要用到这个定理：
When x ~ Normal(mu,S), then t(x)*A*x ~ chisq(df=r, noncentral parameter= lamda)
if and only if ASAS=AS, where r= rank(A), and lamda=t(mu)*A*mu/2
绿色的荷叶 发表于 2010-2-19 02:03
书上说total sum of squares TSS，和ESS(Error sum of squares) , Regression sum of squares 都是服从卡方分布的，自由度分别是n-1, n-k, k-1.
怎样证明啊？
求各位大师指点~~~
急求~~~
谢谢了
sum of Normal ^2 is chi square. Some book put it as a definition. This is so basic theorem in Statistics. You can find answer (proof) in the link below.
扫码加好友，拉您进群
全部版块
我的主页
收藏
