空间非稳定性模型及其在产学联盟研发创新中的应用

Spatially non-stationary model and its application in industry-university alliance R&D innovation

针对经典常系数计量经济模型采用普通最小二乘估计(OLS)的不足,提出了空间非稳定性模型及基于加权最小二乘法的空间变系数地理加权回归估计,并将该模型应用于中国省域企业研发和产学联盟研发创新的实证研究中.通过建立的产学联盟研发创新空间非稳定性模型实证估计结果,发现得到的地理加权回归模型(GWR)的拟合优度在96%左右,明显优于OLS估计的拟合优度72%.最后检验了企业自主研发投入及产学联盟研发对企业创新的作用.

In this paper,Geographically Weighted Regression（GWR） of the spatially non-stationary model is proposed.It is based on the regression estimate of weighted least squares（WLS） method.We use the GWR model to analyze the firm＇s RD and industry-university alliance RD innovation of Chinese provinces.According to the tests of GWR model of spatially non-stationary,the fitted degree of the model is about 96%which is obviously better than that of OLS estimation.At last,this paper tests the role of the firm＇s domestic RD investment and industry-university alliance RD to the innovation of Chinese provinces.