摘要
该文研究了小样本统计学习在生物识别中的数学模型估计问题,并探讨了先验风险优化替代实际风险优化的可行性以及机器学习的深度与广度的矛盾性,最后在小样本采集,识别,建立样本数据库等方面进行了分析研究。
In the paper we research the mathematics model of small sampling statistics study on biological recognition, then we study the possibility of using priori risk optimization instead of the actual risk optimization and the contradiction between the deepness and the extent of machine study, finally analyze the small sample collection, the recognize and the building sample data-base etc.
出处
《电脑知识与技术》
2017年第6X期178-179,181,共3页
Computer Knowledge and Technology
关键词
统计学习
学习深度
SVM
小样本采样
statistics study
study depth
SVM
small sample collection
