Andrei Kelarev

Dr. Andrei Kelarev

School of Computing Technologies, STEM College, RMIT University, Melbourne, VIC 3000, Australia
+1
orcid

Research Focus & Metrics

Academic Career Length

30Years (1997-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
28 April 2022 in Electronics
Verified
Journal Article
Open access
PDF file
RAHMAN ZIAURRAHMAN ZIAUR
RAHMAN ZIAUR
XYXun Yi
Xun Yi
Sk Tanzir MehediSk Tanzir Mehedi
Sk Tanzir Mehedi
AIAProf Rafiqul Islam
AProf Rafiqul Islam
Andrei KelarevAndrei Kelarev
Andrei Kelarev
Blockchain has recently drawn wide attention in the research community. Since its emergence, the world has seen the expansion of this new technology, which was initially developed as a digital currency more than a decade ago. A self-administering ledger that ensures extensive data immutability over a peer-to-peer network has made it attractive for cybersecurity applications, including sensor-enabled Internet of Things (IoT) systems. Brand new challenges and questions now demand solutions, as IoT devices are now online in a distributed fashion to simplify our everyday lives. Motivated by those challenges, the work here has detailed issues from which an IoT infrastructure can suffer if the wrong blockchain technology is chosen. Unlike a typical review, this paper focuses on security challenges of the blockchain-IoT ecosystem through critical findings and applicable use cases. The contribution directs Blockchain architects, designers, and researchers in the domain to select an unblemished combination of Blockchain-powered IoT applications. In addition, the paper provides insight into the state-of-the-art Blockchain platforms, namely Ethereum, Hyperledger, and IOTA, to exhibit their respective challenges, their constraints, and their prospects in terms of performance and scalability.

2 March 2017 in IEEE Transactions on Big Data
Verified
Journal Article
Closed access
Jemal H. AbawajyJemal H. Abawajy
Jemal H. Abawajy
Andrei KelarevAndrei Kelarev
Andrei Kelarev
Malicious software (malware) pose serious challenges for security of big data. The number and complexity of malware targeting Android devices have been exponentially increasing with the ever growing popularity of Android devices. To address this problem, multi-classifier fusion systems have long been used to increase the accuracy of malware detection for personal computers. However, previously developed systems are quite large and they cannot be transferred to Android platform. To this end, we propose Iterative Classifier Fusion System (ICFS), which is a system of minimum size, since it applies a smallest possible number of classifiers. The system applies classifiers iteratively in fusion with new iterative feature selection (IFS) procedure. We carry out extensive empirical study to determine the best options to be employed in ICFS and to compare the effectiveness of ICFS with several other traditional classifiers. The experiments show that the best outcomes for Android malware detection have been obtained by the ICFS procedure using LibSVM with polynomial kernel, combined with Multilayer Perceptron and NBtree classifier and applying IFS feature selection based on Wrapper Subset Evaluator with Particle Swarm Optimization.

1 January 2019 in Future Generation Computer Systems
Verified
Journal Article
Closed access
Andrei KelarevAndrei Kelarev
Andrei Kelarev
XYXun Yi
Xun Yi
SBShahriar Badsha
Shahriar Badsha
XYXuechao Yang
Xuechao Yang
LRLeanne Rylands
Leanne Rylands
JSJennifer Seberry
Jennifer Seberry

1 July 1997 in Proceedings of the American Mathematical Society
Verified
Journal Article
Open access
Andrei KelarevAndrei Kelarev
Andrei Kelarev

25 April 2012 in Optimization
Verified
Journal Article
Open access
GBGleb Beliakov
Gleb Beliakov
Andrei KelarevAndrei Kelarev
Andrei Kelarev
JYJohn Yearwood
John Yearwood
Large outliers break down linear and nonlinear regression models. Robust regression methods allow one to filter out the outliers when building a model. By replacing the traditional least squares criterion with the least trimmed squares (LTS) criterion, in which half of data is treated as potential outliers, one can fit accurate regression models to strongly contaminated data. High-breakdown methods have become very well established in linear regression, but have started being applied for non-linear regression only recently. In this work, we examine the problem of fitting artificial neural networks (ANNs) to contaminated data using LTS criterion. We introduce a penalized LTS criterion which prevents unnecessary removal of valid data. Training of ANNs leads to a challenging non-smooth global optimization problem. We compare the efficiency of several derivative-free optimization methods in solving it, and show that our approach identifies the outliers correctly when ANNs are used for nonlinear regression.