摘要
分析了人工免疫网络聚类基本原理,论证了模糊计算方法在聚类中的准确性及高效性,提出了将模糊计算应用于免疫网络的智能动态聚类算法。通过引入"对阈值的自动确定"和"对抗体群的自动进化机制",避免了外部参数对聚类结果的人为影响,使聚类结果根据期望的聚类数目自动调整,记录调整依据,增强了实用性和移植性。实验结果表明,算法能有效地发现指定数目的聚类结果,在抗体群的进化过程中,提供了发展趋势和决策依据。
By analyzing the basic principles of artificial immune network, accuracy and high performance of fuzzy computation for information intelligent processing, an intelligent algorithm of fuzzy dynamic clustering based on immune network is proposed. By importing the principles of auto-.iu'~tif~,ing threshold and auto-evolution, the effect of man-made outer parameters is avoided, and while the results are automatically asserted on charge of the required clustering number, the reasons are recorded, by which the practicability and replant ion of the algorithm are both enhanced. Application results of cooperation credit rating show the searching of clustering is rational and feasible, while the developing trend and the reference of decision-making are also intelligently provided through the procedures of evolution.
出处
《计算机工程》
CAS
CSCD
北大核心
2007年第16期168-171,共4页
Computer Engineering
基金
四川金财科技集团基金资助项目
关键词
模糊人工免疫网络
聚类
亲和力
最大生成树
资信评估
fuzzy artificial immune network
clustering
affinity
maximum-spanning tree
credit rating