Scholarlink

<p>Phishing attacks are a major concern of cyber-attacks of intruding personal information and financial data of the internet users. Website URLs are used to attack the user’s information by replacing the original URLs with malicious ones. These attacks are rising with the increase in the number of internet users. Therefore, as URL is the main target to overcome phishing attacks where many researches are taking place to detect malicious URLs. Machine learning and deep learning algorithms are adopted to efficiently detect phishing attacks in URLs by training and testing them in their own way. CapsNet is one of the CNN-based architectures where the spatial relationship is given more importance. As URLs contain many features it’s very important to obtain the spatial relationship of URL features. The self-attention model helps to obtain the cross-interaction of various features in the URLs. Therefore, this paper proposes the anomaly detection of URLs using self-attention-based CapsNet architecture that detects malicious and benign URLs of accuracy 99.2%. </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

Anomaly detection based capsnet for malicious Url detection system

  • 期刊论文

  • 发表日期：

    2025年5月20日

查看全文
  • Rakan A. Alsowail

摘要

Phishing attacks are a major concern of cyber-attacks of intruding personal information and financial data of the internet users. Website URLs are used to attack the user’s information by replacing the original URLs with malicious ones. These attacks are rising with the increase in the number of internet users. Therefore, as URL is the main target to overcome phishing attacks where many researches are taking place to detect malicious URLs. Machine learning and deep learning algorithms are adopted to efficiently detect phishing attacks in URLs by training and testing them in their own way. CapsNet is one of the CNN-based architectures where the spatial relationship is given more importance. As URLs contain many features it’s very important to obtain the spatial relationship of URL features. The self-attention model helps to obtain the cross-interaction of various features in the URLs. Therefore, this paper proposes the anomaly detection of URLs using self-attention-based CapsNet architecture that detects malicious and benign URLs of accuracy 99.2%.