从表1的数值可以看出: 半参数三次非线性回归的估计效果最好。参数线性回归的估计的效果最差。其次, 半参数模型线性回归与半参数模型非线性的估计效果大致相当, 非线性的效果略微好点, 二次与三次的估计值已基本相等。从比较结果来看, 焦炭>40mm价格半参数模型的估计精度较高, 拟合效果要明显好于参数回归模型。究其原因, 半参数模型的非参数分量起到了调整作用, 降低了模型估计误差。实际估计效果与半参数的理论说明相吻合, 说明半参数模型对焦炭价格估计具有优势广州翻译公司。
From the values in the Table 1 it can be seen that non-linear semi-parametric cubic regression gives the best estimation results. The linear parametric regression gives the poorest estimation results. And, the linear regression of semi-parametric model and the nonlinear semi-parametric model have roughly a similar effect. The nonlinear version gives slightly better results. Estimation values for square and cubic are practically equal. As far as the comparison results are concerned, the semi-parametric model for the coke>40mm offers comparatively better accuracy. Their fit effects are markedly better than the parametric regression model. The causes are the non-parametric parts of semi-parametric model make some adjustment and thus to cut down on the errors of the model. The actual estimation results tally with the theoretical specifications of semi-parameters, which show that the semi-parametric model is advantageous for estimation of coke price.