学习过程收敛速度的界是统计学习理论的重要组成部分,这些界决定了学习机器的推广能力.以机会理论和Hybrid变量的概念为基础,讨论了基于Hybrid样本的学习过程一致收敛速度的界,并给出了这些界和函数容量之间的关系.
Bounds on the rate of uniform convergence of learning process are important component part of statistical learning theory and the bounds determine the generalization abilities of learning machines.Based on the chance theory and the definition of Hybrid variable,bounds on the rate of uniform convergence of learning process based on hybrid samples are discussed and the relationship between the bounds and the capacity of the set of functions is given.
[1] SUN Xiaojing,WANG Chao,HA Minghu,et al.The key theorem of learning theory based on hybrid variable[C] //Guilin:Proceedings of the International Conference on Machine Learning and Cybenetics,2011:1141-1145.
[2] LIU Baoding.Uncertainty theory [DB/OL].(2010)[2015-09-20].http://www.doc88.com/p-34688422/032.html.
[3] VAPNIK V N.Statistical learning theory[M].New York:A Wiley-Interscience Publication,1998.
[4] 张学工.关于统计学习理论与支持向量机[J].自动化学报,2000,26(1):32-44. ZHANG Xuegong.Introduction to statistical learning theory and support vector machines[J].Acta Automatica Sinica,2000,26(1):32-44.
[5] 哈明虎,王超,张植明,等.不确定统计学习理论[M].北京:科学出版社,2010.
[6] 哈明虎,王鹏.可能性空间中学习过程一致收敛速度的界[J].河北大学学报(自然科学版),2004,24(1):1-6. HA Minghu,WANG Peng.Bounds on the rate of uniform convergence of learning process on possibility spaces[J].Journal of Hebei University(Natural Science Edition),2004,24(1):1-6.
[7] 田景峰,张植明.可信性空间上基于复模糊变量的学习过程一致收敛速度的界[J].华北电力大学学报,2009,36(5):106-112. TIAN Jingfeng,ZHANG Zhiming.The bounds on the rate of uniform convergence of learning process based on complex fuzzy variable on credibility space[J].Journal of North China Electric Power University,2009,36(5):106-112.
