PC恶意软件检测python源代码

Malware is one of the most serious security threats on the Internet today. Unfortunately, the number of new malware samples has explosively increased: anti-malware vendors are now confronted with millions of potential malware samples per year. Consequently, many studies have been reported on using data mining and machine learning techniques to develop intelligent malware detection systems. Lots of works use different feature and different data set to train a classification model. Although they show a high percent of accuracy on their own test data, most of model become rapidly antiquated as malware continues to evolve. When using the obfuscation techniques or polymorphism techniques, they can not work very well. In this work, we propose a effective malware detection approach using data-mining techniques based on opcode, data structure and the imported libraries. We also use different classifiers and conduct some experiments to evaluate our approach. In addition, we provide empirical validation that our method is capable of detecting new unknown malware, also fresh malware collected in 2017. In addition, we use obfuscation on malware to test our model. their own test data, most of model become rapidly antiquated as malware continues to evolve. When using the obfuscation techniques or polymorphism techniques, they can not work very well. In this work, we propose a effective malware detection approach using data-mining techniques based on opcode, data structure and the imported libraries. We also use different classifiers and conduct some experiments to evaluate our approach. In addition, we provide empirical validation that our method is capable of detecting new unknown malware, also fresh malware collected in 2017. In addition, we use obfuscation on malware to test our model.