TELK
O
M
N
I
K
A
Te
l
e
c
ommu
n
i
c
at
i
on
,
C
omp
u
ti
n
g
,
El
e
c
t
r
on
i
c
s
an
d
C
on
tr
o
l
V
ol
.
18
,
N
o
.
2
,
A
pri
l
2020
,
p
p.
7
7
6
~
7
8
2
IS
S
N
:
1693
-
6930
,
a
c
c
r
e
di
t
e
d
F
i
rs
t
G
r
a
de
by
K
e
m
e
n
ri
s
t
e
kdi
k
t
i
,
D
e
c
r
e
e
N
o:
21
/
E
/
K
P
T
/
2
018
D
O
I:
10
.
129
28/
T
E
L
K
O
M
N
IK
A
.
v1
8i
2
.
1482
6
776
Jou
r
n
al
h
o
m
e
pa
ge
:
ht
t
p:
/
/
j
our
nal
.
uad
.
a
c
.
i
d/
i
nd
e
x
.
php
/
T
E
L
K
O
MNIKA
PSO
o
p
t
i
m
i
z
a
t
i
o
n
o
n
b
a
c
k
p
r
o
p
a
g
a
t
i
o
n
f
o
r
f
i
sh
c
a
t
c
h
p
r
o
d
u
c
t
i
o
n
p
r
e
d
i
c
t
i
o
n
Y
u
s
l
e
n
a
S
a
r
i
,
Ek
a
S
e
t
ya
Wi
ja
ya,
A
n
d
r
e
yan
R
i
z
k
y
Bas
k
ar
a,
R
i
c
o
S
i
l
as
D
w
i
K
as
an
d
a
F
a
c
u
l
t
y
o
f
E
ng
i
ne
e
r
i
ng
,
U
ni
ve
r
s
i
t
a
s
L
a
m
bun
g
M
a
ngku
r
a
t
,
I
ndo
ne
s
i
a
A
r
ti
c
l
e
I
n
fo
A
BS
TR
A
C
T
Ar
t
i
c
l
e
h
i
s
t
or
y
:
Re
c
e
i
v
e
d
J
ul
20
,
20
19
Re
vi
s
e
d
J
a
n
16
,
202
0
A
c
c
e
pt
e
d
Fe
b
4
,
2020
G
l
oba
l
c
l
i
m
a
t
e
c
ha
nge
i
s
a
n
i
s
s
ue
t
h
a
t
i
s
e
noug
h
t
o
g
r
a
b
t
he
a
t
t
e
nt
i
on
of
t
h
e
w
o
r
l
d
c
o
m
m
un
i
t
y.
T
h
i
s
i
s
m
a
i
nl
y
b
e
c
a
us
e
o
f
t
h
e
i
m
pa
c
t
i
t
ha
s
on
h
um
a
n
l
i
f
e
.
T
he
i
m
pa
c
t
t
ha
t
i
s
f
e
l
t
a
l
s
o
oc
c
u
r
s
i
n
w
a
t
e
r
s
on
t
he
S
o
ut
h
K
a
l
i
m
a
n
t
a
n
r
e
g
i
on
.
T
hi
s
i
s
o
f
c
our
s
e
c
a
n
di
s
r
u
pt
t
he
pr
o
duc
t
i
vi
t
y
of
f
i
s
h
i
n
t
he
m
a
r
i
ne
w
a
t
e
r
s
of
S
ou
t
h
K
a
l
i
m
a
nt
a
n
.
T
h
i
s
s
t
ud
y
a
i
m
s
t
o
m
a
k
e
f
i
s
h
c
a
t
c
h
p
r
odu
c
t
i
on
pr
e
d
i
c
t
i
on
m
ode
l
s
ba
s
e
d
on
c
l
i
m
a
t
e
c
h
a
ng
e
i
n
t
h
e
S
out
h
K
a
l
i
m
a
n
t
a
n
P
r
o
vi
n
c
e
be
c
a
us
e
t
he
a
m
ou
nt
o
f
p
r
od
uc
t
i
vi
t
y
of
m
a
r
i
n
e
f
i
s
h
ha
s
f
l
uc
t
ua
t
e
d
.
T
hi
s
s
t
ud
y
us
e
s
c
l
i
m
a
t
e
da
t
a
a
s
i
n
pu
t
a
nd
f
i
s
h
pr
oduc
t
i
on
a
s
ou
t
pu
t
w
hi
c
h
i
s
di
vi
de
d
i
n
t
o
t
w
o,
n
a
m
e
l
y
t
r
a
i
n
i
ng
a
n
d
t
e
s
t
i
n
g
da
t
a
.
T
he
n
t
he
p
r
e
di
c
t
i
on
i
s
c
o
ndu
c
t
e
d
us
i
ng
B
a
c
kp
r
op
a
ga
t
i
on
m
e
t
hod
c
o
m
b
i
n
e
d
w
i
t
h
P
a
r
t
i
c
l
e
S
w
a
r
m
O
p
t
i
m
i
z
a
t
i
on
m
e
t
ho
d.
T
he
r
e
s
ul
t
s
of
t
he
s
t
ud
y
p
r
od
uc
e
d
a
p
r
e
d
i
c
t
i
on
m
ode
l
w
i
t
h
R
M
S
E
of
0
.
09
09
w
i
t
h
a
c
o
m
b
i
na
t
i
on
o
f
p
a
r
a
m
e
t
e
r
s
u
s
e
d
,
n
a
m
e
l
y,
C
1:
2,
C
2
:
2,
w
:
0
.
7
,
l
e
a
r
n
i
ng
r
a
t
e
:
0
.
5,
M
o
m
e
nt
u
m
:
0
.
1
,
P
a
r
t
i
c
l
e
s
:
5,
a
nd
e
po
c
h:
5
00.
W
h
i
l
e
t
he
m
od
e
l
us
e
d
w
he
n
p
r
e
d
i
c
t
i
ng
t
e
s
t
i
n
g
da
t
a
pr
oduc
e
s
R
M
S
E
of
0
.
1
448
.
Ke
y
w
or
d
s
:
Ba
c
k
prop
a
ga
t
i
on
Cl
i
m
a
t
e
c
h
a
ng
e
F
i
s
h
prod
uc
t
i
on
pr
e
di
c
t
i
on
P
a
rt
i
c
l
e
s
w
a
rm
op
t
i
m
i
z
a
t
i
on
RM
S
E
T
hi
s
i
s
an
op
e
n
ac
c
e
s
s
ar
t
i
c
l
e
u
nde
r
t
he
C
C
B
Y
-
SA
l
i
c
e
n
s
e
.
Cor
r
e
s
pon
di
n
g
Au
t
h
or
:
A
ndre
y
a
n
Ri
z
ky
Ba
s
k
a
ra
,
F
a
c
ul
t
y
of
E
ng
i
n
e
e
r
i
ng
,
U
ni
ve
rs
i
t
a
s
L
a
m
bung
M
a
n
gkur
a
t
,
Bri
gj
e
n
H
.
H
a
s
a
n
B
a
s
ri
S
t
.
,
P
a
n
ge
r
a
n,
B
a
nj
a
r
m
a
s
i
n
U
t
a
r
a
,
Ba
nj
a
r
m
a
s
i
n,
K
a
l
i
m
a
n
t
a
n
S
e
l
a
t
a
n
,
Indon
e
s
i
a
.
E
m
a
i
l
:
a
ndr
e
y
a
n.
ba
s
k
a
ra
@
ul
m
.
a
c
.
i
d
1.
I
N
TR
O
D
U
C
TI
O
N
F
i
s
h
produ
c
t
i
on
i
n
S
out
h
K
a
l
i
m
a
n
t
a
n
a
re
produ
c
e
s
fro
m
va
ri
o
us
a
c
t
i
vi
t
i
e
s
s
uc
h
a
s
i
n
t
h
e
f
orm
of
c
u
l
t
i
va
t
i
on
a
nd
c
a
t
c
h
.
T
he
a
m
oun
t
of
fi
s
h
fro
m
m
a
r
i
ne
c
a
t
c
h
i
n
2016
di
v
i
de
d
i
nt
o
qua
r
t
e
r
y
e
a
r
.
I
t
r
e
a
c
he
d
34,
161
t
ons
i
n
t
h
e
f
i
rs
t
qu
a
rt
e
r
,
47
,
260
t
ons
i
n
t
h
e
s
e
c
ond
q
u
a
rt
e
r,
36,
7
85
t
o
ns
i
n
t
he
t
hi
rd
qu
a
rt
e
r,
a
nd
58,
299
t
ons
i
n
t
h
e
f
our
t
h
qua
rt
e
r
[1
–
3
]
.
