Presenter(s)

Kayla D Chisholm

Download

Download Project (241 KB)

Description

Over the last two years, machine learning has become rapidly utilized in cybersecurity, rising from almost non-existent to currently over half of cybersecurity techniques utilized commercially. Machine learning is advancing at a rapid rate, and the application of new learning techniques to cybersecurity have yet to be investigated. The presented work discusses the use of novel machine learning algorithms for network intrusion detection. Typically, this type of network intrusion detection system operates based on a set of rules programmed to recognize known attacks and intrusion techniques. However, this detection method does little to prevent new, or ‘zero day’ attacks. On the other hand, a detection system that uses machine learning could analyze patterns in network data in real time to determine attack likelihood. To test the efficacy of machine learning and neural network algorithms for network security, we use publicly available data sets to ‘teach’ these neural systems what an attack may look like. After training, the system will be tested to determine how well it learned the features contained within the input data. Our results show accuracy and error rates of the algorithms that have been implemented.

Publication Date

4-24-2019

Project Designation

Honors Thesis

Primary Advisor

Chris G. Yakopcic

Primary Advisor's Department

Electrical and Computer Engineering

Keywords

Stander Symposium project

Machine Learning for Cyberattack Detection

Share

COinS