摘要
本文提出的模糊加权距离,反映了模式识别与分类中每一样本的各个分量对“距离”贡献的差异性.这种差异性由本文定义的特征优势的概念所描述,而特征优势的概念则由模糊集理论导出并按照人们的思维方式所确定.本文还就新距离的合理性进行了讨论.实验结果及实用结果表明了新距离的优越性.
The fuzzy weighted distance presented here shows that every feature of every sample has different effect on distance in pattern recognition and classification. The difference is discribed by using the concept of feature odds defind here which is derived from fuzzy set theory and is determined by the way of human thinking in recognizing things.The reasonableness of this new distance is discussed in this paper.The experiment results show the advantages of this new distance.
出处
《北方交通大学学报》
CSCD
北大核心
1990年第2期24-29,共6页
Journal of Northern Jiaotong University
关键词
模糊加权
距离
模式识别
分类
聚类
pattern recognition
weighting
feature extraction
