Global Positioning System Reference
In-Depth Information
N
n
−
r
w
2
,
0
T
−
1
2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
z
H
a
∼
σ
(7.198)
See Equation (4.272) for the corresponding expression for the zero hypothesis. The
matrix
T
has the same meaning as in Section 4.9.4, i.e.,
=
A
2
N
−
1
A
2
−
1
T
(7.199)
1
The next step is to diagonalize the covariance matrix of
Z
H
a
and to compute the sum
of the squares of the transformed random variables. These newly formed random
variables have a unit variate normal distribution with a nonzero mean. According to
Section A.5.2, the sum of the squares has a noncentral chi-square distribution. Thus,
z
H
a
Tz
H
a
σ
v
T
Pv
σ
∆
2
n
2
,
=
∼ χ
(7.200)
[29
λ
2
0
2
0
where the noncentrality parameter is
Lin
—
7
——
No
*PgE
w
2
Tw
2
σ
λ =
(7.201)
2
0
The reader is referred to the statistical literature, such as Koch (1988), for additional
details on noncentral distributions and their respective derivations. Finally, the ratio
v
T
Pv
v
T
Pv
∗
∆
n
1
−
r
∼
F
n
2
,n
1
−
r,
λ
(7.202)
[29
n
2
has a noncentral
F
distribution with noncentrality
λ
. If the test statistic computed
under the specifications of
H
0
fulfills
F
≤
F
n
2
,n
1
−
r,
α
, then
H
0
is accepted with a
type-I error of
α
. The alternative hypothesis
H
a
can be separated from
H
0
with the
power 1
− β
(
α
,
λ
)
. The type-II error is
F
n
2
,n
1
−
r,
1
−
α
β
(
α
,
λ
)
=
F
n
2
,n
1
−
r,
λ
dx
(7.203)
0
Th
e integration is taken over the noncentral
F
-distribution function from zero to the
va
lue
F
n
2
,n
1
−
r,
α
, which is specified by the significance level
.
Because the noncentrality is different for each alternative hypothesis according to
(7
.201), the type-II error
α
)
also varies with
H
a
. Rather than using the individual
ty
pe-II errors to make decisions, Euler and Schaffrin (1990) propose using the ratio of
no
ncentrality parameters. They designate the float solution as the common alternative
hypothesis
H
a
, for all null hypotheses. In this case, the value
w
2
in (7.195) is
w
2
=−
A
2
x
∗
+
2
β
(
α
,
λ
(7.204)
and the noncentrality parameter becomes