mahendradata / ab-ht

The implementation of AB-HT: An Ensemble Incremental Learning Algorithm for Network Intrusion Detection Systems

AB-HT

This repository is a proof of concept of algorithms described in AB-HT: An Ensemble Incremental Learning Algorithm for Network Intrusion Detection Systems paper.

Installation

Clone repo and install requirements.txt in a $ python>=3.8.0 environment.

$ git clone https://github.com/mahendradata/ab-ht.git # clone
$ cd ab-ht
$ pip install -r requirements.txt  # install

Prepare the dataset

Before running the main program, we need to prepare the CICIDS2017 dataset. Run the following commands to prepare the dataset.

$ cd datasets/
$ bash step1-download.sh
$ python step2-preprocessing.py
$ python step3-merging-monday.py

Running other programs

After the dataset is ready, you can run other programs in this repository. But first, let create a folder for the output file.

$ cd ..  #back to the root folder of the repository
$ mkdir outputs

Then run the experimental programs, for example:

Running the AB-HT program:

$ python AB-HT.py conf/AB-HT.json outputs/AB-HT.log

Running the HT program:

$ python HT.py conf/HT.json outputs/HT.log

Running the HATT program:

$ python HATT.py conf/HATT.json outputs/HATT.log

Running the DT program:

$ python DT.py conf/DT.json outputs/DT.log

Running the AB-DT program:

$ python AB-DT.py conf/AB-DT.json outputs/AB-DT.log

Folder Structure

• datasets: contains the preprocessed dataset.
• conf: contains configuration files in json format.
• lib:
  • dataset: contains libraries to preprocess and manipulate the dataset.
  • model: contains libraries to create the models.
  • util: contains utility libraries.

These are additional folders. You need to create this folder on your own.

• outputs: a folder to save the experimental data.