|
| 摘要: |
| 文章研究半参数变系数空间误差回归模型下的参数估计和变量选择问题.利用了局部多项式的方法对变系数函数进行估计,其次分别构造参数部分和非参数部分的最大经验对数似然比估计,同时使用惩罚经验似然(PEL)进行选择变量,并且用平方再求和的方法估计空间系数及误差项的方差.得到了参数与非参数的估计值,以及PEL方法选择变量的优越性的结论.在合适的条件下,推广了惩罚经验似然估计具有Oracle特征且在零假设下服从渐近卡方分布. |
| 关键词: 部分线性模型|惩罚经验似然|空间自回归|变量误差 |
| DOI： |
| 分类号:O212.7 |
| 基金项目:“行政记录人口普查”的数据质量评估框架研究(21CTJ005);金融资产泡沫的形成、识别及其价格有效性测度研究(72271003). |
|
| PENALIZED EMPIRICAL LIKELIHOOD FOR SEMIPARAMETRIC VARYING COEFFICIENT SPATIAL ERROR REGRESSION MODEL |
|
WANG Yi-heng,HE Bang-qiang
|
| Abstract: |
| The article considers the problem of parameter estimation and variable selection in semi-parametric variable coefficient spatial error regression model. The local linear estimation method was used to estimate the variable coefficient function, then the maximum empirical log-likelihood ratio estimation of parametric and non-parametric components is constructed, and PEL was suggested to select the variables, and the square and sum method was used to estimate the variance of the spatial coefficient and error term. The estimated values of pa-rameters and non-parameters and the superiority of PEL method in selecting variables are obtained. Under suitable conditions, the Penalized empirical likelihood has oracle charac-teristics and obeys the asymptotic chi-square distribution under the null hypothesis. |
| Key words: partially linear model|penalized empirical likelihood|SAR|Errors-in-variables |