T
h
e
re
i
s
a
s
i
gn
i
fi
c
a
n
t
de
c
re
a
s
e
a
nd
i
n
c
re
a
s
e
i
n
t
h
e
a
m
ount
of
f
i
s
h
produ
c
t
i
on
.
T
hus
,
i
t
i
s
d
i
ffi
c
ul
t
t
o
p
rovi
de
po
l
i
c
i
e
s
r
e
l
a
t
e
d
t
o
fi
s
he
ri
e
s
pro
duc
t
i
on
w
i
t
h
s
u
c
h
be
ha
v
i
or
.
T
h
e
re
a
re
a
l
s
o
probl
e
m
s
re
l
a
t
e
d
t
o
t
he
m
a
rk
e
t
i
ng
of
f
i
s
h
i
t
s
e
l
f
i
n
ord
e
r
t
o
i
n
c
re
a
s
e
r
e
gi
o
na
l
i
nc
o
m
e
fr
om
t
h
e
provi
nc
e
i
t
s
e
l
f.
T
he
r
e
a
r
e
a
nu
m
b
e
r
of
p
robl
e
m
s
re
l
a
t
e
d
t
o
f
i
s
h
pr
oduc
t
i
o
n
i
n
S
out
h
K
a
l
i
m
a
n
t
a
n
w
h
i
c
h
ha
ve
re
s
u
l
t
e
d
i
n
a
de
c
l
i
ne
i
n
l
o
c
a
l
f
i
s
h
produ
c
t
i
on
a
s
s
t
a
t
e
d
by
t
h
e
s
e
c
r
e
t
a
r
y
of
t
he
S
out
h
K
a
l
i
m
a
nt
a
n
M
a
r
i
n
e
a
nd
F
i
s
h
e
ri
e
s
A
ge
nc
y
(D
K
P
)
t
ha
t
s
e
a
fi
s
he
ri
e
s
pr
oduc
t
i
on
i
n
S
ou
t
h
K
a
l
i
m
a
n
t
a
n
h
a
s
fl
u
c
t
ua
t
e
d
fro
m
y
e
a
r
t
o
ye
a
r
du
e
t
o
fi
s
he
r
m
e
n
f
rom
o
t
he
r
re
g
i
ons
i
l
l
e
ga
l
fi
s
hi
ng
i
n
c
l
ud
i
ng
ov
e
rf
i
s
hi
ng
or
no
t
ha
v
i
ng
a
pe
r
m
i
t
i
n
t
he
w
a
t
e
rs
of
S
ou
t
h
K
a
l
i
m
a
nt
a
n.
T
he
r
e
i
s
a
l
s
o
a
n
ot
h
e
r
i
s
s
ue
re
l
a
t
e
d
t
o
c
l
i
m
a
t
e
,
n
a
m
e
l
y
gl
ob
a
l
w
a
r
m
i
ng.
T
hi
s
of
c
ours
e
a
l
s
o
a
ffe
c
t
s
t
he
c
o
ndi
t
i
on
of
m
a
ri
ne
fi
s
h
e
ri
e
s
t
ha
t
o
c
c
ur
i
n
S
out
h
K
a
l
i
m
a
n
t
a
n
.
In
a
s
t
udy
c
ond
uc
t
e
d
b
y
M
uha
m
m
a
d
A
z
ha
r
Ra
z
a
k
a
nd
E
dw
i
n
R
i
ks
a
k
om
a
ra
,
a
bout
fore
c
a
s
t
i
ng
t
he
a
m
oun
t
of
f
i
s
h
produ
c
t
i
o
n
i
n
B
a
nj
a
rm
a
s
i
n
us
i
ng
t
he
b
a
c
kprog
a
t
i
on
m
e
t
h
od.
T
he
re
s
u
l
t
s
gi
v
e
n
i
n
us
i
ng
t
he
m
e
t
hod
s
uc
c
e
e
de
d
i
n
pro
duc
i
ng
a
roun
d
20
%
e
rrors
i
n
t
he
t
e
s
t
i
ng
p
roc
e
s
s
w
i
t
h
t
h
e
s
up
port
i
ng
va
r
i
a
b
l
e
s
na
m
e
l
y
a
i
r
t
e
m
pe
ra
t
ure
[4]
.
O
t
he
r
s
t
ud
i
e
s
b
y
P
ut
ri
a
nd
A
g
us
[5]
i
n
t
he
pre
di
c
t
i
on
of
c
a
t
f
i
s
h
pr
odu
c
t
i
on
i
n
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
K
O
M
N
IK
A
T
e
l
e
c
om
m
un
Co
m
put
E
l
Con
t
rol
P
SO
o
pt
i
m
i
z
at
i
on
on
bac
k
pr
opag
at
i
on
f
or
f
i
s
h
c
a
t
c
h
pr
od
uc
t
i
on
pr
e
di
c
t
i
on
(
Y
us
l
e
na
Sar
i
)
777
S
l
e
m
a
n
D
i
s
t
ri
c
t
w
i
t
h
t
h
e
n
e
uro
fu
z
z
y
m
od
e
l
c
a
n
c
on
c
l
u
de
t
ha
t
t
he
r
e
s
ul
t
s
o
bt
a
i
n
e
d
by
m
e
a
s
uri
n
g
M
A
P
E
a
r
e
qui
t
e
go
od
a
t
20%
i
n
t
r
a
i
n
i
ng
da
t
a
a
nd
d
a
t
a
t
e
s
t
i
ng.
M
od
e
l
obt
a
i
ne
d
i
s
t
h
e
s
e
c
ond
m
ode
l
w
i
t
h
i
npu
t
m
od
e
l
s
ba
s
e
d
o
n
c
a
t
f
i
s
h
pro
duc
t
i
on
da
t
a
f
our
m
ont
hs
e
a
rl
i
e
r
.
O
t
h
e
r
re
s
e
a
rc
h
a
l
s
o
m
e
nt
i
ons
t
h
a
t
t
h
e
ba
c
kpr
opa
g
a
t
i
on
m
e
t
hod
a
l
on
g
w
i
t
h
opt
i
m
i
z
a
t
i
on
w
i
t
h
P
S
O
i
n
t
he
s
t
udy
of
ve
ge
t
a
bl
e
pri
c
e
pre
d
i
c
t
i
o
ns
c
ond
uc
t
e
d
b
y
Y
E
L
u,
e
t
a
l
.
obt
a
i
n
e
d
a
good
r
e
s
ul
t
.
I
t
c
a
n
be
s
e
e
n
t
he
r
e
s
ul
t
s
of
t
h
i
s
s
t
udy
s
t
a
t
i
ng
t
ha
t
t
h
e
a
c
c
ura
c
y
i
n
t
he
f
orm
o
f
M
S
E
gi
ve
n
t
hroug
h
r
e
s
e
a
rc
h
be
t
w
e
e
n
p
re
d
i
c
t
i
ons
us
i
ng
onl
y
b
a
c
k
p
ropa
g
a
t
i
on
a
nd
b
a
c
kprop
a
ga
t
i
o
n
a
l
o
ng
w
i
t
h
P
S
O
re
s
ul
t
e
d
i
n
a
v
a
l
u
e
of
0
.
0029
a
nd
0
.
001
0.
I
t
c
a
n
be
s
e
e
n
fr
om
t
h
e
re
s
e
a
r
c
h
t
h
a
t
P
S
O
c
a
n
be
qui
t
e
goo
d
a
t
re
du
c
i
ng
t
he
e
x
i
s
t
i
ng
M
S
E
s
o
t
h
a
t
b
e
t
t
e
r
pre
di
c
t
i
on
re
s
ul
t
s
a
re
ob
t
a
i
ne
d
[6]
.
Ba
s
e
d
o
n
w
ha
t
ha
s
b
e
e
n
de
s
c
ri
b
e
d
a
bove
,
t
he
r
e
s
e
a
rc
h
c
on
duc
t
e
d
i
n
t
h
i
s
pa
p
e
r
i
s
a
i
m
i
n
g
t
o
pre
d
i
c
t
fi
s
h
c
a
t
c
h
pr
o
duc
t
i
on
a
ga
i
ns
t
c
l
i
m
a
t
e
c
ha
nge
us
i
ng
t
h
e
Ba
c
kpro
pa
ga
t
i
on
m
e
t
h
od
i
n
S
ou
t
h
K
a
l
i
m
a
nt
a
n
pro
vi
n
c
e
,
but
by
ha
v
i
ng
s
u
pport
i
ng
va
r
i
a
b
l
e
s
or
m
ore
i
np
ut
s
s
uc
h
a
s
ra
i
nf
a
l
l
a
nd
w
i
nd
s
pe
e
d
c
oup
l
e
d
w
i
t
h
P
a
r
t
i
c
l
e
S
w
a
r
m
opt
i
m
i
z
a
t
i
on
.
