文献类型：期刊文献

中文题名：基于自组织特征映射神经网络的局部矢量量化算法

英文题名：A LOCAL LEARNING VECTOR QUANTIZATION ALGORITHM BASED ON SELF-ORGANIZING FEATURE MAPPING NEURAL NETWORKS

作者：缪青[1];高大启[1]

机构：[1]华东理工大学计算机系,上海200237

年份：2006

卷号：23

期号：7

起止页码：108

中文期刊名：计算机应用与软件

外文期刊名：Computer Applications and Software

收录：CSTPCD;;北大核心:【北大核心2004】；CSCD:【CSCD_E2011_2012】；

语种：中文

中文关键词：LVQ;SOFM;LSOFM;隶属度;领域

外文关键词：LVQ SOFM LSOFM Membership degree Neighbourhood

摘要：提出了一种基于自组织特征映射神经网络的局部矢量量化算法(Local vector quantization algorithm based on Self-O rgan i-zing Feature M app ing neural networks,LSOFM),LSOFM算法是对SOFM算法的一种改进,它将隶属关系引入到参考点权值的修改中,自组织特征映射神经网络的领域大小的确定依赖于训练矢量与参考点之间的隶属关系。
This paper presents a kind of local learning vector quantization algorithm which bases on self-organizing feature mapping neural networks （LSOFM）. Improving on SOFM, The LSOFM algorithm introduces subjection relation into the process of prototype design. The domain used in self-organizlng feature mapping neural networks depends on the subjection relation between the learning vector and the prototype.