192 Chapter 5: Special Random Variables
following equalities:
α=P
{
χn^2 /n
χm^2 /m
>Fα,n,m
}
=P
{
χm^2 /m
χn^2 /n
<
1
Fα,n,m
}
= 1 −P
{
χm^2 /m
χn^2 /n
≥
1
Fα,n,m
}
or, equivalently,
P
{
χm^2 /m
χn^2 /n
≥
1
Fα,n,m
}
= 1 −α (5.8.3)
But because (χm^2 /m)/(χn^2 /n) has anF-distribution with degrees of freedommandn,it
follows that
1 −α=P
{
χm^2 /m
χn^2 /n
≥F 1 −α,m,n
}
implying, from Equation 5.8.3, that
1
Fα,n,m
=F 1 −α,m,n
Thus, for instance,F.9,5,7=1/F.1,7,5=1/3.37=.2967 where the value ofF.1,7,5was
obtained from Table A4 of the Appendix.
Program 5.8.3 computes the distribution function ofFn,m.
EXAMPLE 5.8f DetermineP{F6,14≤1. 5}.
SOLUTION Run Program 5.8.3 to obtain the solution .7518. ■
*5.9The Logistics Distribution
A random variableXis said to have alogisticsdistribution with parametersμandv>0if
its distribution function is
F(x)=
e(x−μ)/v
1 +e(x−μ)/v
, −∞<x<∞
* Optional section.