线性模型有偏估计的一种新算法  

A new algorithm of biased estimation in linear model

作　　者：梁飞豹[1] 林同华[1] 

机构地区：[1]福州大学数学与计算机科学学院,福建福州350108

出　　处：《合肥工业大学学报（自然科学版）》2016年第6期854-858,共5页Journal of Hefei University of Technology：Natural Science

基　　金：国家自然科学基金资助项目(11301084);福建省自然科学基金资助项目(2014J01010)

摘　　要：文章以降低预测残差平方和为目标,基于岭估计增大回归系数矩阵的对角元素的思想,提出一种利用高斯消去变换工具的线性模型参数估计法,并进行数据模拟实验,最后通过平均预测残差平方和以及平均残差平方和的箱线图来对比新算法和最小二乘估计及岭估计的优良性,说明满足一定条件时,新算法在估计精度和稳定性上优于这2种方法。In order to reduce the prediction residual sum of squares ,and based on the method of ridge estimation that increases the diagonal elements of the regression coefficients matrix ,a linear model pa-rameter estimation method using the tool of Gaussian elimination transform is proposed .And the data experiment is carried out .Finally ,its boxplots about the mean prediction residual sum of squares and the mean residual sum of squares are compared with those of the least squares estimation and the ridge estimation ,and the results show that the new algorithm performs better than these two estimations in the estimation precision and stability under certain conditions .

关 键 词：预测残差平方和 参数估计 岭估计 高斯消去 箱线图 

分 类 号：O212.1[理学—概率论与数理统计]