Institutional Repository of Vilnius University
Analysis and automated detection of cybersecurity vulnerabilities /
Abstract
dc:descriptionPhishing attacks remain one of the most common cybersecurity threats. Unlike traditional malware that can be detected through signature based methods, phishing exploits human psychology and trust relationships, making it effective at bypassing technical defenses. Modern phishing campaigns employ sophisticated techniques including typosquatting, homoglyph substitution, and dynamic content generation to evade detection by browser security features and email filters. This bachelor work examines phishing attack lifecycle using established cybersecurity frameworks including the Lockheed Martin Cyber Kill Chain and MITRE ATT\&CK . The practical part of the work presents a machine learning based phishing detection system that analyzes URL characteristics and visual brand impersonation. The system combines four supervised algorithms (Random Forest, Gradient Boosting, Logistic Regression, Multi-Layer Perceptron) with one unsupervised model (Isolation Forest). The models are trained on URLs from legitimate sources and phishing databases, with 43 features extracted per URL including domain characteristics, lexical patterns, and brand impersonation indicators. The system is implemented as a Flask REST API with both a command line interface and Chrome browser extension for practical deployment. Detection performance is also compared against Chrome's native Safe Browsing protection. The results of this investigation aims to provide insights into the strengths and current gaps of native endpoint defenses, inform practical recommendations for improving user protection, and suggest further directions for research in the field on the phishing detection.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Urbanovič, Edgar,
Subjects
dc:subject × 1Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- eng
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD277691480&prefLang=en_US