Skip to content

loki52501/autoencode_detection_research-Project

Repository files navigation

UNSW Categorical Analysis – Final Project

This repository contains a Jupyter Notebook for analyzing and testing categorical features of the UNSW-NB15 dataset, a benchmark dataset widely used in network security and intrusion detection research.


📌 Project Overview

  • Focused on categorical preprocessing and evaluation for the UNSW dataset.
  • Explores encoding techniques, feature separation, and classification performance.
  • Provides a reproducible framework for testing categorical handling methods on security-related data.

📂 Repository Structure


unsw-categorical-final-separate-test.ipynb   # Main notebook with preprocessing and experiments
data/                                        # (Optional) Folder for dataset storage
results/                                     # (Optional) Folder for experiment outputs
README.md                                    # Project documentation


⚙️ Requirements

Make sure you have the following installed before running the notebook:

  • Python 3.8+
  • Jupyter Notebook or JupyterLab
  • Required libraries:
    • pandas
    • numpy
    • scikit-learn
    • matplotlib
    • seaborn

Install dependencies with:

pip install -r requirements.txt

(If no requirements.txt is provided, install the libraries manually.)


🚀 How to Run

  1. Clone the repository:

    git clone https://github.com/YOUR_USERNAME/unsw-categorical-analysis.git
    cd unsw-categorical-analysis
  2. Launch Jupyter Notebook:

    jupyter notebook
  3. Open unsw-categorical-final-separate-test.ipynb and run the cells in order.


📊 Key Features

  • Data Cleaning – handling missing values, categorical separation.
  • Encoding Methods – one-hot encoding, label encoding, frequency encoding.
  • Model Training & Testing – supervised learning models on processed categorical features.
  • Evaluation – accuracy, precision, recall, and F1-score metrics.

📈 Results

  • Insights into the effect of different categorical encoding techniques.
  • Model performance benchmarks on the UNSW-NB15 dataset.
  • Framework extendable to other cybersecurity datasets.

🧠 Inspiration

This project was inspired by the need to better understand how categorical data preprocessing impacts intrusion detection and cybersecurity analytics.


🤝 Contributing

Contributions are welcome! If you’d like to extend the project (e.g., new encoding techniques, model architectures, or visualizations), please fork the repo and submit a PR.


📜 License

This project is licensed under the MIT License. Feel free to use, modify, and share with attribution.

About

No description, website, or topics provided.

Resources

Stars

1 star

Watchers

0 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors