Comparison of Estimates on Covariance Matrix
-
摘要: 本文在一般线性回归模型误差异方差情况下,通过计算机模拟对回归系数最小二乘估计的协方差矩阵的估计进行了比较.结果表明,当样本大小大于50时,回归系数的最小二乘估计具有较高的估计精度;其协方差矩阵的五种估计以普通最小二乘估计的协方差矩阵为最优.Abstract: For linear regression models with heteroscedastical errors, this paper compares the estimates of covariance matrix on the least square estimate of regression coefficiencies by computer simulation. Our research shows that when sample size is greater than 50, the least square estimate of regression coefficiencies has high estimate precision. Among five estimates of its covariance, that of the ordinary least squares estimate is the best.
-
计量
- 文章访问数: 7
- HTML全文浏览量: 2
- PDF下载量: 3
下载:
