CAJ | 학술논문

在生物免疫系统中抗体的产生受抗原免疫原性的影响，对此，提出一种非均匀免疫网络的聚类算法：将免疫原性引入到人工免疫网络的进化过程中，生成一个反映数据样本特征的非均匀免疫网络。实验结果证明该算法对数据聚类中类边界不清晰以及有数据噪声状况有很好的辨别效果。
재생물면역계통중항체적산생수항원면역원성적영향，대차，제출일충비균균면역망락적취류산법：장면역원성인입도인공면역망락적진화과정중，생성일개반영수거양본특정적비균균면역망락。실험결과증명해산법대수거취류중류변계불청석이급유수거조성상황유흔호적변별효과。
In this paper an inhomogeneous artificial immune net(IaiNet) for clustering is presented according to the immunogenicity of the antigen in biology immune system. The algorithm introduces the immunogenicity of the antigen into the evolution of artificial immune net(aiNet). The algorithm treats the data sample as the antigen, and treats the characteristic of the data sample as the immunogenicity of the antigen. The resultant network is an inhomogeneous internal image which reflects the characteristic of the data samples. The test results show that the algorithm achieves better performances on the clustering problem of the ambiguity clusters boundary and the noises.