Title: Evaluating the behaviour of stream learning algorithms for detecting invasion on wireless networks

Authors: Cláudio Alves; Flávia Bernardini; Edwin B. Mitacc Meza; Leandro Sousa

Addresses: Institute of Science and Technology – ICT, Fluminense Federal University – UFF, Rio das Ostras, RJ, Brazil ' Institute of Computing – IC, Fluminense Federal University – UFF, Niterói, RJ, Brazil ' Institute of Science and Technology – ICT, Fluminense Federal University – UFF, Rio das Ostras, RJ, Brazil ' Institute of Science and Technology – ICT, Fluminense Federal University – UFF, Rio das Ostras, RJ, Brazil

Abstract: Ensuring protection in computer networks is an increasingly difficult task because of the sheer number and variability of threats currently encountered. Intrusion detection systems (IDSs) is usually used to improve the security of information in computers networks, including any content that has value to a person or company. IDS monitor computers or networks to identify malicious activity or unauthorised access. An open issue is how much data is necessary for constructing models for predicting invasion in wireless networks, specially considering that are some scenarios that dataset is not promptly available. Our approach should consider constructing classifiers given a dataset and, as the dataset grows, new classifiers are constructed. Other strategy is explore stream learning algorithms that adapt models along the time. In addition to studying the applicability of stream learning algorithms. This work aims to investigate whether in terms of processing time, stream algorithms are more efficient than batch ones.

Keywords: stream learning; intrusion detection systems; IDSs; wireless networks.

DOI: 10.1504/IJSN.2020.109691

International Journal of Security and Networks, 2020 Vol.15 No.3, pp.133 - 140

Received: 15 Aug 2019
Accepted: 31 Aug 2019

Published online: 11 Sep 2020 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article