Scholarlink

<p>The expansion of Internet of Things (IoT) devices has generated considerable apprehension regarding their security. Insufficient security protocols enable the exploitation of IoT devices by adversaries. The distributed denial-of-service attack is one instance of such assaults. Therefore, the presence of intrusion detection systems in the Internet of Things is of considerable importance. This work use the group approach of majority voting, a particular class of machine learning algorithms, to detect and predict attacks. The justification for utilizing this method is to achieve improved detection accuracy and a minimal rate of false positives by integrating multiple machine learning classification algorithms across various Internet of Things networks. The proposed method was assessed utilizing the augmented and updated CICDDOS2019 dataset. The simulation results indicate that the majority voting group method, when applied to five attacks from this dataset, attains detection rates of 99.97%, 99.98%, 100%, and 99.97%, respectively. A success rate of 99.67% was achieved in detecting DNS, NETBIOS, LDAP, UDP, and SNMP attacks, illustrating exceptional and constant efficacy in identifying and predicting assaults relative to conventional models.</p>

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

Detection of DDoS attacks in IoT networks using a majority voting mechanism

  • 期刊论文

  • 发表日期：

    2025年10月17日

查看全文
  • Suhua Chen,
  • Xu Fang

摘要

The expansion of Internet of Things (IoT) devices has generated considerable apprehension regarding their security. Insufficient security protocols enable the exploitation of IoT devices by adversaries. The distributed denial-of-service attack is one instance of such assaults. Therefore, the presence of intrusion detection systems in the Internet of Things is of considerable importance. This work use the group approach of majority voting, a particular class of machine learning algorithms, to detect and predict attacks. The justification for utilizing this method is to achieve improved detection accuracy and a minimal rate of false positives by integrating multiple machine learning classification algorithms across various Internet of Things networks. The proposed method was assessed utilizing the augmented and updated CICDDOS2019 dataset. The simulation results indicate that the majority voting group method, when applied to five attacks from this dataset, attains detection rates of 99.97%, 99.98%, 100%, and 99.97%, respectively. A success rate of 99.67% was achieved in detecting DNS, NETBIOS, LDAP, UDP, and SNMP attacks, illustrating exceptional and constant efficacy in identifying and predicting assaults relative to conventional models.