In this Jupyter Notebook project, modern machine learning libraries are applied onto an older dataset - the KDD Cup 1999 dataset. The dataset is a simulation of a military computer network; the records are comprised of internet connections that are classified as either normal connections or detected intrusion (with a specified attack type). A supervised machine learning model is built to perform binary classifications on whether a record is normal or an intrusion. The motivations of this project were to investigate the effectiveness of machine learning algorithms on a dataset that historically was much harder to build a model for.
While GitHub supports rendering Jupyter Notebook projects, it does not support rendering them as default README page. Furthermore, there are performance issues when rendering non-trivial notebooks. For best results, the notebook can be viewed using Jupyter's nbviewer.
The dataset for this project is the KDD Cup 1999 Dataset.