I
t
i
s
be
c
a
us
e
t
he
m
e
t
hod
a
ppl
i
e
d
us
i
n
g
b
a
c
kpr
opa
g
a
t
i
on
a
n
d
opt
i
m
i
z
e
d
w
i
t
h
P
S
O
c
a
n
prov
i
de
t
he
be
s
t
pr
e
d
i
c
t
i
v
e
r
e
s
ul
t
s
b
a
s
e
d
on
t
h
e
e
xi
s
t
i
ng
r
e
s
e
a
rc
h
.
2.
LI
TER
A
TU
R
E
R
EV
I
EW
2.
1
.
F
i
s
h
p
r
od
u
c
ti
on
A
s
c
l
i
m
a
t
e
c
on
di
t
i
ons
c
h
a
ng
e
i
n
t
he
w
orl
d,
t
he
re
s
u
l
t
i
ng
fi
s
h
pro
duc
t
i
on
e
xp
e
ri
e
nc
e
s
a
ri
s
e
a
nd
fa
l
l
due
t
o
t
hi
s
phe
nom
e
non
.
In
S
out
h
K
a
l
i
m
a
nt
a
n
i
t
s
e
l
f,
c
l
i
m
a
t
e
c
h
a
ng
e
a
ff
e
c
t
s
t
he
c
ond
i
t
i
on
of
e
x
i
s
t
i
ng
fi
s
h
produc
t
i
o
n.
M
a
ny
s
t
udi
e
s
h
a
ve
s
ugg
e
s
t
e
d
t
h
e
e
ffe
c
t
of
f
i
s
h
pr
oduc
t
i
on
on
c
l
i
m
a
t
e
,
na
m
e
l
y
t
h
e
di
re
c
t
i
nt
e
ra
c
t
i
on
of
c
l
i
m
a
t
e
for
fi
s
h
p
rodu
c
t
i
on
t
h
a
t
o
c
c
urs
[7
–
12]
.
M
a
r
i
ne
fi
s
h
pr
oduc
t
i
o
n
i
s
i
nf
l
ue
nc
e
d
by
c
l
i
m
a
t
e
e
l
e
m
e
nt
s
i
n
t
he
for
m
o
f
r
a
i
n
fa
l
l
,
w
i
nd
s
p
e
e
d
,
a
nd
a
i
r
t
e
m
p
e
r
a
t
ur
e
.
2.
2
.
C
l
i
mat
e
c
h
an
ge
W
i
t
h
t
h
e
i
m
p
a
c
t
of
c
l
i
m
a
t
e
c
h
a
ng
e
,
t
he
l
e
ve
l
s
o
f
CO
2
i
n
t
h
e
a
i
r
l
a
y
e
r
c
a
n
a
l
s
o
e
xp
e
ri
e
nc
e
a
n
i
n
c
re
a
s
e
w
hi
c
h
d
i
r
e
c
t
l
y
i
nc
r
e
a
s
e
s
t
h
e
t
e
m
p
e
r
a
t
ur
e
of
t
he
e
a
rt
h
i
nc
l
ud
i
ng
a
qu
a
t
i
c
c
o
m
pon
e
nt
s
,
s
uc
h
a
s
ri
v
e
rs
,
l
a
ke
s
a
nd
t
he
s
e
a
.
Cl
i
m
a
t
e
c
ha
nge
i
s
c
h
a
ra
c
t
e
ri
z
e
d
by
s
e
v
e
r
a
l
phe
nom
e
na
,
s
uc
h
a
s
c
h
a
ng
e
s
i
n
a
ve
ra
g
e
or
m
e
d
i
a
n
v
a
l
u
e
s
a
nd
v
a
ri
a
t
i
ons
i
n
c
l
i
m
a
t
e
e
l
e
m
e
n
t
s
[13
–
1
5]
.
T
e
m
p
e
ra
t
ur
e
i
n
c
re
a
s
e
s
i
n
t
h
e
l
ong
run
a
nd
t
e
nds
t
o
i
nc
r
e
a
s
e
ov
e
r
t
i
m
e
.
In
a
ddi
t
i
on
,
c
h
a
nge
s
i
n
ra
i
nf
a
l
l
p
a
t
t
e
rns
t
h
a
t
a
r
e
m
a
rke
d
by
t
he
l
a
t
e
s
t
a
r
t
of
t
he
w
e
t
s
e
a
s
on
a
nd
t
h
e
e
nd
of
t
he
ra
i
ny
s
e
a
s
on
a
r
e
f
a
s
t
e
r.
W
he
re
t
h
e
ra
i
ny
s
e
a
s
o
n
i
s
s
hor
t
e
r
b
ut
w
i
t
h
a
h
i
gh
i
n
t
e
ns
i
t
y
of
r
a
i
nf
a
l
l
,
c
l
i
m
a
t
e
c
ha
n
ge
c
a
n
be
s
a
i
d
[
16
–
18]
.
T
h
e
oc
c
urre
nc
e
of
t
he
phe
nom
e
non
of
c
l
i
m
a
t
e
c
ha
ng
e
i
n
Indon
e
s
i
a
,
e
s
p
e
c
i
a
l
l
y
i
n
S
out
h
K
a
l
i
m
a
nt
a
n
c
a
n
be
o
bs
e
rv
e
d
fro
m
t
h
e
c
h
a
ng
e
s
i
n
t
h
e
a
v
e
ra
g
e
l
o
ng
-
t
e
r
m
ra
i
nfa
l
l
i
n
t
h
e
re
g
i
on
a
s
s
how
n
i
n
F
i
gu
re
1.
F
i
gure
1
.
G
r
a
ph
of
nor
m
a
l
r
a
i
nf
a
l
l
c
h
a
nge
s
A
ve
r
ag
e
R
ai
n
fa
l
l
p
e
r
D
e
c
ad
e
of
S
yams
u
d
i
n
N
oor
M
e
t
e
o
r
ol
og
i
c
al
S
tati
on
Ban
ja
r
mas
i
n
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
1693
-
6930
T
E
L
K
O
M
N
IK
A
T
e
l
e
c
om
m
un
Co
m
put
E
l
Con
t
rol
,
V
ol
.
18
,
N
o.
2
,
A
pri
l
2
020:
7
7
6
-
7
8
2
778
2.
3
.
N
or
mal
i
z
ati
on
D
a
t
a
a
t
t
ri
bu
t
e
v
a
l
u
e
s
t
ha
t
va
ry
i
n
ra
ng
e
oft
e
n
ne
e
d
t
o
b
e
nor
m
a
l
i
z
e
d
or
s
t
a
nda
rdi
z
e
d
s
o
t
h
a
t
t
h
e
da
t
a
m
i
n
i
ng
pro
c
e
s
s
i
s
no
t
bi
a
s
e
d
.
N
or
m
a
l
l
y
d
a
t
a
nor
m
a
l
i
z
a
t
i
o
n
i
s
c
a
rr
i
e
d
out
i
n
s
m
a
l
l
ra
ng
e
s
,
s
u
c
h
a
s
0
-
1
or
1
-
(
-
1),
s
o
t
h
a
t
a
l
l
a
t
t
ri
b
ut
e
s
w
i
l
l
ha
v
e
t
he
s
a
m
e
w
e
i
ght
.
N
or
m
a
l
i
z
a
t
i
on
t
e
c
hni
que
s
a
re
v
e
ry
i
m
p
ort
a
nt
i
n
d
a
t
a
m
i
n
i
ng
,
e
s
p
e
c
i
a
l
l
y
c
l
a
s
s
i
f
i
c
a
t
i
on
a
nd
c
l
us
t
e
ri
ng
[19
,
20]
.
O
f
t
h
e
m
a
ny
s
t
r
a
t
e
gi
e
s
us
e
d
i
n
no
rm
a
l
i
z
a
t
i
on,
t
h
i
s
s
t
udy
us
e
d
t
he
no
rm
a
l
i
z
a
t
i
on
m
e
t
ho
d
c
a
l
l
e
d
m
i
n
-
m
a
x.
A
s
t
h
e
na
m
e
i
m
p
l
i
e
s
,
t
h
i
s
m
e
t
ho
d
us
e
s
m
i
ni
m
u
m
a
nd
m
a
x
i
m
u
m
v
a
l
ue
s
t
o
c
onv
e
rt
da
t
a
l
i
n
e
a
r
l
y
.
