基于免疫优势的克隆选择聚类算法

 基于克隆选择原理和免疫优势理论，本文提出一种新的基于免疫优势的克隆选择聚类算法（Immunodomaince based Clonal Selection Clustering Algorithm，IDCSCA），该算法通过在经典的克隆选择算法框架中，引入基于免疫优势理论的免疫优势算子实现了在线自适应动态获得先验知识和个体间的信息共享。新算法首先通过对群体中若干最优抗体的分析，提取免疫优势，然后将其推广到整个抗体群，通过在进化过程中利用积累的先验知识，在保证抗体种群多样性的基础上加快收敛速度。采用个5个数据集对算法性能进行了测试，与模糊C均值算法（Fuzzy C-means, FCM）、基于遗传算法的模糊聚类算法(Genetic Algorithm based Fuzzy C-means, GAFCM)以及基于克隆选择的模糊聚类算法（Clonal Selection Algorithm based Fuzzy C-means, CSAFCM）比较，结果表明IDCSCA能有效避免聚类中心迭代过程中陷入局部最优点的问题，而且聚类性能更稳定。

Immunodomaince Based Clonal Selection Clustering Algorithm

Based on clonal selection principle and the immunodominance theory,a new immune clustering algorithm,Immunodomaince based Clonal Selection Clustering Algorithm(IDCSCA) is proposed in this paper.An immunodomaince operator is introduced to the clonal selection algorithm,which can realize on-line gaining priori knowledge and sharing information among different individuals.Firstly,the gene of elites in antibody population can be extracted and generalized to ordinary antibodies,by the interaction,the whole antib...