Cyber Times International Journal of Technology & Management

To achieve accuracy in finding these malicious softwares different methods has been used such as through Machine Learning (ML) which works on supervised learning, Deep Learning (DL) which works on unsupervised learning, and Artificial Intelligence (AI)which works reinforcement learning on the basis of these techniques framework of malware detection on Operating system, malicious softwares includes viruses, worms, trojans, ransomwares, spyware, rootkits and many more softwares which interrupt’s other softwares, steal information, creates blockage on processor or try to harm any user are considered as a malware. These variants of malwares have different types of objectives, different methods of breaching and for further growth. To resolve a variants of problems variants of solutions are required. On the way to solutions initially it is retaliated through ML models which works on static analysis with static feature selection which is constant and further it evolved to DL which works on dynamic analysis with no constant features defined at early stage and AI works on enhanced signature based techniques, ensemble learning, feature engineering and selection and many more. At any time whenever analysis is required there is a requirement of benchmark on the basis of which a difference can be notified. For this malware’s history is considered to prepare a dataset on malware’s features, availability, methods of intrusions and many more to train different models for prevention from these malwares in future.