基于模糊C-均值的原型模式选择及其在核爆地震识别中的应用
详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文
摘要
协同模式识别是一种全新的有着抗噪声、抗缺损等诸多优良特性的模式识别方法,但将其用于核爆地震和天然地震的分类时,采用现有的原型模式选择方法识别效果并不理想。本文提出了一种基于模糊C-均值的原型模式选择方法,该方法首先对每一类训练样本采用模糊C-均值聚类的方法聚为c类,然后选取这c类的c个重心或c个聚类中心作为该类的原型模式进行核爆地震的协同模式识别。实验结果表明,同现有的原型模式选择方法相比,该方法使识别率有了较大提高。
The synergetic pattern recognition is a new way of pattern recognition with many excellent features such as noise resistance and deformity resistance.But when it is used in the discrimination between nuclear explosion and earthquake using existing methods of prototype selection,the results are not satisfying.A new method of prototype selection based on FCM is proposed in this paper.First,each group of training samples is clustered into c groups using FCM;then c barycenters or centers are chosen as prototypes.Experiment results show that compared with existing methods of prototype selection this new method is effective and it increases the recognition ratio greatly.
引文
[1]Haken H.协同计算机和认知——神经网络的自上而下方法[M].杨家本,译.北京:中国计算机学会学术著作丛书,清华大学出版社,1994.
    [2]Wagner T,Bobel F G.Testing synergetic Algo-rithms with Industrial Classification Problems[J],Neural Networks,1994,17(8):1313-1321.
    [3]王海龙,戚飞虎.基于聚类法的协同神经网络学习算法[J].上海交通大学学报,1998,32(10):39-41.
    [4]王海龙,戚飞虎.基于信息叠加的学习算法.红外与毫米波学报,19(6),2000:205-208.
    [5]王海龙,戚飞虎,詹劲峰.基于遗传算法的原型模式选择算法[J].计算机工程,2000:26(10),19-20.
    [6]宋仔标,李夕海,刘代志.基于主成分分析的原型模式选择及其应用[C].国家安全与军事地球物理学术研讨会,西安,2005:91-93.
    [7]孙即祥,等.现代模式识别[M].长沙:国防科技大学出版社,2001:268-269.
    [8]何志文,李夕海,刘代志,等.基于相关性分析的特征选择方法研究[J].核电子学与探测技术,2005:25(6),729-731,749.
    [9]Daizhi Liu,Xihai Li,and Bin zhang.Feature Se-lection and Identification of Underground NuclearExplosion and Natural Earthquake Based on Gam-ma Test and BP Neural Network[P].Advances inNeural Networks-ISNN 2005,J.Wang,X.Liao,and Z.Yi(Eds.),LNCS 3497,2005:393 398.

版权所有:© 2023 中国地质图书馆 中国地质调查局地学文献中心