NSL-KDD Dataset Analysis and Model Deployment

This repository contains the code and resources for analyzing the NSL-KDD dataset, a network traffic dataset designed for cybersecurity analysis and intrusion detection. The dataset has been preprocessed and analyzed using various machine learning techniques, and a model has been deployed using Flask and Docker.

Data Preprocessing Steps

• Outlier Removal: Outliers were removed from the data using the median replacement method.
• One-Hot Encoding: Categorical variables were encoded using one-hot encoding to prepare them for modeling.
• Feature Scaling: RobustScaler was applied to scale the features to mitigate the impact of outliers.
• Dimensionality Reduction: Principal Component Analysis (PCA) was explored to reduce the dimensionality of the dataset.

Machine Learning Models

Several machine learning models were trained and evaluated using the preprocessed data, including:

• Random Forest Classifier
• Decision Tree Classifier

Model evaluation metrics were used to compare the performance of these models, and hyperparameter tuning was performed to optimize their performance.

Model Deployment

The trained model was deployed using Flask, a lightweight web framework, and Docker, a containerization platform. This allows for easy deployment and scalability of the model in production environments.

Repository Structure

model_training/: Jupyter notebooks used for data preprocessing, model training, and evaluation. app.py: Flask application for serving the deployed model. Dockerfile: Docker configuration file for containerizing the Flask application. requirements.txt: Python dependencies required for running the Flask application.

Usage

1. Clone the repository:

git clone https://github.com/elifsare/Anomaly-Detection.git
cd nsl-kdd-analysis

2. Install dependencies:

pip install -r requirements.txt

3. Run the Flask application:

python app.py

4. Access the deployed model via http://localhost:5000 in your web browser.