The Internet of Things (IoT) is one of the technical advancements that is progressing swiftly. which promises to be revolutionary soon. IoT systems are convenient due to its device centralized and computerized control. This technology allows various physical devices, home applications, vehicles, appliances, etc., to be interconnected and exposed to the Internet. On the other hand, it entails the fundamental need to protect the network from adversarial and unwanted alterations. Machine-to-machine protocols like Message Queuing Telemetry Transport (MQTT) are typically used by IoT devices to communicate. Numerous techniques to attack networks employ the lightweight messaging protocol known as MQTT (Message Queuing Telemetry Transport). Due to its heterogeneous nature and the lack of security approaches, the publish-subscribe strategy utilized by the MQTT protocol increases the number of potential network attacks. This thesis presents a novel approach to detecting cybersecurity breaches in MQTT-IoT networks using machine learning techniques. We suggest a detection system to address the issue of cybersecurity threats in MQTT-IoT networks. Our method involves cleaning the data to pull out relevant features, training the ensemble machine learning models on these features, and then using these models to find anomaly behavior that could indicate a cyberattack. We implemented our plan by using Machine Learning Ensemble techniques and Feature selection. To test our system, we ran many experiments using MQTT-IoT-IDS2020, a dataset that included both normal MQTT-IoT network activity and simulated attacks of different types. Our experimental findings indicate that our detection system, grounded in machine learning, can identify cybersecurity threats on MQTT-IoT networks with notable accuracy, precision, F1-score, and recall. The obtained results for binary and multiclass classification indicate that the proposed system can bring a remarkable layer of security. We show how Machine Learning Ensemble Techniques applied to small low-cost devices are an efficient and versatile combination characterized by a bright future ahead. This thesis advances the application of machine learning methodologies in cybersecurity and contributes to the enhancement of security protocols within MQTT-IoT networks.

Library of Congress Subject Headings

Cyberterrorism--Prevention; Internet of things--Security measures; Computer networks--Security measures; Machine learning

Publication Date


Document Type


Student Type


Degree Name

Computing Security (MS)


Wesam Almobaideen

Advisor/Committee Member

Huda Saadeh

Advisor/Committee Member

Kevser Akpinar


RIT Dubai

Plan Codes