Scholarlink

Despite the perceived security of wireless networks to be secure from several threats, they are vulnerable to more serious threats such as DDoS attacks. This research explores the feasibility of using artificially generated network datasets to train predictors to predict continuous threats targeting wireless networks. Over and over again, hackers take advantage of these networks’ complex and constantly evolving nature, with varied weaknesses originating from various causes. Several sectors are critical infrastructures, often the first to be attacked because of their significance and the devastating impacts they bring when shaken. Among the various network security threats, two have been selected in this paper: DDoS attacks and ARP Spoofing because of their high prevalence and impact. DDoS attacks suddenly flood the network with constant traffic; thus, the network and computer resources cannot respond to genuine users. ARP Spoofing enables the attacker to intercept or alter messages between two devices in the same broadcast domain. These attacks cause interference with the actual functioning of the network and make other confidential information in the network open for further attacks. Thus, to mitigate these threats, the current study utilizes a Random Forest algorithm combined with the Principal Component Analysis (RFPCA). This way, only the most relevant features are included, enhancing efficiency and increasing the detection Model's performance. Due to its ability to handle complex data structures in conjunction with PCA for dimensionality reduction, Random Forest is a good framework for identifying such patterns of attacks. Moreover, an entropy-based classification method is also incorporated to fine-tune the detection precision to identify the inconsistency in the network traffic patterns. Therefore, this study reveals an extensive process for using the said approaches in threat identification and prevention in wireless networks. The results are valuable in illustrating how machine learning algorithms may be applied to enhancing the robustness of networks against such threats as DDoS and ARP Spoofing by offering insights for further investigation and a catalog for real-world applications of defending wireless communication systems.

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

Recognition and Classification of ARP Spoofing and DDoS Attack Using Machine Learning Approach

  • 章节

  • 发表日期：

    2025年

查看全文
  • Saswati Chatterjee,
  • Suneeta Satpathy,
  • Deepthi Godavarthi

摘要

Despite the perceived security of wireless networks to be secure from several threats, they are vulnerable to more serious threats such as DDoS attacks. This research explores the feasibility of using artificially generated network datasets to train predictors to predict continuous threats targeting wireless networks. Over and over again, hackers take advantage of these networks’ complex and constantly evolving nature, with varied weaknesses originating from various causes. Several sectors are critical infrastructures, often the first to be attacked because of their significance and the devastating impacts they bring when shaken. Among the various network security threats, two have been selected in this paper: DDoS attacks and ARP Spoofing because of their high prevalence and impact. DDoS attacks suddenly flood the network with constant traffic; thus, the network and computer resources cannot respond to genuine users. ARP Spoofing enables the attacker to intercept or alter messages between two devices in the same broadcast domain. These attacks cause interference with the actual functioning of the network and make other confidential information in the network open for further attacks. Thus, to mitigate these threats, the current study utilizes a Random Forest algorithm combined with the Principal Component Analysis (RFPCA). This way, only the most relevant features are included, enhancing efficiency and increasing the detection Model's performance. Due to its ability to handle complex data structures in conjunction with PCA for dimensionality reduction, Random Forest is a good framework for identifying such patterns of attacks. Moreover, an entropy-based classification method is also incorporated to fine-tune the detection precision to identify the inconsistency in the network traffic patterns. Therefore, this study reveals an extensive process for using the said approaches in threat identification and prevention in wireless networks. The results are valuable in illustrating how machine learning algorithms may be applied to enhancing the robustness of networks against such threats as DDoS and ARP Spoofing by offering insights for further investigation and a catalog for real-world applications of defending wireless communication systems.