Zero-day attacks detection using machine learning techniques

Currently, we live in a world where technology is ubiquitous, and it has become an integral part of our daily lives. Technology continually advances to become more robust and to provide greater privileges to those who consume its services. However, as technology continues to grow, the Internet expands, and so do the vulnerabilities and cyberattacks to which we, as users, may be exposed within a network. It's a network where our data and personal information can be discovered by malicious users. Among these attacks, there are zero-day attacks, which fall within the realm of cybersecurity. They are the most dangerous threats one can encounter within a network or software today. This is because the security measures in place to prevent such cyberattacks lack knowledge or records of them. Zero-day attacks rely on the injection of malicious code, exploiting vulnerabilities that have not yet been discovered by users or the creators of the said software or network. The objective of this graduation project is to propose a new algorithm that, by employing various machine learning techniques, can detect zero-day attacks through anomaly recognition within a dataset containing benign and malicious network traffic.

• Autor Principal: Armijos Inga, Arianna Belen (author)
• Formato: bachelorThesis
• Idioma: eng
• Publicado: 2023
• Subjects: Ciberseguridad; Aprendizaje automático; Cibersecurity; Isolation forest; Machine learning
• Acceso en liña: http://repositorio.yachaytech.edu.ec/handle/123456789/684