Scholarlink

<p>The article proposes an enhanced negative selection algorithm to detect anomalies in computer programs. It relies on the adoption of a decreasing activation threshold in the binary template algorithm, which classically features a fixed activation threshold. The algorithm has two steps: first, a minimal set of receptors providing the desired probability of anomaly detection is determined. Second, this set of receptors is adapted to an individual computer program and used to train the intrusion detection system. The final set of receptors is applied to detect anomalies. To verify the proper functioning of the algorithm, practical experiments to detect anomalies using the EMBER dataset and VirusShare repository were carried out. The probabilities of anomaly detection of the experimental results are nearly the same as those calculated analytically.</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

Enhancing computer program anomaly detection: a decreasing-threshold approach

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年6月1日

查看全文
  • Krzysztof Wawryn,
  • Patryk Widulinski

摘要

The article proposes an enhanced negative selection algorithm to detect anomalies in computer programs. It relies on the adoption of a decreasing activation threshold in the binary template algorithm, which classically features a fixed activation threshold. The algorithm has two steps: first, a minimal set of receptors providing the desired probability of anomaly detection is determined. Second, this set of receptors is adapted to an individual computer program and used to train the intrusion detection system. The final set of receptors is applied to detect anomalies. To verify the proper functioning of the algorithm, practical experiments to detect anomalies using the EMBER dataset and VirusShare repository were carried out. The probabilities of anomaly detection of the experimental results are nearly the same as those calculated analytically.