免疫进化否定选择算法

发布时间：2018-01-19 23:05

本文关键词： 人工免疫否定选择算法检测器免疫进化　出处：《计算机应用研究》2017年05期 　论文类型：期刊论文

【摘要】：当训练样本分布密集交错时,传统的否定选择算法难以将检测器生成在正/反样本间的有效区域,导致检测器集合对这些样本的识别率降低,影响了算法性能。为使检测器能有效地识别分布密集交错的样本,提出了免疫进化否定选择算法(IENSA)。IENSA通过加入两个免疫进化过程,首先在样本分布密集的区域引导检测器在正/反样本之间有效地生成,然后在样本分布稀疏的区域对冗余检测器进行抑制。实验结果表明在二维人工数据集Rectangle与三维标准数据集Skin segmentation上,相对于经典的RNSA与V-detector算法,IENSA均能以较少的检测器达到较高的检测率。
[Abstract]:When the distribution of training samples is dense and interlaced, it is difficult for the traditional negative selection algorithm to generate the detector in the effective region between the positive and inverse samples, which leads to the reduction of the recognition rate of these samples by the detector set. In order to effectively identify the dense and interlaced samples, an immune evolutionary negative selection algorithm (IENSAN. IENSA) is proposed. Two immune evolution processes are added to the algorithm. First, the detector is effectively generated between positive and inverse samples in the region where the samples are densely distributed. Then the redundancy detector is suppressed in the region where the sample distribution is sparse. The experimental results show that in the two-dimensional artificial data set Rectangle and the three-dimensional standard data set Skin. On segmentation. Compared with the classical RNSA and V-detector algorithms, RNSA can achieve higher detection rate with fewer detectors.
【作者单位】：西华师范大学教育信息技术中心;
【基金】：国家自然科学基金资助项目(61402308) 四川省教育厅自然科学重点资助项目(15ZA0146,15ZB0142)
【分类号】：TP301.6
【正文快照】： 0引言生物免疫系统是生物体抵御外来病原体侵害的重要屏障,具有保护自生组织、识别并清除外来有害抗原的能力[1]。受此启发,大量专家学者尝试在计算机系统中模拟生物免疫系统以抵抗来自系统外部的病毒、攻击,并由此逐渐形成了人工免疫这一研究领域。在人工免疫的众多理论模型

【相似文献】