h******3 发帖数: 190 | 1 Help... Somehow I am stuck and cannot get the expected answer which is -xbar
*var(slope estimate) | k*******a 发帖数: 772 | 2 inthat=ybar-xbar*slopehat
so cov(inthat,slopehat)=cov(ybar,slopehat)-xbar*var(slopehat)
可以证明第一项=0
因为 ybar=sigma(yi)/n
slopehat=sigma(ki*yi)
ki=(xi-xbar)/Sx^2
所以第一项~sigma((xi-xbar)*var(yi))
因为var(yi)都一样,而且sigma(xi-xbar)=0
所以第一项=0 |
|