【摘要】 基于对多级免疫学习算法(M ILA)的批评性研究,提出了单级免疫学习算法(SILA)。该算法适用于低维度特征量的异常检测,提高了探测器训练的效率和效益,而且对M ackey-G lass时间序列数据的检测取得了很好的实验结果。更多还原

【Abstract】 Anomaly detection is one of the main issues in ensuring computer security.Various artificial immune system(AIS) algorithms,from negative selection algorithm (NSA) to multilevel immune learning algorithm(MILA),are therefore developed to serve this purpose.Based on a critical study of the MILA approach,this paper proposes a single-level immune learning algorithm(SILA),which extends the ideas of MILA and pays more attentions to the problem space.The proposed algorithm contributes mainly to(improving) the effectiveness and efficiency of detector training,which is of great concern in all artificial(immune) systems.更多还原