| 摘要: |
| 在变系数测量误差模型的估计方法基础上，不假设误差结构类型，利用核实方法估计系数函数；然后根据经典的波士顿住房数据拟合带有测量误差的变系数模型，利用精确测量的小部分无误差的数据核实带有测量误差的住房数据；结果表明：变系数模型能够更好地解释房价中位数的波动趋势，且在实际应用中，核实方法能够处理复杂的误差结构模型以及节省很多成本。 |
| 关键词: 测量误差 替代数据 核实样本 |
| DOI： |
| 分类号: |
| 基金项目: |
|
| Analysis on Varying Coefficient Measurement Errors Model with Boston Housing Data By Validation Method |
|
YIN Wen wen
|
| Abstract: |
| On the basis of the estimation method of varying coefficient model with measurement errors， we estimate coefficient function without assuming any error structures. Then， according to Boston housing data ， we use the varying coefficient model with measurement errors to fit these data with validation methods. There， we use a small part of housing data accurately measured to verify other data measured with errors. The final fitted results also can explain the fluctuation trend of the median housing price very well. And in practical applications， validation method can handle complex error structure models and save much costs. |
| Key words: measurement errors surrogate data validation samples |
