Homework 2
Homework 2
線性模式作業二 080501
β j β j' j j' are estimable, prove that the cell means μ ij μ α i β j are
integration by parts.
0 1 2
5. Let Y a BX with a 1 , B 1 0 , and X ~ N2 (0, I 2 ) .
1 0 1
Homework 2
2. Since i p and j q are estimable for all i p, j q, i 1...a , j 1...b.
0a 0b
1
1
0b
a
C I a 1 0b1 , = a , r (C ) a b 2
1
0a 1b
0
a 1 I b1
b
Let two-way crossed model matrix is X .
we have r (C ) a b 2 r ( X ) r ( X ) a b 1
Thus for all i 1...a, j 1...b, at least one p q is estimable for some p, q.
and we know linear combinations of estimable functions are estimable.
i j ( p q ) ( i p ) ( j q )
must be estimable i, j.
3. Let Py X GLS
0
X ( X V 1 X ) X V 1 y
we have known that P X ( X V 1 X ) X V 1 is a projection matrix onto C ( X ).
Claim Py , ( I P ) y 0
Py , ( I P ) y y PV 1 ( I P ) y
X ( X V 1 X ) X V 1 y V 1 I X ( X V 1 X ) X V 1 y
y (V 1 ) X ( X V 1 X ) X V 1 y y (V 1 ) X ( X V 1 X ) X V 1 X ( X V 1 X ) X V 1 y
y (V 1 ) X ( X V 1 X ) X V 1 y y (V 1 ) X ( X V 1 X ) X V 1 y
0 done!
4. Pr[U c | U ~ k2 2 ]
k 2 x
1
x 2 e 2
k 2
dx
c
k 2 2
( )2
2
k k
k 2 1
u x 2 du x dx
Let 2
x
x
dv e dx
2 v 2 e 2
3
線性模式作業二 080501
k x k x k c k x
1 1
2 x 2 e 2 x2 e 2 c2e 2 x2 e 2
k 2
k k 2
k
k 2
dx
c c
k 2 2 k 2 k 2 2 k
( )2 ( )2 2 ( )2 ( )2 2
2 c 2 2 2
k c
2
c e 2
k
Pr[U C | U ~ k2 ]
k 2 2
( )2
2
k c
2
c e 2
Pr[U c | U ~ k2 2 ] Pr[U C | U ~ k2 ] k
.
k 2 2
( )2
2
5. Y a BX , X ~ N (0, I )
2 2
0 5 1 2
Y ~ N 3 ( a, BB) N 3 1 , 1 1 0
1 2 0 1
a).
V cov(Y ) BB r (V) 2 N (V ) span 1 1 2 .
b).
by (a) (1,1, 2) is one of the normal vector of Y
Y1 Y2 2Y3 c, the space though (0, 1,1) Y1 Y2 2Y3 3.
c). Pr(2Y1 Y2 2)
Note that: the linear combination of multivariate normal distribution
still normal distribution.
Let W 2Y1 Y2
w 2 0 ( 1) 1, w2 4 5 1 4 ( 1) 25
3
W ~ N (1, 23) Pr(W 2) Pr( Z ) 0.7257.
25
d). Pr(Y1 Y2 2Y3 0) 0
Y distribution lies on Y1 Y2 2Y3 3
the space is parallel to Y1 Y2 2Y3 0.