It
c
a
n
b
e
s
e
e
n
t
hro
ugh
c
a
l
c
ul
a
t
i
ons
or
t
h
e
fo
l
l
ow
i
ng
f
orm
u
l
a
.
A
f
t
e
r
t
ha
t
t
he
r
e
i
s
a
f
orm
ul
a
for
no
rm
a
l
i
z
i
ng
norm
a
l
i
z
e
d
da
t
a
[21
–
2
3]
:
m
a
x
()
X
'
'
'
'
(
X
)
m
in
m
a
x
m
in
m
in
m
in
XX
X
X
X
X
−
=
−
+
−
(1)
w
he
re
,
X
’
=
no
rm
a
l
i
z
a
t
i
on
r
e
s
ul
t
X
=
d
a
t
a
t
ha
t
w
i
l
l
b
e
nor
m
a
l
i
z
e
d
X
max
=
m
á
x
i
m
um
va
l
u
e
of
ov
e
ra
l
l
d
a
t
a
X
m
i
n
=
m
i
n
i
m
u
m
va
l
ue
of
ov
e
ra
l
l
d
a
t
a
X’
m
a
x
=
ne
w
m
a
xi
m
u
m
v
a
l
ue
X’
m
i
n
=
n
e
w
m
i
n
i
m
um
v
a
l
u
e
=
(
′
×
(
−
))
+
(2)
w
he
re
,
X
=
d
e
nor
m
a
l
i
z
a
t
i
on
r
e
s
ul
t
X
’
=
d
a
t
a
t
ha
t
w
i
l
l
b
e
de
no
rm
a
l
i
z
e
d
X
max
=
m
a
x
i
m
um
va
l
u
e
of
ov
e
ra
l
l
d
a
t
a
X
m
i
n
=
m
i
n
i
m
u
m
va
l
ue
of
ov
e
ra
l
l
d
a
t
a
2.
4
.
R
oot
m
e
an
s
q
u
ar
e
e
r
r
o
r
(R
S
M
E)
Root
M
e
a
n
S
qu
a
re
E
rro
r
(RM
S
E
)
i
s
a
m
e
a
s
ure
t
o
c
a
l
c
ul
a
t
e
t
he
m
a
g
ni
t
ude
o
f
e
rro
rs
i
n
pr
e
di
c
t
i
ons
,
RM
S
E
h
a
s
be
e
n
us
e
d
a
s
a
s
t
a
nd
a
rd
s
t
a
t
i
s
t
i
c
a
l
m
e
t
ri
c
f
or
m
e
a
s
ur
i
ng
m
o
de
l
p
e
rfor
m
a
nc
e
i
n
m
e
t
e
oro
l
ogy
,
a
i
r
qua
l
i
t
y
,
a
nd
c
l
i
m
a
t
e
r
e
s
e
a
r
c
h
s
t
ud
i
e
s
[24
,
2
5]
:
RM
S
E
=
1
2
n
=
ˆ
()
Y
i
Y
i
i
n
−
(3)
w
he
re
,
Yi
=
a
c
t
ua
l
d
a
t
a
Ŷ
i
=
f
i
na
l
da
t
a
(e
s
t
i
m
a
t
e
d
da
t
a
)
N
=
a
m
o
unt
of
da
t
a
2.
5
.
Bac
k
p
r
op
ag
ati
on
Ba
c
k
prop
a
ga
t
i
on
i
s
a
n
a
l
gor
i
t
h
m
o
n
a
rt
i
fi
c
i
a
l
n
e
ur
a
l
n
e
t
w
o
rks
t
ha
t
i
s
oft
e
n
us
e
d
i
n
s
e
a
rc
hi
ng
for
opt
i
m
a
l
w
e
i
gh
t
s
.
In
t
he
b
a
c
kp
ropa
ga
t
i
on
ne
t
w
ork
t
h
e
re
a
r
e
t
h
e
d
e
s
i
r
e
d
i
nput
p
a
t
t
e
rns
a
nd
ou
t
put
pa
t
t
e
rns
.
W
he
n
a
n
e
t
w
or
k
i
s
gi
v
e
n
a
pa
t
t
e
rn
,
t
h
e
va
l
ue
s
o
f
w
e
i
ght
s
a
re
c
ha
n
ge
d
i
n
ord
e
r
t
o
m
i
ni
m
i
z
e
t
he
di
ff
e
re
nc
e
be
t
w
e
e
n
t
he
out
put
p
a
t
t
e
rn
of
t
he
ne
t
w
ork
a
nd
t
h
e
de
s
i
r
e
d
out
put
p
a
t
t
e
rn
.
N
e
t
w
o
rk
t
r
a
i
n
i
ng
i
s
c
a
rri
e
d
out
re
pe
a
t
e
dl
y
un
t
i
l
a
l
l
o
ut
pu
t
pa
t
t
e
rns
fr
om
t
h
e
ne
t
w
ork
c
a
n
r
e
c
o
gni
z
e
t
he
d
e
s
i
r
e
d
out
pu
t
pa
t
t
e
rn
[6,
21
,
26]
.
T
he
t
e
rm
i
n
a
t
i
on
c
ond
i
t
i
on
us
e
d
i
n
t
hi
s
a
l
gori
t
hm
i
s
t
h
e
m
a
xi
m
u
m
num
be
r
of
i
t
e
r
a
t
i
ons
a
nd
e
r
ror
t
a
rg
e
t
s
.
T
r
a
i
ni
ng
w
i
l
l
b
e
s
t
op
pe
d
i
f
t
he
nu
m
be
r
of
t
ra
i
ni
n
g
i
t
e
r
a
t
i
ons
e
x
c
e
e
ds
t
he
m
a
xi
m
u
m
nu
m
be
r
of
i
t
e
ra
t
i
ons
or
i
f
t
he
e
rr
or
obt
a
i
n
e
d
i
n
t
h
e
t
ra
i
ni
ng
i
s
s
m
a
l
l
e
r
t
ha
n
t
he
t
a
rg
e
t
e
rro
r.
2.
6
.
P
ar
t
i
c
l
e
s
w
ar
n
op
ti
m
i
z
ati
on
P
S
O
i
s
a
popul
a
t
i
on
-
b
a
s
e
d
opt
i
m
i
z
a
t
i
on
t
e
c
hni
qu
e
d
e
ve
l
op
e
d
by
E
b
e
rha
rt
a
nd
K
e
n
ne
dy
i
n
19
95,
w
h
i
c
h
w
a
s
i
ns
p
i
re
d
by
t
he
s
oc
i
a
l
b
e
ha
v
i
or
of
bi
rds
or
fi
s
h
fl
o
c
ks
[2
2,
2
3]
.
P
S
O
c
a
n
be
a
s
s
u
m
e
d
a
s
a
gro
up
o
f
b
i
rds
l
ooki
ng
f
or
food
i
n
a
n
a
re
a
.
B
i
rds
t
h
a
t
a
r
e
l
ook
i
ng
for
food
d
o
no
t
k
now
e
xa
c
t
l
y
w
he
r
e
t
h
e
f
ood
i
s
l
oc
a
t
e
d
,
s
o
w
ha
t
t
h
e
bi
rd
do
e
s
i
s
l
o
ok
for
t
h
e
b
e
s
t
s
t
ra
t
e
gy
t
o
fi
nd
t
he
s
e
foo
ds
b
y
fo
l
l
ow
i
ng
t
h
e
c
l
os
e
s
t
bi
rds
t
o
t
h
e
s
e
foods
[6]
.
F
ol
l
ow
i
ng
a
re
t
h
e
e
qu
a
t
i
ons
i
n
t
h
e
P
S
O
f
orm
u
l
a
:
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
K
O
M
N
IK
A
T
e
l
e
c
om
m
un
Co
m
put
E
l
Con
t
rol
P
SO
o
pt
i
m
i
z
at
i
on
on
bac
k
pr
opag
at
i
on
f
or
f
i
s
h
c
a
t
c
h
pr
od
uc
t
i
on
pr
e
di
c
t
i
on
(
Y
us
l
e
na
Sar
i
)
779
v
i
,
d
=
w
*
v
i
,
d
+
c
1
*
R
*
(pb
e
s
t
i
,
d
-
x
i
,
d)
+
c
2
*
R
*
(gbe
s
t
d
-
x
i
,
d
)
(4)
x
i
d
=
x
i
,
d
+
v
i
,
d
(5)
W
=
In
e
rt
i
a
l
w
e
i
g
ht
i
ng
f
a
c
t
or
c
1,
c
2
=
l
e
a
rni
n
g
r
a
t
e
c
ons
t
a
n
t
R
=
r
a
ndo
m
num
be
r
(0
-
1)
X
i
,
d
=
Cu
rre
n
t
pos
i
t
i
on
of
t
h
e
i
-
t
h
p
a
rt
i
c
l
e
i
n
t
he
i
-
t
h
i
t
e
ra
t
i
o
n
P
be
s
t
i
=
P
r
e
vi
ous
b
e
s
t
pos
i
t
i
on
of
i
-
t
h
p
a
r
t
i
c
l
e
G
be
s
t
i
=
T
he
b
e
s
t
pa
r
t
i
c
l
e
a
m
o
ng
a
l
l
p
a
rt
s
i
n
a
gr
oup
or
popul
a
t
i
on
n
=
N
u
m
be
r
of
p
a
rt
i
c
l
e
s
i
n
a
grou
p
d
=
D
i
m
e
ns
i
on
3.
