Scholarlink

With the advent of the Internet of Things (IoT), machines are now able to communicate, collect data, and make decisions, revolutionizing everyday life. However, the increasing number of IoT devices put the latter at risk for the following cyberattacks; Aggressive scan, SSH Brute Force, UDP scan Sparta, MQTT Brute Force, Flood DoS, and SlowITe. This research aims to establish how models such as Naïve Bayes (NB), Decision Tree (DT), Random Forest (RF), Bagging (BG), AdaBoost (AB), Histogram-Based Gradient Boosting (HGB), XGBoost, and Stacking perform in detecting these attacks. Evaluation measures that are used are F1 score, accuracy, precision, and recall. The results show that XGBoost excels in accuracy (96.42%) and Recall (96.42%), BG in precision (98.56%), and AB in F1 Score (96.42%). The goal and problem of this research are aimed at improving the security level of IoT environments, proposing an effective method to identify MQTT-based threats, and establishing a reliable connection to IoT applications.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Lightweight MQTT Protocol Security Analysis Based on ML

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Shrutika Sablok,
  • Surjit Singh

摘要

With the advent of the Internet of Things (IoT), machines are now able to communicate, collect data, and make decisions, revolutionizing everyday life. However, the increasing number of IoT devices put the latter at risk for the following cyberattacks; Aggressive scan, SSH Brute Force, UDP scan Sparta, MQTT Brute Force, Flood DoS, and SlowITe. This research aims to establish how models such as Naïve Bayes (NB), Decision Tree (DT), Random Forest (RF), Bagging (BG), AdaBoost (AB), Histogram-Based Gradient Boosting (HGB), XGBoost, and Stacking perform in detecting these attacks. Evaluation measures that are used are F1 score, accuracy, precision, and recall. The results show that XGBoost excels in accuracy (96.42%) and Recall (96.42%), BG in precision (98.56%), and AB in F1 Score (96.42%). The goal and problem of this research are aimed at improving the security level of IoT environments, proposing an effective method to identify MQTT-based threats, and establishing a reliable connection to IoT applications.