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.
出处
《纯粹数学与应用数学》
CSCD
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)
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
semiparametric regression model, least absolute deviation, local linear estimation,algorithm, robustness
