摘要
针对手写体数字提取的特征维数过高且有冗余从而影响识别速度的问题,提出了基于遗传算法的高维特征选择方法.遗传算法采用类内类间比作为适应度函数,识别率高但速度较慢;而对手写体数字识别的仿真实验表明,特征选择方法虽然识别率在一定程度上有所下降,但提高了识别速度.
Aimed to the phenomenon that the extracted feature dimension of the handwritten numeral is too high and redundant, a high-dimensional feature selection method was proposed using genetic algorithms whose fitness function is the ratio of intra-class and inter-class, which has high recognition rate but low speed. The simulation results on the handwritten digital recognition showed that although the rocognition rate of feature selection decreased to some extent, the speed of the recognition increased.
出处
《郑州轻工业学院学报（自然科学版）》
CAS
2010年第2期75-78,共4页
Journal of Zhengzhou University of Light Industry:Natural Science
关键词
类内类间比
特征选择
遗传算法
手写体
数字识别
the ratio of intra-class and inter-class
feature selection
genetic algorithm
handwritten numeral
digital recognition
作者简介
吴进文（1985-），女，河南省禹州市人，河南财经学院硕士研究生，主要研究方向为模式识别
