Towards a Better Understanding of Machine Learning based Network Intrusion Detection Systems in Industrial Networks

Installation

Ubuntu 18.04.

• sudo apt install python-rpy2
• removed all dependencies from the file which are listed under "pip"
• pip install matplotlib==3.2.2
• pip install pdpbox==0.2.0
• pip install xlrd==1.2.0
• in pdp_plot_utils.py (something line anaconda3/envs/lukas-thesis/lib/python3.7/site-packages/pdpbox/pdp_plot_utils.py), replace the following:

    # inter_ax.clabel(c2, contour_label_fontsize=fontsize, inline=1) # old
    inter_ax.clabel(c2, fontsize=fontsize, inline=1) # new

Windows 10

• The code was tested on Windows 10 Version 1909 with the VS Code IDE
• The dependencies which need to be installed are listed in 'dependencies_windows.yml' If using Anaconda, the following command creates a new environment with the needed dependencies: conda env create -f dependencies_windows.yml

Run

• run conda env create -f dependencies.yml
• to start conda env: conda activate lukas-thesis
• to run the starter code: python src/_playground.py
• to stop the conda env: conda deactivate