Scholarlink

The Internet of Things (IoT) has revolutionized different businesses, advertising unparalleled comfort and proficiency. In any case, this broad appropriation has too uncovered IoT frameworks to various security dangers, counting malevolent interruptions, and information breaches. Conventional Intrusion Detection Systems (IDSs) regularly battle to keep up with the energetic and complex nature of these assaults, requiring inventive approaches to upgrade IoT security. This paper proposes a novel IDS system utilizing Generative Adversarial Networks (GANs) custom fitted for IoT situations. By learning the typical behavior designs from the created information, the IDS can successfully identify deviations characteristic of potential security breaches. As we went on with our refinement, we used Random Forest computation, which gave the best result with a precision of 1.00, and the Decision Tree, which gave an accuracy of 99.9%.

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

Enhancing Cybersecurity with Generative Adversarial Networks (GANs): A Novel Approach to Intrusion Detection in IoT Attacks

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Aishwarya Priya,
  • Diya Jeph,
  • Aarushi Garg,
  • Deepak Kumar Sharma,
  • Koyel Datta Gupta

摘要

The Internet of Things (IoT) has revolutionized different businesses, advertising unparalleled comfort and proficiency. In any case, this broad appropriation has too uncovered IoT frameworks to various security dangers, counting malevolent interruptions, and information breaches. Conventional Intrusion Detection Systems (IDSs) regularly battle to keep up with the energetic and complex nature of these assaults, requiring inventive approaches to upgrade IoT security. This paper proposes a novel IDS system utilizing Generative Adversarial Networks (GANs) custom fitted for IoT situations. By learning the typical behavior designs from the created information, the IDS can successfully identify deviations characteristic of potential security breaches. As we went on with our refinement, we used Random Forest computation, which gave the best result with a precision of 1.00, and the Decision Tree, which gave an accuracy of 99.9%.