论文部分内容阅读
针对现有检测器生成算法存在效率低、自适应性差、生成的检测器集庞大且冗余等问题,借鉴生物免疫系统中抗体的克隆机制和亲和度变异机制,并融合小生境策略以及检测器的变异和优化等,构建基于免疫软件人(ISM)特性的检测器生成算法及模型.与传统算法相比,该算法能够降低检测器的冗余度,减少检测器集的规模,保持检测器的多样性;通过合理地改变其匹配阈值,能够实现以较小的检测器集检测出更多的异常行为的目的.实验结果表明,所提出的算法具有较强的自适应性,且拥有较高的检测效率和性能.
In order to solve the problems of low efficiency, poor adaptability and large detector set generated by existing detector generation algorithms, the cloning mechanism and affinity variation mechanism of antibodies in biological immune system are used for reference, and the niche strategy and detector , This paper constructs a detector generation algorithm and model based on immune software human (ISM) characteristics.Compared with the traditional algorithm, the algorithm can reduce the redundancy of the detector, reduce the size of the detector set, keep the detector By changing the matching threshold reasonably, the purpose of detecting more abnormal behaviors with smaller detector sets can be achieved.Experimental results show that the proposed algorithm has strong self-adaptability, High detection efficiency and performance.