摘要
提出了一种改进的LSM-ALSM子空间模式识别方法,将LSM的旋转策略引入ALSM,使子空间之间互不关联的情况得到改善,提高了ALSM对相似样本的区分能力。讨论中以性能函数代替经验函数来确定拒识规则的参数,实现了识别率、误识率与拒识率之间的最佳平衡;通过对有限字符集的实验结果表明,LSM-ALSM算法有效地改善了分类器的识别率和可靠性。
A new subspace algorithm(LSM-ALSM) is discussed in this paper. By rotating the subspaces of ALSM, LSM changes their independent relations and improves the ALSM classifiers recognition capacity in alike characters. To realize the best balance between recognized rate, rejected rate and error rate, we presents performance function instead of experience function to determine the refuse rules. The experiment with limited character set shows that LSM-ALSM method improves the recognized rate and credibility effectively.
出处
《电子科技大学学报》
EI
CAS
CSCD
北大核心
2004年第1期5-8,共4页
Journal of University of Electronic Science and Technology of China
关键词
学习子空间
性能函数
散布矩阵
最小描述长度
study subspace
performance function
spread matrix
minimum description length
