TRAP@NCI

Network based Anomaly Detection: An ensemble approach

Bency, Johan (2017) Network based Anomaly Detection: An ensemble approach. Masters thesis, Dublin, National College of Ireland.

[img]
Preview
PDF (Master of Science)
Download (927kB) | Preview

Abstract

Intrusion detection is a relatively mature domain. Network Intrusion Detection Systems have been used for over a long period of time for detecting potential intruders or attackers in a network or similar environment. Detecting and eliminating such threats is essential for the growth of any company in this cashless economy era where more and more users are moving to online banking, ecommerce and other related domains. With new technologies such as machine learning, neural networks, the advancements in computers such as processors and memory, and programming languages and libraries custom tailored for machine learning such as R, Python, matplotlib, scikit-learn etc. NIDS instead of being a vague Intrusion detection approach is now an accurate Intrusion detection system. The anomalies in network data are effectively predicted using such machine learning and neural network algorithms. In this project we are trying to create a model which works with a classic NIDS benchmark dataset in its backbone to create an accurate system which predicts network anomalies. Here we are using ensembled approaches for both sampling and for prediction phases for getting a superior result. For the sampling process we are using SMOTETomek algorithm which combine regular SMOTE for over-sampling and Tomek link for under-sampling. For the classification approach we are using Support Vector Machines (SVM), K-Nearest Neighbors, Nave Bayes. The method is not claiming an absolute accuracy or performance, and it can be still be improved upon with better resources and tools.

Item Type: Thesis (Masters)
Subjects: Q Science > QA Mathematics > Electronic computers. Computer science
T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science
Divisions: School of Computing > Master of Science in Data Analytics
Depositing User: CAOIMHE NI MHAICIN
Date Deposited: 28 Aug 2018 14:46
Last Modified: 28 Aug 2018 14:46
URI: http://trap.ncirl.ie/id/eprint/3097

Actions (login required)

View Item View Item