Project Title: The Use of Machine Learning in The Detection of Anomaly Network Intrusion
This Project is part of the Course : CS 584 Machine Learning
In this Repository, The Project Source Code includes the notebook files for different datasets :
- Project Code for Training Dataset.
- Project Code for Testing Dataset.
The Dataset folder includes the 'UNSW-NB15 dataset' containing the different parts of the dataset:
- NUSW-NB15_features - Describes the Features in the Dataset
- UNSW-NB15_testing-set - 82,332 records in the testing set
- UNSW-NB15_training-set (in .zip as filesize>25MB) - 175,341 records in the training set
Both of the Python notebooks of the project source code were run primarily on the Local System's Juypter Notebook, as well as in the Google Colab.
The Project contains two different jupyter notebooks implemented for each dataset i.e., Training and Testing Dataset, Which Can Be Implemented in a Single file, But due of GPU and CPU constraints, and also for making it more easier to implement. It has been split into two different notebooks.
-
Akshat Behera : Responsible for Unsupervised Learning Models (Neural Network (using MLP Classifier), CNN, RCNN) and a Supervised Model - (Random Forest) Implementation along with Feature Selection and EDA Implementations.
-
Nagarjuna Bolla : Responsible for Supervised Learning Models (Logistic Regression, Decision Tree, SVM, XGBoost) Implementation and Data-Preprocessing(Data Cleaning, Encoding and Standardization).