Section 7.2 (p.385)
1.
.
If n=1, log f(x) = const.
- ½ log x – x/2. It doesn’t have a mode because that as x ® 0, log f(x) increases without a bound.
If n=2, log f(x) = const.
- x/2. The mode is 0.
If n ³ 3, let
. We have that x = n-2 maximizes log f(x). So the mode is
n-2.
2. Try to draw them yourself.
3. We want to find the smallest radius r such that Pr(X2
+ Y2 £ r2) ³ 0.99. Since X2 + Y2 has a c2 distribution
with 2 degrees of freedom (Theorem 2), r2 ³ 9.210 based on table on p.690.
4. X2 + Y2 + Z2 has a c2 distribution
with 3 degrees of freedom. Pr(X2 + Y2 + Z2 £ 1) »
0.20 based on the Chi-square table.
6. Fi (Xi) has a uniform distribution on
(0,1) (see p.154 for details).
Let Zi = -2
log Fi (Xi), we can show that Zi has pdf f(z) = ½ e-z/2, for z > 0 (You
should be able to figure this out by yourself. If not, review Sec. 3.8). In
other words, Zi has a Gamma distribution with parameters a = 1 and b
= ½. From Theorem 3. of Sec. 5.9, we know that Y = -2 S log Fi (Xi) = S Zi has a Gamma distribution with a = n and b
= 1/2., or a c2 distribution
with 2n degrees of freedom
8.
has a standard normal distribution. Therefore
has a c2 distribution
with 1 degrees of freedom
9. Both (X1 + X2 + X3) and (X4
+ X5 + X6) have normal distributions with mean 0 and
variance 3. Therefore (X1 + X2 + X3)2/3
and (X4 + X5 + X6)2/3 both have c2 distributions
with 1 degrees of freedom. Let c = 1/3, Y has a c2 distributions
with 2 degrees of freedom.