模式识别的改进算法研究

Study of improvement algorithm for pattern recognition

针对模式识别问题,设计了一种基于字符串编码的匹配算法,实现了对样本的二次筛选和两种不同的模式匹配算法的交叉匹配,不仅可以有效地减少"洞"的存在,而且可以生成高质量的检测样本,从而增强识别效率、降低时间消耗。对数字识别的实验结果表明,该匹配算法具有很高的识别率,能有效降低错误率,得到稳定的满意解,具有良好的适应性、学习性、开放性和鲁棒性。

In this paper,we designed a kind of matching algorithm based on string coding,to realize the sample quadratic sieve and the crossmatching of two different matching algorithms.This designed algorithm could avoid the ＂hole＂effectively,generate detection samples of high quality ,strength the identification efficiency and reduce the time wasted.Experiments result shows that this algorithm has good adqptability,openess and robustness.