节点文献
基于矩生成函数的多元响应降维子空间估计
Estimate of dimension reduction subspace for multivariate responses based on moment-generating function
【摘要】 提出一类新的估计多元响应降维子空间方法,即基于矩生成函数(GF)、绝对值矩生成函数(G-A)与矩生成函数海塞主方向方法(G-PHD).给出了该方法估计量的相合性以及渐近性性质,并进行了实例模拟.
【Abstract】 A new class of estimators for dimension reduction subspace with multivariate responses are proposed based on some related literature,which are termed as moment-generating function(GF),absolute moment-generating function(G-A)and moment-generating function of principal Hessian directions(G-PHD)respectively.Consistency and asymptotic property of estimators are given.
【关键词】 充分降维;
切片逆回归;
矩生成函数;
绝对值矩生成函数;
海塞主方向;
【Key words】 sufficient dimension reduction; sliced inverse regression; moment generating function; absolute moment generating function; principal Hessian directions;
【Key words】 sufficient dimension reduction; sliced inverse regression; moment generating function; absolute moment generating function; principal Hessian directions;
【基金】 国家自然科学基金资助项目(61473329);福建省自然科学基金资助项目(2015J01009);福建省中青年教师教育科研项目(JAT160566)
- 【文献出处】 东北师大学报(自然科学版) ,Journal of Northeast Normal University(Natural Science Edition) , 编辑部邮箱 ,2017年01期
- 【分类号】O212
- 【下载频次】8