华东交通大学学报Issue(4)：59-64,6.
非线性LTS估计的截断凝聚光滑化方法
Truncated Aggregate Smoothing Method for Nonlinear LTS Estimator
摘要
Abstract
The computing of the nonlinear least trimmed squares (LTS) estimator is considered. LTS is a robust esti-mator and can be converted to a min-min non-convex and non-smooth programming problem. For the data set with size m , the objective function is the minimum of all the m͂-subsets' residual sum of squares. Even if m is not big, the number of the subsets may be very large which makes computing LTS estimator difficult. For such a special kind of problem, an appropriate truncated criteria standard is given and then an efficient truncated smooth-ing Newton method is proposed. The numerical results show the efficiency.关键词
LTS估计/凝聚函数/截断凝聚光滑化Key words
LTS estimator/aggregate function/truncation smoothing分类
数理科学引用本文复制引用
肖瑜..非线性LTS估计的截断凝聚光滑化方法[J].华东交通大学学报,2014,(4):59-64,6.基金项目
国家自然科学基金资助项目 （）
