Intrusion Detection Systems

CSCI735 - Foundations of Intelligent Security Systems.
Rochester Institute of Technology, Fall 2023.

Contributors:
1. Archit Joshi
2.Parijat Kawale

Instructor :
Dr.Leon Reznik,
Professor, Golisano College of Computing and Information Sciences

Project Overview

This course project deals with initial exploration and modelling of Intrusion Detection Systems(IDS). As a part of the project we aim to utilise machine learning and AI techniques to model a misuse based and anomaly based Intrusion detection system (IDS). The project is broken into 3 phases with corresponding documentation for the same.

---

Dataset - https://www.kaggle.com/datasets/galaxyh/kdd-cup-1999-data
Location in project - datasetKDD

NOTE -

• All code in the project can be executed directly from the main() method. To run specific IDS functions one has to comment out the function calls in the script as the script is meant to be ran as a whole for metrics and continuity.
• The code contains multiple dependencies on external libraries, a few of which are keras, tensorflow, pandas, matplotlib. Please review the code and make sure the dependencies are satisfied before execution.

---

Project Phases

• Phase 1
  This phase focuses on identifying the data to work on, data preperation and cleaning as well as experimenting with current state of the art IDS like Snort and Suricata. From our initial exploration with the data set we had, we resorted to using classifier techniques to develop generic IDS in python.

  Phase 1 Code - idsClassifier.py
  Phase 1 Report - Project_Phase_1.pdf




• Phase 2
  This phase focuses on realising the shortcomings in Phase 1 and using Artificial Neural Networks (ANN) to recreate the IDS models for misuse and anomaly based IDS.

  Phase 2 Code - annClassifier.py
  Phase 2 Report - Project_Phase_2.pdf




• Phase 3
  This phase deals with comparing the approaches in Phase 1 and 2, drawing up statistics and addressing shortcomings and room for improvement.

  Phase 3 Report - Project_Phase_3.pdf

---