RE
S
EA
R
C
H
M
ET
H
O
D
O
L
O
G
Y
T
he
d
e
s
i
gn
of
t
h
e
e
x
i
s
t
i
n
g
m
od
e
l
i
s
a
n
i
l
l
us
t
ra
t
i
o
n
o
f
t
he
m
ode
l
t
h
a
t
w
i
l
l
be
m
a
d
e
a
c
c
ord
i
ng
t
o
t
he
i
ni
t
i
a
l
n
e
e
ds
b
a
s
e
d
on
da
t
a
a
na
l
ys
i
s
t
h
a
t
h
a
s
be
e
n
do
n
e
be
fo
re
.
D
a
t
a
t
o
be
us
e
d
a
re
c
l
i
m
a
t
e
da
t
a
a
nd
produc
t
i
o
n
da
t
a
e
ve
ry
3
m
ont
hs
.
T
he
d
a
t
a
c
o
l
l
e
c
t
e
d
w
i
l
l
b
e
di
v
i
de
d
i
n
t
o
t
w
o
na
m
e
l
y
t
r
a
i
n
i
ng
d
a
t
a
t
o
ge
t
t
he
m
o
de
l
a
nd
t
e
s
t
i
n
g
d
a
t
a
t
o
e
v
a
l
u
a
t
e
t
he
r
e
s
ul
t
s
of
t
h
e
m
od
e
l
.
T
he
t
r
a
i
ni
ng
d
a
t
a
w
i
l
l
b
e
n
orm
a
l
i
z
e
d
a
nd
t
h
e
n
t
ra
i
ne
d
us
i
ng
ba
c
kpr
opa
g
a
t
i
on
a
nd
P
S
O
,
a
ft
e
r
g
e
t
t
i
ng
t
h
e
b
e
s
t
m
ode
l
t
he
n
w
e
a
pp
l
y
t
he
m
ode
l
t
o
t
he
norm
a
l
i
z
e
d
t
e
s
t
i
ng
da
t
a
t
o
g
e
t
r
e
s
ul
t
s
.
T
h
e
de
s
c
ri
p
t
i
on
of
t
he
m
ode
l
de
s
i
gn
t
ha
t
w
i
l
l
be
m
a
d
e
c
a
n
b
e
s
e
e
n
t
hroug
h
t
he
fol
l
ow
i
ng
F
i
gur
e
2
.
F
i
gure
2
.
M
od
e
l
d
e
s
i
gn
i
ng
4.
R
ES
U
LT
A
N
D
A
N
A
LY
S
I
S
4.
1
.
D
ata
p
r
o
c
e
s
s
i
n
g
T
he
d
a
t
a
t
h
a
t
w
i
l
l
be
us
e
d
i
n
t
h
i
s
s
t
udy
i
s
i
n
t
h
e
for
m
of
m
on
t
hl
y
c
l
i
m
a
t
e
da
t
a
w
h
i
c
h
i
n
c
l
ud
e
s
re
l
a
t
e
d
i
ndi
c
a
t
ors
t
h
a
t
a
r
e
ne
e
de
d.
C
l
i
m
a
t
e
da
t
a
w
a
s
ob
t
a
i
ne
d
f
rom
t
he
Cl
a
s
s
I
B
a
nj
a
rb
a
ru
M
e
t
e
or
ol
ogy
a
nd
G
e
o
phys
i
c
s
A
ge
nc
y
,
w
he
re
t
he
c
l
i
m
a
t
e
d
a
t
a
obt
a
i
n
e
d
fro
m
S
ya
m
s
udi
n
N
oor
s
t
a
t
i
on
a
nd
t
e
c
hn
i
c
a
l
i
m
p
l
e
m
e
nt
i
ng
uni
t
s
i
n
a
re
a
s
c
l
os
e
t
o
t
h
e
s
e
a
.
T
h
e
c
l
i
m
a
t
e
d
a
t
a
i
s
f
rom
2008
-
2016
a
n
d
t
h
e
c
l
i
m
a
t
e
da
t
a
obt
a
i
n
e
d
i
s
10
2
d
a
t
a
.
T
h
e
n
e
xt
da
t
a
r
e
l
a
t
e
d
t
o
fi
s
h
produ
c
t
i
on
w
a
s
ob
t
a
i
n
e
d
t
hroug
h
t
he
S
out
h
K
a
l
i
m
a
nt
a
n
P
r
ovi
n
c
e
'
s
O
ff
i
c
e
o
f
M
a
ri
n
e
a
nd
F
i
s
h
e
ri
e
s
w
he
r
e
d
a
t
a
obt
a
i
ne
d
i
s
fi
s
h
produ
c
t
i
on
d
a
t
a
i
n
t
he
f
orm
of
pro
duc
t
i
on
qua
r
t
e
r
t
o
t
a
l
i
ng
36
d
a
t
a
.
T
he
fo
l
l
ow
i
ng
i
npu
t
da
t
a
i
s
s
how
n
i
n
T
a
bl
e
1
a
nd
t
he
t
a
rg
e
t
d
a
t
a
i
s
s
how
n
i
n
T
a
bl
e
2
.
T
a
b
l
e
1
.
In
put
da
t
a
us
e
d
No
Y
e
a
r
M
o
n
t
h
Ra
i
n
fa
l
l
W
i
n
d
V
e
l
o
c
i
t
y
T
e
m
p
e
ra
t
u
re
1
2008
J
a
n
u
a
ry
465
4
26
2
2008
F
e
b
ru
a
ry
81
3
27
…
..
..
..
..
..
107
2016
N
o
v
e
m
b
e
r
231
3
28
108
2016
D
e
c
e
m
b
e
r
723
3
27
Ta
b
l
e
2
.
T
a
rg
e
t
d
a
t
a
t
o
be
us
e
d
No
Y
e
a
r
F
i
s
h
P
ro
d
u
c
t
i
o
n
(Q
u
a
rt
e
rl
y
)
P
ro
d
u
c
t
i
o
n
i
n
T
o
n
1
2008
J
a
n
u
a
ry
-
M
a
rc
h
3
5
2
1
1
2
2008
A
p
ri
l
-
J
u
n
e
3
0
4
3
1
…
..
..
..
35
2016
J
u
l
y
-
S
e
p
t
e
m
b
e
r
3
6
7
8
5
36
2016
O
c
t
o
b
e
r
-
D
e
c
e
m
b
e
r
5
8
2
9
9
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
1693
-
6930
T
E
L
K
O
M
N
IK
A
T
e
l
e
c
om
m
un
Co
m
put
E
l
Con
t
rol
,
V
ol
.
18
,
N
o.
2
,
A
pri
l
2
020:
7
7
6
-
7
8
2
780
4.
2
.
D
i
s
c
u
s
s
i
on
4.
2
.
1
E
xp
e
r
i
m
e
n
t
t
o
ob
ta
i
n
b
e
s
t
m
od
e
l
In
t
h
e
Ba
c
kprop
a
g
a
t
i
on
a
nd
P
S
O
m
e
t
ho
ds
,
t
he
re
a
r
e
p
a
ra
m
e
t
e
rs
t
h
a
t
w
i
l
l
be
us
e
d
a
n
d
t
he
purpos
e
of
t
hi
s
s
t
e
p
i
s
t
o
fi
nd
t
h
e
r
i
ght
c
o
m
bi
n
a
t
i
on
of
pa
r
a
m
e
t
e
rs
t
h
a
t
w
e
c
a
n
a
l
s
o
s
e
e
t
h
e
r
e
s
ul
t
s
b
e
for
e
.
