Skip to content

A Lightweight Cascade-Based Farmework for Real-time Ze-ro-Day Attack Detection

Pre-release
Pre-release

Choose a tag to compare

@FRakhmatov FRakhmatov released this 27 Jan 10:51
· 3 commits to main since this release
b6a18b8

This repository contains Python code demonstrating a two-stage XGBoost model for zero-day network intrusion detection. The approach is designed to identify novel attack types (zero-day attacks) that were not present in the training data.

  1. Test results in the CSE-CIC-IDS2018 dataset

  2. Test results in the CSIC-2012 dataset

  3. Test results in the UNSW-NB15 dataset