摘要
考虑到微型计算机内存空间有限而汉字个数较多的特点,本文提出了一个适合于微型计算机处理的汉字特征选择的可分性准则函数.使用此方法进行汉字特征选择能保证不同汉字的分类特征距离较大并且同一汉字聚类性较好,使得汉字特征选取十分合理.
In order to solve the contradiction between the limited memory space in amicrocomputer and the large number of Chinese characters,this paper presents a sepa-rable standard function to extract features of Chinese characters.In the processing offeature selection in a microcomputer,this function is characterized by distinctly differ- entiating the classified features of different Chinese characters and better clustering ofsame Chinese characters.Obviously,this makes the selection of Chinese character fea-tures reasonable.
出处
《哈尔滨工业大学学报》
EI
CAS
CSCD
北大核心
1989年第2期77-84,共8页
Journal of Harbin Institute of Technology
关键词
汉字识别
统计模式识别
特征选择
Printed Chinese character recognition
statistical pattern recognition
feature extraction
feature selection.
