摘要
针对面板数据插值问题,分别从截面数据和时间序列数据的角度,建立截面数据插值的克立格(Kriging)模型和时间序列数据插值的遗传神经网络(GABP)模型。在此基础上建立面板数据插值的线性组合模型,提出空间漂移度法确定组合插值模型的加权系数。对2007年福建部分市县人均GDP的插值的实证研究结果表明:面板数据的线性组合模型插值效果优于单项模型插值的效果。
A panel data is considered as a set of cross - sectional data or a set of time series data. The Kriging algorithm interpolation model based on cross- sectional data, and genetic algorithm back- propagation （GABP） neural network interpolation model based on time series data are established respectively. A combination interpolation model is constructed by the results of the two models. The weights of the combination model are obtained by spatial drift. An empirical research is carried out by interpolation some cities＇ GDP per capita in Fujian, 2007. It is found that the effective of the linear combination interpolation model is better than that of the single interpolation model on panel data.
出处
《统计与信息论坛》
CSSCI
2009年第4期3-6,共4页
Journal of Statistics and Information
基金
福建省教育厅A类科技项目<空间数据线性组合插值模型研究>(JA08263)
关键词
面板数据
组合插值模型
空间漂移度
panel data
combination interpolation model
spatial drift
