基于半参数回归模型的最小一乘局部线性算法  

Least absolute deviation local linear algorithm based on semiparameteric regression model

作　　者：苏正军[1] 刘迎照[1] 

机构地区：[1]洛阳师范学院数学科学学院,河南洛阳471022

出　　处：《纯粹数学与应用数学》2013年第5期513-519,共7页Pure and Applied Mathematics

基　　金：河南省基础与前沿技术研究计划项目(102300410216)

摘　　要：根据最小一乘准则,推导出最小一乘局部线性估计的计算方法,并通过对模拟数据的计算和分析,对比最小一乘核算法和最小二乘局部线性算法,验证了最小一乘局部线性算法是一种有效的,稳健的估计方法,并且有降低边界效应的作用.Based on the least absolute deviation estimation, local linear least absolute deviation algorithm is derived. The effectiveness and robustness of our method are verified by simulation compared with the least absolute deviation kernel algorithm and local linear least squares algorithm. The model can also reduce the boundary effect.

关 键 词：半参数回归模型 最小一乘 局部线性估计 算法 稳健性 

分 类 号：O242.1[理学—计算数学]