Strengthening intrusion detection system for adversarial attacks improved handling of imbalance classification problem

Name: Chutipon Pimsarn

keyword: Intrusion detection system

MeSH: Cluster analysis -- 12879

Classification :.DDC: e-Theses

; Adversarial attack, Machine learning, Imbalance classification, Data clustering

MeSH: Database security -- 78717

MeSH: Data protection -- 289872

MeSH: Computers -- Access control -- 289873

LCSH: Cluster analysis -- 12879

LCSH: Database security -- 78717

LCSH: Data protection -- 289872

LCSH: Computers -- Access control -- 289873

Abstract: A network-based intrusion system or NIDS is the best defense mechanism, often sup-optimal for detecting an unseen malicious pattern. In response, many studies have attempted to promote NIDS that uses machine learning to improve their ability to recognize an opponent's attack. According to this research line, the current work focuses on non-payload connections at the TCP stack level, which is generalized and applicable to different network applications. As a complement to the recently published investigation that searches for the most informative feature space for classifying obfuscated connections, the problem of class imbalance is examined herein. Especially, a multiple-clustering-based undersampling framework is proposed to determine the set of cluster centroids that best represent the majority class to reduce par size to be equal to the minority. Initially, a pool of centroids is created using ensemble clustering to obtain a group of precise and diverse groupings. Then select the last representative from this group. Three different objective functions are formed for this optimization-driven process, thus leading to three variants of FF-Majority, FF-Minority, and FF-Overall. Based on a detailed assessment of the published dataset, four classification models, and different settings, these new methods show better predictive performance than baseline data: the single-clustering undersampling counterpart and state-of-the-art techniques. Parameter analysis and implications for analyzing an extreme case are also guidelines for future applications.

Mae Fah Luang University. Learning Resources and Educational Media Center

Address: CHIANG RAI

Email: library@mfu.ac.th

Name: Tossapon Boongoen

Created: 2021

Modified: 2022-08-08

Issued: 2022-08-08

วิทยานิพนธ์/Thesis

application/pdf

CallNumber: e-Theses

eng

DegreeName: Master of Science

Level: Master's Degree

Descipline: Information Technology

Grantor: Mae Fah Luang University

©copyrights Mae Fah Luang University

ลำดับที่.	ชื่อแฟ้มข้อมูล	ขนาดแฟ้มข้อมูล	จำนวนเข้าถึง	วัน-เวลาเข้าถึงล่าสุด
1	08 b.135049 6151301002 (6-8-65).pdf	2.19 MB	2	2022-10-18 21:21:40

ใช้เวลา

0.016567 วินาที