T
h
e
pa
ra
m
e
t
e
rs
c
ont
a
i
n
e
d
i
n
t
he
P
S
O
m
e
t
hod
c
1,
c
2
,
a
nd
w
h
a
v
e
be
e
n
d
e
t
e
r
m
i
n
e
d
i
n
a
d
va
n
c
e
or
d
e
fa
u
l
t
b
e
c
a
us
e
t
h
e
r
e
s
ul
t
s
i
n
t
he
w
e
i
ght
s
e
a
rc
h
i
ng
a
r
e
fo
und
t
o
be
ve
r
y
go
od.
T
h
e
va
l
u
e
s
gi
ve
n
a
r
e
c
1
=
2,
c
2
=
2,
w
=
0
.
7
,
l
e
a
rni
ng
ra
t
e
=
1
-
5
,
m
om
e
nt
um
=
0
.
1
-
0
.
5
,
nu
m
b
e
r
of
pa
r
t
i
c
l
e
s
=
5
-
7
a
nd
e
po
c
h
=
5
00.
Ba
s
e
d
on
re
s
e
a
r
c
h
c
ondu
c
t
e
d
b
y
H
a
rry
G
a
nda
N
ugra
ha
a
nd
A
z
ha
ri
i
n
t
he
re
s
e
a
rc
h
t
i
t
l
e
"
T
ra
i
ni
ng
of
A
rt
i
fi
c
i
a
l
N
e
ura
l
N
e
t
w
ork
W
e
i
gh
t
U
s
i
n
g
P
a
rt
i
c
l
e
S
w
a
rm
O
p
t
i
m
i
z
a
t
i
on
fo
r
F
or
e
c
a
s
t
i
ng
Inf
l
a
t
i
on
L
e
v
e
l
s
"
i
t
i
s
s
a
i
d
t
h
a
t
h
i
dde
n
l
a
ye
rs
c
a
n
a
ff
e
c
t
t
h
e
va
l
ue
of
t
he
ob
t
a
i
ne
d
RM
S
E
.
T
he
fo
l
l
ow
i
ng
i
s
a
n
a
t
t
e
m
pt
t
o
ge
t
t
he
r
e
s
ul
t
f
rom
d
i
ffe
re
n
t
a
r
c
hi
t
e
c
t
ur
e
t
h
a
t
h
a
s
t
he
nu
m
b
e
r
of
ne
ur
ons
3
t
o
7
i
n
t
he
h
i
dde
n
l
a
y
e
r
a
nd
t
h
e
re
s
ul
t
s
of
t
he
r
e
s
e
a
rc
h
c
a
n
b
e
s
e
e
n
b
e
l
ow
w
i
t
h
t
he
e
xp
e
ri
m
e
n
t
a
l
d
a
t
a
t
r
a
i
n
i
ng
ra
t
i
o
a
nd
t
he
t
e
s
t
i
s
8
0:
20
.
W
i
t
h
t
he
gi
v
e
n
l
e
a
rni
ng
r
a
t
e
,
t
he
num
be
r
of
p
a
rt
i
c
l
e
s
a
nd
fi
x
e
d
m
o
m
e
nt
u
m
,
t
h
e
e
xp
e
r
i
m
e
nt
i
s
c
on
duc
t
e
d
t
o
s
e
e
t
h
e
e
ff
e
c
t
gi
v
e
n
by
t
h
e
nu
m
b
e
r
of
di
ff
e
re
nt
ne
u
rons
i
n
t
he
hi
dde
n
l
a
y
e
r.
A
s
s
e
e
n
i
n
T
a
b
l
e
3
,
t
h
e
v
a
l
u
e
ge
n
e
r
a
t
e
d
on
t
h
e
a
rc
h
i
t
e
c
t
ure
w
i
t
h
t
h
e
nu
m
b
e
r
of
n
e
rve
s
i
n
t
h
e
hi
dd
e
n
l
a
y
e
r
i
s
3
ha
s
t
h
e
s
m
a
l
l
e
s
t
RM
S
E
va
l
ue
t
ha
n
t
he
o
t
he
rs
.
T
h
i
s
pro
ve
s
t
h
a
t
t
he
nu
m
b
e
r
of
ne
rve
s
or
nod
e
s
i
n
t
he
h
i
dd
e
n
l
a
y
e
r
ha
s
a
n
e
ffe
c
t
i
n
m
i
n
i
m
i
z
i
ng
t
he
e
rrors
t
h
a
t
o
c
c
u
r,
o
nl
y
t
he
va
l
ue
gi
v
e
n
do
e
s
not
ha
v
e
a
s
i
gni
f
i
c
a
nt
e
ffe
c
t
.
T
h
e
n
t
h
e
e
xp
e
ri
m
e
nt
w
i
l
l
be
c
a
rr
i
e
d
ou
t
a
g
a
i
n
by
l
o
oki
n
g
a
t
t
he
e
f
fe
c
t
of
t
h
e
gi
v
e
n
l
e
a
rn
i
ng
ra
t
e
on
t
he
va
l
u
e
of
t
he
pr
e
di
c
t
i
ons
.
T
h
e
re
s
u
l
t
of
pre
d
i
c
t
i
o
n
w
i
t
h
di
ff
e
r
e
nt
l
e
a
rn
i
ng
ra
t
e
c
a
n
be
s
e
e
n
be
l
ow
.
A
s
s
e
e
n
i
n
T
a
b
l
e
4
t
h
e
s
m
a
l
l
e
s
t
e
rro
r
va
l
u
e
i
s
g
e
n
e
ra
t
e
d
by
a
l
e
a
rni
n
g
ra
t
e
w
i
t
h
a
va
l
ue
of
4
.
T
he
r
e
s
ul
t
s
obt
a
i
ne
d
t
hro
ugh
t
hi
s
va
l
u
e
a
r
e
0.
1075
on
t
he
da
t
a
t
ra
i
ni
n
g
l
i
s
t
.
In
t
hi
s
c
a
s
e
t
he
l
e
a
rni
ng
ra
t
e
i
t
s
e
l
f
i
s
t
he
v
a
l
ue
of
t
h
e
ra
t
e
of
l
e
a
rn
i
ng
t
o
c
ont
r
ol
c
h
a
ng
e
s
i
n
w
e
i
gh
t
o
bt
a
i
ne
d
t
hrough
c
a
l
c
u
l
a
t
i
ons
i
n
e
a
c
h
i
t
e
ra
t
i
on
.
T
hus
,
t
h
e
l
e
a
rn
i
ng
r
a
t
e
va
l
ue
of
4
c
a
n
c
o
nt
ro
l
t
he
r
a
t
e
of
l
e
a
rni
ng
i
n
w
e
l
l
-
ob
t
a
i
ne
d
w
e
i
gh
t
s
no
t
l
i
k
e
t
he
ot
he
r
va
l
u
e
s
gi
v
e
n.
N
e
xt
,
t
h
e
e
xpe
r
i
m
e
nt
i
s
don
e
a
g
a
i
n
by
m
a
ki
ng
c
ha
n
ge
s
t
o
t
he
m
o
m
e
n
t
um
v
a
l
ue
,
t
h
i
s
a
l
s
o
a
i
m
s
a
t
w
he
t
h
e
r
t
he
va
l
ue
gi
v
e
n
c
a
n
a
f
fe
c
t
t
he
va
l
ue
of
t
he
e
r
rors
t
ha
t
oc
c
ur
.
T
h
e
fo
l
l
ow
i
ng
r
e
s
ul
t
s
fro
m
t
hi
s
e
xpe
ri
m
e
n
t
c
a
n
be
s
e
e
n
t
hro
ugh
t
h
e
f
ol
l
ow
i
ng
T
a
b
l
e
5
.
T
a
b
l
e
3
.
E
xpe
ri
m
e
n
t
r
e
s
ul
t
of
d
i
ff
e
re
n
t
hi
dd
e
n
l
a
ye
r
A
rc
h
i
t
e
c
t
u
re
α
M
o
m
e
n
t
u
m
M
o
m
e
n
t
u
m
RM
S
E
o
f
D
a
t
a
T
ra
i
n
i
n
g
RM
S
E
o
f
D
a
t
a
T
e
s
t
i
n
g
9
-
3
-
1
1
0
.
1
5
0
.
1
4
6
9
0
.
2
0
0
1
9
-
4
-
1
1
0
.
1
5
0
.
1
4
7
1
0
.
2
0
1
4
9
-
5
-
1
1
0
.
1
5
0
.
1
4
4
4
0
.
2
2
2
6
9
-
6
-
1
1
0
.
1
5
0
.
1
4
8
1
0
.
2
2
7
0
9
-
7
-
1
1
0
.
