Date of Award

1-1-2022

Publication Type

Thesis

Degree Name

M.A.Sc.

Department

Electrical and Computer Engineering

Keywords

Water storage, Dimensionality reduction, Intrusion detection

Supervisor

J. Ahamed

Supervisor

B. Rashidzadeh

Rights

info:eu-repo/semantics/openAccess

Abstract

Supervisory control and data acquisition (SCADA) systems are often imperiled bycyber-attacks, which can often be detected using intrusion detection system (IDSs).However, the performance and efficiency of IDSs can be affected by several factors,including the quality of data, curse of dimensionality of the data, and computationalcost. Feature reduction techniques can overcome most of these challenges by eliminatingthe redundant and non-informative features, thereby increasing the detectionaccuracy. This study aims to shows the importance of feature reduction on the intrusiondetection performance. To do this, a multi-modular IDS is designed that isconnected to the SCADA system of a water storage tank. A comparative study isalso performed by employing advanced feature selection and dimensionality reductiontechniques. The utilized feature reduction techniques improves the IDS efficiency byreducing the memory usage and using data with better quality, which in turn increasethe detection accuracy. The obtained results have been analyzed in terms of F1-scoreand accuracy.

Share

COinS