记忆克隆分类及其在齿轮箱故障识别中的应用*

作者：李刚，庄健，侯洪宁，于德弘 日期：2009-08-18/span> 浏览：3093 查看PDF文档

记忆克隆分类及其在齿轮箱故障识别中的应用*

李刚，庄健，侯洪宁，于德弘
（西安交通大学 机械工程学院，陕西 西安 710049）

摘要：为了高效地挖掘分类模型，提出了一种基于抗体克隆选择学说和免疫记忆理论的有监督分类算法MCIC。采用模糊C均值聚类产生的初始抗体和样本特征维的信息熵权重为算法提供了先验知识；在抗体种群进行全局搜索的基础上，设计了动态记忆单元局部搜索算子，用以加快抗体亲和力成熟速度；最终，根据最近邻原则实现了未知样本的类属划分，并采用美国加州大学标准数据集中的4个数据集分类和齿轮箱故障识别问题进行了仿真实验。研究结果表明，与CLOALG算法相比，该算法不仅具有更快的收敛速度，而且获得了更高分类准确率。
关键词：人工免疫系统；记忆克隆；有监督学习；数据分类
中图分类号：TP183文献标识码：A文章编号：1001-4551(2009)08-0004-05

Memory clonal based supervised artificial immune classifier and its application
LI Gang, ZHUANG Jian, HOU Hongning, YU Dehong
(School of Mechanical Engineering, Xian Jiaotong University, Xian 710049, China)
Abstract: To efficiently mining the classification model, a novel classification algorithm was put forward based on antibody clonal selection and immune memory principle. Initial antibody population from fuzzy C means clustering and feature weightings computed through the information entropy were employed to provide priori knowledge. To improving the hyper mutation rate, a dynamic memory unit evolvement operator was designed for finetune search while the antibody population carry out global search. Finally classification was performed in a nearest neighbor approach. Experimental results on four benchmark datasets from UCI data set repository and gear box failure data demonstrate that, compared with CLOALG artificial immune classifiers, the new classifier not only has faster convergence speed, but also can achieve higher classification accuracy.
Key words: artificial immune systems; memory clonal; supervised learning; data classification
参考文献(References):
［1］MITRAKIS N E, THEOCHARIS J B, PETRIDIS V. A multilayered neurofuzzy classifier with selforganizing properties［J］. Fuzzy Sets and Systems,2008,159(23):3132-3159.
［2］KHOSHGOFTAAR T M, LIU Y. A Multiobjective software quality classification model using Genetic programming［J］. IEEE Transactions on Reliability,2007,56(2):237-245.
［3］ZHONG Y F, ZHANG L P, GONG J Y. A supervised artificial immune classifier for remotesensing imagery［J］. IEEE Transactions on Geosciences and Remote Sensing,2007,45(12):3957-3966.
［4］张进华，张鄂.基于免疫算法的圆柱螺旋弹簧模糊优化设计［J］.机电工程,2004,21(2):45-48.
［5］TIMMIS J. Artificial immune systemtoday and tomorrow［J］. Nat. Comput.,2007(6):1-18.
［6］DE CASTRO L N, VON ZUBEN F J. Learning and optimization using the clonal selection principle［J］. IEEE Transactions on Evolutionary Computation,2002,6(3):239-251.
［7］WATKINS A, TIMMIS J, BOGGESS L. Artificial immune recognition system (AIRS): an immuneinspired supervised learning algorithm［J］. Genetic Programming and Evolvable Machines,2004,5(3):291-317.
［8］HART E, TIMMIS J. Application areas of AIS: The past, the present and future［J］. Applied Soft Computing,2008(8):191-201.
［9］焦李成,杜海峰,刘芳,等.免疫优化计算、学习与识别［M］.北京:科学出版社,2006.
［10］邱菀华.管理决策与应用熵学［M］.北京:机械工业出版社,2002.
［11］NEWMAN D J, HEITICHS, BLAKE C L, et al. UCI Repository of Machine Learning Databases［EB/OL］.［1998-07-09］.http://archive.ics.uci.edu/ml/.
［12］陈强,郑德玲.基于免疫原理的齿轮箱故障检测和诊断方法研究［J］.矿山机械，2005,33(5):75-77.
［13］张军，陆森林，和卫星，等.滚动轴承振动诊断的BP神经网络方法［J］.轻工机械，2007，25（2）：90-93.