Questions tagged «prediction»

3
当响应变量为时预测
我的估计模型是 ln^(yt)=9.873−0.472ln(xt2)−0.01xt3ln^(yt)=9.873−0.472ln⁡(xt2)−0.01xt3\hat \ln(y_t)=9.873-0.472\ln(x_{t2})-0.01x_{t3} 当和时,我被要求以的平均值找到95%置信度的预测CI 。我们假设,其中。y0y0y_0x02=250x02=250x_{02}=250x03=8x03=8x_{03}=8s2x0(XTX)−1xT0=0.000243952s2x0(XTX)−1x0T=0.000243952s^2 x_0(X^TX)^{-1}x_0^T=0.000243952x0=(250,8)x0=(250,8)x_0=(250,8) 我有上一年的解决方案,如下所示: 我发现 ,其中是分布的位数,而 。这给了我。CI(E[ln(y0)|x0])=[ln^(yt)−tα/2sE,ln^(yt)+tα/2sE]CI(E[ln(y0)|x0])=[ln^(yt)−tα/2sE,ln^(yt)+tα/2sE]\text{CI}(E[ln(y_0)|x_0])=\left[\hat\ln(y_t)-t_{\alpha/2}s_E,\hat \ln(y_t)+t_{\alpha/2}s_E\right]tttα/2α/2\alpha/2t(n−k)t(n−k)t(n-k)sE=0.000243952−−−−−−−−−−√sE=0.000243952s_E=\sqrt{0.000243952}[7.1563,7.2175][7.1563,7.2175][7.1563,7.2175] 然后作者执行 。CI(E[y0|x0])=[e7.1563,e7.2175]=[1282.158,1363.077]CI(E[y0|x0])=[e7.1563,e7.2175]=[1282.158,1363.077]\text{CI}(E[y_0|x_0])=[e^{7.1563},e^{7.2175}]=[1282.158,1363.077] 我不同意最后一步(由于詹森的不平等,我们将低估了这一点)。在Wooldridge的《计量经济学概论》(第212页)中,他指出,如果我们确定误差项是正常的,则一致的估计量为: E^[y0|x0]=es2/2eln^(y0)E^[y0|x0]=es2/2eln^(y0)\hat E[y_0|x_0]=e^{s^2/2}e^{\hat \ln(y_0)} 所以,我在想做 CI(E[y0|x0])=[es2/21282.158,es2/21363.077]=[1282.314,1363.243]CI(E[y0|x0])=[es2/21282.158,es2/21363.077]=[1282.314,1363.243]\text{CI}(E[y_0|x_0])=\left[e^{s^2/2} 1282.158,e^{s^2/2}1363.077 \right] = \left[ 1282.314,1363.243 \right] 它是否正确? 另外,此练习的解决方案指出,这与我得到的任何一种解决方案都相去甚远。CI(E[y0|x0])=[624.020,663.519]CI(E[y0|x0])=[624.020,663.519]\text{CI}(E[y_0|x_0])=[624.020,663.519] 任何帮助,将不胜感激。 PS:我也读到,校正不应该用于CI,而只能用于点估计E^[y0|x0]E^[y0|x0]\hat E[y_0|x_0]
By using our site, you acknowledge that you have read and understand our Cookie Policy and Privacy Policy.
Licensed under cc by-sa 3.0 with attribution required.