Optimum parameter machine learning classification and prediction of Internet of Things (IoT) malwares using static malware analysis techniques

Shaukat, SU 2019, Optimum parameter machine learning classification and prediction of Internet of Things (IoT) malwares using static malware analysis techniques , MRes thesis, University of Salford.

[img]
Preview
PDF - Accepted Version
Download (3MB) | Preview

Abstract

Application of machine learning in the field of malware analysis is not a new concept, there have been lots of researches done on the classification of malware in android and windows environments. However, when it comes to malware analysis in the internet of things (IoT), it still requires work to be done. IoT was not designed to keeping security/privacy under consideration. Therefore, this area is full of research challenges. This study seeks to evaluate important machine learning classifiers like Support Vector Machines, Neural Network, Random Forest, Decision Trees, Naive Bayes, Bayesian Network, etc. and proposes a framework to utilize static feature extraction and selection processes highlight issues like over-fitting and generalization of classifiers to get an optimized algorithm with better performance. For background study, we used systematic literature review to find out research gaps in IoT, presented malware as a big challenge for IoT and the reasons for applying malware analysis targeting IoT devices and finally perform classification on malware dataset. The classification process used was applied on three different datasets containing file header, program header and section headers as features. Preliminary results show the accuracy of over 90% on file header, program header, and section headers. The scope of this document just discusses these results as initial results and still require some issues to be addressed which may effect on the performance measures.

Item Type: Thesis (MRes)
Contributors: Saraee, MH (Supervisor)
Schools: Schools > School of Computing, Science and Engineering
Depositing User: Syed Shaukat
Date Deposited: 11 Feb 2019 12:45
Last Modified: 11 Mar 2019 01:38
URI: http://usir.salford.ac.uk/id/eprint/49550

Actions (login required)

Edit record (repository staff only) Edit record (repository staff only)

Downloads

Downloads per month over past year