1
5
0
.
1
4
9
5
0
.
2
2
8
7
T
a
b
l
e
4
.
E
xpe
ri
m
e
n
t
r
e
s
ul
t
of
d
i
ff
e
re
n
t
l
e
a
rn
i
ng
ra
t
e
A
rc
h
i
t
e
c
t
u
re
L
e
a
r
n
i
n
g
R
a
t
e
M
o
m
e
n
t
u
m
P
a
rt
i
c
l
e
RM
S
E
o
f
D
a
t
a
T
ra
i
n
i
n
g
RM
S
E
o
f
D
a
t
a
T
e
s
t
i
n
g
9
-
3
-
1
1
0
.
1
5
0
.
1
4
6
9
0
.
2
0
0
1
9
-
3
-
1
2
0
.
1
5
0
.
1
3
4
6
0
.
2
1
2
9
9
-
3
-
1
3
0
.
1
5
0
.
1
3
0
4
0
.
2
1
7
4
9
-
3
-
1
4
0
.
1
5
0
.
1
0
7
5
0
.
1
7
9
7
9
-
3
-
1
5
0
.
1
5
0
.
1
1
5
6
0
.
2
0
9
8
T
a
b
l
e
5
.
E
xpe
ri
m
e
n
t
r
e
s
ul
t
of
d
i
ff
e
re
n
t
m
o
m
e
n
t
u
m
A
rc
h
i
t
e
c
t
u
re
α
M
o
m
e
n
t
u
m
P
a
rt
i
c
l
e
RM
S
E
D
a
t
a
T
ra
i
n
i
n
g
RM
S
E
D
a
t
a
T
e
s
t
i
n
g
9
-
3
-
1
4
0
.
1
5
0
.
1
0
7
5
0
.
1
7
9
7
9
-
3
-
1
4
0
.
2
5
0
.
0
9
1
1
0
.
1
5
9
8
9
-
3
-
1
4
0
.
3
5
0
.
0
9
2
3
0
.
1
5
3
2
9
-
3
-
1
4
0
.
4
5
0
.
0
9
6
1
0
.
1
6
4
6
9
-
3
-
1
4
0
.
5
5
0
.
0
9
0
1
0
.
1
5
1
1
A
s
s
e
e
n
i
n
T
a
bl
e
5,
t
h
e
r
e
s
ul
t
s
of
e
xpe
r
i
m
e
n
t
by
m
a
k
i
ng
c
ha
ng
e
s
t
o
t
he
m
o
m
e
nt
u
m
v
a
l
ue
do
not
i
nc
r
e
a
s
e
t
h
e
RM
S
E
va
l
ue
obt
a
i
n
e
d
.
T
h
e
b
e
s
t
r
e
s
ul
t
i
s
ob
t
a
i
n
e
d
w
he
n
t
h
e
m
om
e
nt
um
va
l
ue
g
i
v
e
n
i
s
0
.
5
.
T
hi
s
pro
ve
s
t
ha
t
t
h
e
m
o
m
e
nt
u
m
v
a
l
u
e
do
e
s
not
s
i
gni
f
i
c
a
nt
l
y
a
ff
e
c
t
t
he
RM
S
E
v
a
l
u
e
ob
t
a
i
ne
d
.
I
t
a
l
s
o
prov
e
s
t
ha
t
i
f
t
h
e
g
i
ve
n
m
o
m
e
nt
u
m
v
a
l
u
e
a
ppro
a
c
he
s
t
he
va
l
ue
of
1
,
w
he
re
1
i
s
t
he
m
a
xi
m
u
m
v
a
l
u
e
of
m
o
m
e
nt
u
m
,
ove
rs
hoo
t
w
i
l
l
oc
c
ur
.
T
h
i
s
i
s
a
n
e
rr
or
i
n
for
e
c
a
s
t
i
ng
w
he
r
e
t
he
va
l
ue
g
i
ve
n
e
xc
e
e
ds
t
he
v
a
l
u
e
t
h
a
t
s
hou
l
d
be
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
K
O
M
N
IK
A
T
e
l
e
c
om
m
un
Co
m
put
E
l
Con
t
rol
P
SO
o
pt
i
m
i
z
at
i
on
on
bac
k
pr
opag
at
i
on
f
or
f
i
s
h
c
a
t
c
h
pr
od
uc
t
i
on
pr
e
di
c
t
i
on
(
Y
us
l
e
na
Sar
i
)
781
a
nd
t
hi
s
c
a
us
e
s
t
he
c
h
a
ng
e
i
n
w
e
i
gh
t
obt
a
i
n
e
d
t
o
be
not
va
r
i
e
d
.
T
hi
s
proc
e
s
s
of
gi
v
i
ng
di
f
fe
r
e
nt
m
o
m
e
nt
u
m
va
l
u
e
s
d
e
t
e
rm
i
n
e
s
how
fa
s
t
t
he
pro
c
e
s
s
i
s
ne
e
de
d
[
21
]
.
T
he
ne
xt
e
xp
e
ri
m
e
nt
i
s
t
o
fi
nd
t
he
s
m
a
l
l
e
s
t
RM
S
E
va
l
ue
ob
t
a
i
ne
d
by
c
ha
ngi
n
g
t
he
va
l
ue
of
t
h
e
e
x
i
s
t
i
n
g
pa
rt
i
c
l
e
s
.
T
he
fo
l
l
ow
i
ng
c
o
m
p
a
ri
s
on
c
a
n
be
s
e
e
n
i
n
T
a
bl
e
6.
T
a
b
l
e
6
s
how
s
t
he
re
s
u
l
t
s
obt
a
i
ne
d
a
f
t
e
r
c
ha
ngi
ng
t
he
n
um
b
e
r
o
f
pa
rt
i
c
l
e
s
.
It
i
s
foun
d
t
h
a
t
t
h
e
a
d
di
t
i
on
of
p
a
rt
i
c
l
e
s
c
a
n
a
c
t
ua
l
l
y
m
a
ke
t
h
e
RM
S
E
va
l
u
e
l
a
rge
r,
i
ne
ff
e
c
t
i
v
e
a
n
d
a
l
s
o
i
n
t
h
e
pro
c
e
s
s
c
a
rr
i
e
d
ou
t
t
h
e
t
i
m
e
ne
e
de
d
i
s
l
onge
r.
T
hi
s
i
s
b
e
c
a
us
e
t
he
nu
m
b
e
r
of
p
a
rt
i
c
l
e
s
gi
ve
n
i
s
gr
e
a
t
e
r
s
o
t
ha
t
t
h
e
s
e
a
r
c
h
f
or
t
he
b
e
s
t
va
l
ue
t
a
ke
s
l
o
n
ge
r.
T
a
b
l
e
6
.
E
xpe
ri
m
e
n
t
r
e
s
ul
t
of
d
i
ff
e
re
n
t
a
m
ount
of
p
a
rt
i
c
l
e
s
A
rc
h
i
t
e
c
t
u
re
α
M
o
m
e
n
t
u
m
P
a
rt
i
c
l
e
RM
S
E
D
a
t
a
T
ra
i
n
i
n
g
RM
S
E
D
a
t
a
T
e
s
t
i
n
g
9
-
3
-
1
4
0
.
5
5
0
.
0
9
0
1
0
.
1
5
1
1
9
-
3
-
1
4
0
.
5
6
0
.
0
9
7
2
0
.
1
5
2
3
9
-
3
-
1
4
0
.
5
7
0
.
0
9
0
9
0
.
1
5
4
4
4.
2
.
2
M
e
th
od
c
omp
ar
i
s
on
T
he
c
o
m
pa
r
i
s
on
r
e
s
ul
t
s
of
t
h
e
t
w
o
m
e
t
ho
ds
na
m
e
l
y
B
a
c
kp
rop
a
ga
t
i
on
us
i
ng
P
S
O
a
nd
B
a
c
kp
ropa
ga
t
i
on
w
i
t
hou
t
us
i
n
g
P
S
O
c
a
n
b
e
s
e
e
n
t
h
rough
t
h
e
fol
l
ow
i
ng
T
a
b
l
e
7.
T
h
e
r
e
s
ul
t
s
of
t
h
e
s
e
c
o
m
pa
ri
s
ons
pr
ove
t
ha
t
t
he
r
e
s
ul
t
s
g
i
v
e
n
by
t
he
B
a
c
kprop
a
ga
t
i
o
n
m
e
t
ho
d
us
i
ng
P
S
O
g
e
t
b
e
t
t
e
r
RM
S
E
re
s
u
l
t
s
t
ha
n
m
e
t
hods
t
ha
t
onl
y
us
e
B
a
c
kprop
a
ga
t
i
o
n.
