Scholarlink

Advanced persistent threat (APT) attacks represent a major threat to nearly all types of infrastructure because attackers continuously enhance their complex techniques and strategies. Traditional defense mechanisms struggle to detect these APT attacks effectively, making it challenging to safeguard network information. Detection of APT attacks often becomes entangled with other types of attacks, emphasizing the need for error-free detection solutions. This study proposes an intelligent approach, termed “APT-Bi-LSTM-GRU,” aimed at analyzing, identifying, and preventing APT attacks. The method employs RFFI to find the pertinent features before training with various deep-learning models. The accuracy and detection rate of APT-Bi-LSTM-GRU has reached 99.29%, demonstrating an average improvement of 5% over current approaches.

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 Web-Based Advanced Persistent Threat Detection Through Deep Learning Techniques

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Konda Srikar Goud,
  • T. Divya,
  • A. Vennela,
  • V. Sushma

摘要

Advanced persistent threat (APT) attacks represent a major threat to nearly all types of infrastructure because attackers continuously enhance their complex techniques and strategies. Traditional defense mechanisms struggle to detect these APT attacks effectively, making it challenging to safeguard network information. Detection of APT attacks often becomes entangled with other types of attacks, emphasizing the need for error-free detection solutions. This study proposes an intelligent approach, termed “APT-Bi-LSTM-GRU,” aimed at analyzing, identifying, and preventing APT attacks. The method employs RFFI to find the pertinent features before training with various deep-learning models. The accuracy and detection rate of APT-Bi-LSTM-GRU has reached 99.29%, demonstrating an average improvement of 5% over current approaches.