Malware detection based on TF-(IDF&ICF) method

Abstract As the level of information technology continues to increase, security problems caused by malware are becoming more and serious. An important method detect is analyze software behavior information, such as permissions, API call sequences system calls. In this paper, used research object for detection, traditional feature extraction methods not ideal sequences. a TF-(IDF&ICF) proposed mathematical analysis, which combines document category features. Experiments show that using extractor in followed training, performance improved four different machine learning models, F1 can reach 0.979, while response time significantly reduced, has good practical value.