/Intrusion-Detection

Using a dataset provided by Kaggle, it is vital to make sure people are safe in their houses and so I created an intrusion detection model using SVM (4 different types).

Primary LanguageJupyter Notebook

Intrusion Detection

This project focuses on intrusion detection using machine learning techniques on the NSL-KDD dataset. The dataset used for this project is the NSL-KDD dataset, which can be found here.

Project Overview

  • Exploratory Data Analysis (EDA): Perform a comprehensive analysis of the dataset to understand the distribution, relationships, and characteristics of the features. This step helps in gaining insights into the data and identifying potential patterns.

  • Data Preprocessing: Preprocess the data by handling missing values, scaling features, and encoding categorical variables. This step is crucial for preparing the data for model training.

  • Model Building: Develop four Support Vector Machine (SVM) models with different kernel functions for intrusion detection. These models have achieved an accuracy of 93% on the NSL-KDD dataset. The SVM models will leverage the preprocessed data to make accurate predictions.

Repository Structure

  • data/: Contains the dataset files used for the project.

  • notebooks/: Contains Jupyter notebooks documenting the EDA, data preprocessing, and model building steps.

  • models/: Contains the saved trained SVM models.

  • src/: Contains any additional source code or utility functions used in the project.