Skip to content

osama336/Network-Anomaly-Detection-System

Repository files navigation

Network-Anomaly-Detection-System using Streamlit

The Network Anomaly Detection System (NADS) has been developed to enhance network monitoring and security by transitioning from traditional manual detection methods to an automated, AI-driven approach. This system leverages modern technologies to ensure efficiency, accuracy, and scalability in detecting and addressing network anomalies.

NADS aims to save time, effort, and resources for network administrators and users by offering a user-friendly interface and robust functionality. Feedback on existing systems and their limitations was gathered through surveys to design a solution tailored to user needs. The system has been developed in two main parts:

A Streamlit-based web interface for real-time interaction, allowing administrators to monitor network traffic, view anomaly alerts, and generate reports.

A backend powered by TShark for network traffic analysis, coupled with a Keras-based deep learning model to detect anomalies with high accuracy.

Data preprocessing and analysis are managed using Pandas, while Plotly is employed to visualize packet capture and anomaly patterns interactively. The integration of these tools enables the system to process large volumes of data efficiently and provide actionable insights to users.

In the future, NADS aims to expand its capabilities by incorporating advanced machine learning models and additional features to support broader network environments, ensuring comprehensive protection against evolving security threats.

How to Run

  1. Install dependencies: pip install -r requirements.txt

  2. ReadMe-If-Python-Venv.txt or ReadMe-If-anaconda.txt

About

The system relies on: Collecting and analyzing data using the LUFlow dataset (benign, malicious, outlier). Training an ANN model using Keras to achieve high accuracy and reduce false positives. Developing an interactive interface with Streamlit to monitor live traffic via Tshark.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages