STA 6934 - Spatial Statistics: HW 1
HW 1 - Problem No. 1
a)
Both the contour and perspective plots show a trend toward higher values in
the northeast, or upper right, of LA. Thee are some lower ”peaks” towards the
southeast.
b)
Your eye estimates should be similar with your answer in c).
c)
We choose spherical model because it relates more points than the exponential
one. The still is (0.0253+0.1546=0.1799), range is 121.5, nugget is 0.0253 from R
output.
(see ozone.pdf)
HW 1 - Problem No. 2
The Still = Co + (Cr*Ar)
The Nugget = Co regardless of Cr or Ar
The Range = Ar/2
(see Routput-hw1.2.pdf)
1
This
preview
has intentionally blurred sections.
Sign up to view the full version.
HW 1 - Problem No. 3
Cov
(
Z
(
S
i
, Z
(
S
j
)) =
C
(
S
i
-
Sj
)
V ar
(
Z
) =
V ar
(
N
i
=1
Z
(
S
i
)
/N
)
=
1
/N
2
[
N
i
=1
V ar
(
Z
(
S
i
)) +
i
=
j
Cov
(
Z
(
S
i
, Z
(
S
j
))]
=
1
/N
2
[
N
i
=1
C
(0) +
i
=
j
C
(0)
ρ
(
S
i
-
Sj
)]
=
C
(0)(1
/N
+ 1
/N
2
i
=
j
ρ
(
S
i
-
Sj
))
If the data are independent,
V ar
(
Z
) = 1
/N
2
∑
N
i
=1
V ar
(
Z
(
S
i
)) = 1
/NC
(0)
Since for spatial data, observations closer together tend to be more alike than
observations farther apart,
ρ
(
S
i
-
Sj
) decreases as
|
S
i
-
Sj
|
becomes larger. So we
can obtain the data from the region as scattered as possible.

This is the end of the preview.
Sign up
to
access the rest of the document.
- Fall '08
- YOUNG
- Statistics, Trigraph, Kriging, sj
-
Click to edit the document details