摘要
提出了对自变量个数大于样本容量的特殊数据的广义岭估计的迭代算法,设计了2种Monte Carlo搜索岭脊方法;通过模拟研究说明了该算法简单、速度快,直观,适用于加权情况,同时第2种搜索岭脊法在估计系数、RSS和PRESS方面具有良好的表现。
Presented herein is an iterative algorithm of the GRE for the data that the number of independent variables is more than the sample size, and two Monte Carlo methods of searching the ridge are designed. Simulation studies show that the algorithm is intuitive, simple, speedy and suitable to weighted estimators,and that the second searching method is superior in estimators, RSS and PRESS.
出处
《合肥工业大学学报（自然科学版）》
CAS
CSCD
北大核心
2007年第8期1065-1068,共4页
Journal of Hefei University of Technology：Natural Science
基金
安徽省教育厅自然科学基金资助项目(2006kj252B)
