Scholarlink

With the widespread use of the Android operating system, the proliferation of Android malware has become increasingly severe. To effectively address this growing security challenge, we propose an advanced malware detection method—SEQDroid, which focuses on in-depth analysis of API call sequences. SEQDroid employs an innovative sequence optimization technique that removes unofficial nodes and mitigates the imbalance in API sequence length distribution, achieving precise differentiation between benign and malicious software in Android applications through a neural network model. The experimental results show that SEQDroid has achieved 96.11% in accuracy and 96.12% in F1-score, both surpassing current detection technologies. This detection strategy developed in this study not only significantly improves the accuracy of detection but also greatly reduces the false positive rate, contributing an efficient solution to the field of Android malware detection.

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

SEQDroid: A Deep Learning Approach for Android Malware Detection Based on API Sequences

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Zhuoyi Wu,
  • Guojun Wang,
  • Xuelei Liu,
  • Yuheng Zhang,
  • Mingfei Chen,
  • Jinrong Kong,
  • Wanyi Gu

摘要

With the widespread use of the Android operating system, the proliferation of Android malware has become increasingly severe. To effectively address this growing security challenge, we propose an advanced malware detection method—SEQDroid, which focuses on in-depth analysis of API call sequences. SEQDroid employs an innovative sequence optimization technique that removes unofficial nodes and mitigates the imbalance in API sequence length distribution, achieving precise differentiation between benign and malicious software in Android applications through a neural network model. The experimental results show that SEQDroid has achieved 96.11% in accuracy and 96.12% in F1-score, both surpassing current detection technologies. This detection strategy developed in this study not only significantly improves the accuracy of detection but also greatly reduces the false positive rate, contributing an efficient solution to the field of Android malware detection.