부산대학교 도서관

As today's time is more dependent towards the internet, there are various types of malwares are being developed on daily basis, as per report presented by Kaspersky around 560,000 instances of malware created daily. Malware are being developed in such a way that they can replicate and change their signature based on the types of detection system present in organization. Due to urge of variety of malwares it is enables security researchers to shift towards use of Machine Learning for better protection towards variety of security systems. Use of machine learning makes the whole thing more productive and responsive. In this research we have suggested a system that detects diversity of viruses depending on their types response to the system, for the initial phase we have taken a Dataset of Malwares which contains 96,724 malware samples and 41,323 binaries, executable and dynamic link libraries which are legit files. Also we have contrasted the different types of machine learning Techniques which can be used for creating these types of system.