Network Anomaly Detection

This repository contains code and resources for detecting network anomalies using machine learning techniques.

Table of Contents

• Introduction
• Installation
• Usage
• Datasets

Introduction

Network anomaly detection is crucial for identifying unusual patterns that may indicate security threats or performance issues. This project aims to provide a robust solution for detecting such anomalies in network traffic.

Installation

To get started, clone the repository and install the required dependencies:

git clone git@github.com:phierhager/network-anomaly.git
cd network-anomaly
poetry install
poetry shell

Datasets

To be able to train, download the UNSW-NB15 dataset from here (https://www.kaggle.com/datasets/dhoogla/unswnb15) or the CIC-IDS2017 dataset from here (https://www.kaggle.com/datasets/dhoogla/cicids2017) and save them in a data/ folder on top level.

Usage

To train a detection model refer to the train_tutorial.ipynb notebook. The real-time package inspection is under development.