Malware Materials Detection by Clustering the Sequence using Hidden Markov Model

Abstract

The exponential development in (malware)malicious development in current times and the growing scenario of challenges that malware poses to network atmospheres, like as  the network and smart networks, highlight the need for further investigation in data technology and privacy forensics on computer network protection.The approach described in this scenario curve distinguishes "types" of malware groups that complex is extending, more obfuscated and more varied in nature. We suggest a hybrid strategy integrating signature recognition elements with machine learning-based approaches for classifying families of malware.The approach is performed using PHMM of the behavior features of the species of malware. This research article illustrates the method of modelling and learning developed PHMM using sequences derived from the discovery of the paramount aspects of each malware family, and the recognized orders produced during the development of Multiple Sequences Alignment (MSA). Since all file sets are not dangerous, and the aim would be to separate their malicious parts from the genuine ones along with put greater focus on them to raise the possibility of malware finding by ensuring the least case effect on the genuine parts. Founded on the "consensus series," the investigational findings indicate that even though minimal training data are usable, our proposed method outperforms other HMM-related strategies.