real-Time Malware Detection System Combined With Machine Learning and YARA Rules

As malicious software continues to proliferate at an unprecedented rate and digital devices become increasingly integral to everyday activities, malware detection has becoming more and more important. The traiditional signature-based detection method has already been widely used. However, these rules are becoming increasingly difficult to handle the threat of variant viruses. Researchers use multiple dynamic analysis methods to improve malware detection. However, some of these methods may be costly in terms of time and hardware resources. To address the growing demands of malware detection, developing an analysis system that combines real-time performance and accuracy is crucial. The proposed method presents a fast and efficient malware detection system combing the advantages of machine learning and YARA rules, achieving 97% accuracy and low false positive rate.