An improved real-valued negative selection algorithm for anomaly detection

作者：Zeng Jinquan^*; Tang Weiwen; Liu Caiming; Hu Jianbin; Peng Lingxi

来源：International Journal of Advancements in Computing Technology, 2012, 4(16): 453-459.

DOI：10.4156/ijact.vol4.issue16.53

• Anomaly detection
• Artificial Immune System
• Classification algorithm
• Negative selection algorithm
• Real-valued negative selection algorithms
• Real-world datasets
• Training phase

Conventional classification algorithms need samples for all classes (e.g. normal and abnormal) during the training phase. However, in many anomaly detection application, we can only get a class samples (commonly, normal samples) to train detectors. This paper presents an improved negative selection algorithm (NSA) and its applications to anomaly detection. This approach can only use normal samples to train detectors; at the same time, in order to overcome the drawbacks of the current representation of the self-space in NSAs, a new scheme of the representation of the self-space is introduced with variable-sized self-radius. In our experiments, this approach is tested using the wellknown real world datasets; preliminary results show that the new approach improves the overall performance of detectors and without increase in complexity.

• 出版日期2012
• 单位电子科技大学