Enhancing network intrusion detection with an ensemble of deep learning and machine learning algorithms

Author

Chisero, Tawanda J.

Title

Enhancing network intrusion detection with an ensemble of deep learning and machine learning algorithms

Abstract

This project seeks to enhance network intrusion detection through an Ensemble model that combines Machine and Deep Learning algorithms. The model integrates the Multilayer Perceptron (MLP), Random Forest Classifier and the Support Vector Machines (SVM). This approach is meant to address the limitations of standalone classifiers by combining their complementary strengths to improve the system’s ability in identifying complex intrusion patterns and adapt to evolving cyber threats. The author aimed at developing a reliable network intrusion detection tool that network security professionals can trust in. To evaluate its practical applicability and performance, the trained ensemble model is deployed within a simulated network environment through a simple web application using Streamlit. This interface allows for real-time intrusion prediction and evaluation of the system’s ability to classify network traffic accurately. Ultimately, the project contributes to the broad field of safeguarding networks in a digital landscape where cyberthreats are evolving and becoming more complex to detect with traditional methods.  

Date

June 2025

Publisher

BUSE

Keywords

Network Intrusion Detection

Ensemble

Deep Learning

Machine

Supervisor

Mr. J. Musariwa