T
hi
s
a
l
s
o
pr
ove
s
t
h
a
t
i
n
t
h
e
t
ra
i
ni
ng
proc
e
s
s
t
h
e
P
S
O
m
e
t
hod
i
s
m
o
re
e
f
fe
c
t
i
ve
t
h
a
n
t
he
Ba
c
kpro
pa
g
a
t
i
on
m
e
t
ho
d.
t
h
e
r
e
s
ul
t
s
of
t
h
e
c
om
pa
r
i
s
on
be
t
w
e
e
n
t
h
e
p
re
d
i
c
t
i
v
e
v
a
l
u
e
a
nd
t
h
e
a
c
t
ua
l
va
l
ue
of
fi
s
h
c
a
t
c
h
produ
c
t
i
on
i
n
t
on
g
i
v
e
n
by
t
he
t
w
o
m
e
t
hods
c
a
n
be
s
e
e
n
t
h
rough
t
h
e
fol
l
ow
i
ng
T
a
b
l
e
8
.
T
a
b
l
e
7
.
RM
S
E
c
o
m
pa
r
i
s
on
of
t
w
o
m
e
t
hod
M
e
t
h
o
d
RM
S
E
T
ra
i
n
i
n
g
RM
S
E
T
e
s
t
i
n
g
Ba
c
k
p
ro
p
a
g
a
t
i
o
n
+
P
S
O
0
.
0
9
0
9
0
.
1
4
4
8
Ba
c
k
p
ro
p
a
g
a
t
i
o
n
0
.
1
4
8
3
0
.
2
1
2
1
T
a
b
l
e
8
.
Co
m
p
a
ri
s
on
of
A
c
t
ua
l
D
a
t
a
a
nd
P
re
d
i
c
t
i
o
n
D
a
t
a
fr
o
m
T
w
o
M
e
t
h
ods
A
c
t
u
a
l
D
a
t
a
(T
o
n
)
P
re
d
i
c
t
i
o
n
Re
s
u
l
t
Ba
c
k
p
ro
p
a
g
a
t
i
o
n
+
P
S
O
(T
o
n
)
Ba
c
k
p
ro
p
a
g
a
t
i
o
n
(T
o
n
)
3
4
1
6
1
4
0
5
3
1
4
0
2
0
4
4
7
2
5
9
4
1
2
5
0
4
0
1
4
2
3
6
7
8
5
4
1
7
0
4
4
2
3
1
2
5
8
2
9
9
5
8
2
5
7
5
8
0
2
1
D
i
ffe
re
n
c
e
A
v
e
ra
g
e
s
4
.
2
5
7
,
7
5
4
.
7
4
1
,
2
5
T
a
b
l
e
8
s
h
ow
s
t
ha
t
t
he
B
a
c
kprop
a
ga
t
i
o
n
onl
y
m
e
t
h
od
gi
v
e
s
t
he
m
os
t
di
ff
e
r
e
nc
e
re
s
u
l
t
be
t
w
e
e
n
t
he
a
c
t
ua
l
va
l
u
e
a
nd
t
h
e
pr
e
di
c
t
i
ve
v
a
l
ue
.
T
h
i
s
c
a
n
b
e
s
e
e
n
fr
om
t
h
e
r
e
s
ul
t
s
o
f
t
h
e
a
ve
r
a
ge
di
ff
e
re
nc
e
gi
v
e
n
by
t
he
t
w
o
m
e
t
hods
t
o
t
h
e
a
c
t
u
a
l
v
a
l
u
e
.
Ba
c
kprop
a
g
a
t
i
on
+
P
S
O
m
e
t
hod
ha
s
a
n
a
v
e
r
a
ge
d
i
ff
e
re
n
c
e
o
f
4,
2
57.
7
5
w
hi
l
e
t
he
a
v
e
ra
g
e
di
ffe
r
e
n
c
e
r
e
s
ul
t
s
g
i
ve
n
by
t
h
e
B
a
c
kp
ropa
g
a
t
i
on
onl
y
m
e
t
ho
d
a
r
e
4.
741.
25.
T
hus
,
t
h
e
re
s
u
l
t
s
of
pr
e
d
i
c
t
e
d
v
a
l
u
e
t
h
a
t
a
re
c
l
os
e
r
t
o
t
h
e
a
c
t
u
a
l
v
a
l
u
e
a
r
e
o
bt
a
i
ne
d
b
y
t
he
B
a
c
kprop
a
ga
t
i
on
+
P
S
O
m
e
t
hod
,
be
c
a
us
e
t
h
e
di
ff
e
re
nc
e
b
e
t
w
e
e
n
pr
e
di
c
t
e
d
a
nd
a
c
t
ua
l
v
a
l
u
e
i
s
s
m
a
l
l
e
r
.
5.
C
O
N
C
LU
S
I
O
N
Ba
s
e
d
o
n
t
h
e
r
e
s
ul
t
s
of
t
h
e
a
na
l
ys
i
s
a
nd
d
i
s
c
us
s
i
on
of
t
h
e
e
xpe
r
i
m
e
n
t
s
t
h
a
t
h
a
v
e
be
e
n
c
a
rr
i
e
d
out
,
i
t
c
a
n
b
e
c
on
c
l
u
de
d
t
h
a
t
t
he
re
s
u
l
t
s
of
pre
d
i
c
t
i
ons
ob
t
a
i
ne
d
a
r
e
qu
i
t
e
good
.
I
t
i
s
s
e
e
n
i
n
t
h
e
a
v
e
ra
ge
di
f
fe
r
e
nc
e
be
t
w
e
e
n
t
he
pre
d
i
c
t
e
d
a
nd
t
h
e
a
c
t
u
a
l
va
l
ue
of
4
,
257
.
75
.
A
s
of
pre
d
i
c
t
i
ons
of
fi
s
h
pr
oduc
t
i
o
n
on
c
l
i
m
a
t
e
c
h
a
ng
e
c
a
n
be
i
m
pl
e
m
e
n
t
e
d
us
i
n
g
t
he
Ba
c
kpr
opa
g
a
t
i
on
m
e
t
ho
d
c
om
b
i
n
e
d
w
i
t
h
P
S
O
.
P
r
e
di
c
t
i
on
m
od
e
l
s
us
i
ng
m
ont
hl
y
c
l
i
m
a
t
e
da
t
a
fro
m
t
he
M
e
t
e
or
ol
og
i
c
a
l
S
t
a
t
i
on
S
y
a
m
s
ud
i
n
N
o
or
p
rodu
c
e
RM
S
E
of
0.
14
48
w
i
t
h
a
c
o
m
bi
n
a
t
i
on
of
p
a
ra
m
e
t
e
rs
us
e
d,
n
a
m
e
l
y,
C1
:
2,
C
2:
2
,
w
:
0
.
7,
l
e
a
rni
ng
ra
t
e
:
4,
M
o
m
e
n
t
u
m
:
0
.
5,
P
a
rt
i
c
l
e
s
:
5
,
a
nd
e
po
c
h:
500
.
F
rom
t
he
s
e
re
s
u
l
t
s
i
t
c
a
n
a
l
s
o
be
c
on
c
l
u
de
d
t
h
a
t
t
o
obt
a
i
n
t
h
e
a
ppropri
a
t
e
pr
e
d
i
c
t
i
on
m
o
de
l
,
t
h
e
op
t
i
m
a
l
c
om
b
i
n
a
t
i
on
of
pa
r
a
m
e
t
e
rs
a
nd
t
h
e
s
ui
t
a
b
i
l
i
t
y
of
t
h
e
da
t
a
w
i
t
h
t
h
e
pa
r
a
m
e
t
e
rs
a
r
e
n
e
e
de
d
t
o
g
e
t
t
he
be
s
t
pre
di
c
t
i
ve
m
o
de
l
re
s
u
l
t
s
.
In
a
dd
i
t
i
on,
t
h
e
a
m
oun
t
of
da
t
a
us
e
d
a
l
s
o
ha
s
i
nfl
u
e
n
c
e
,
b
e
c
a
us
e
t
he
m
or
e
d
a
t
a
us
e
d
i
n
t
h
e
t
ra
i
ni
ng
pro
c
e
s
s
a
nd
t
h
e
m
or
e
da
t
a
va
r
i
a
t
i
ons
c
a
n
b
e
l
e
a
rne
d
by
t
he
m
od
e
l
t
h
e
a
c
c
u
ra
c
y
r
e
s
ul
t
s
obt
a
i
n
e
d
w
i
l
l
b
e
hi
ghe
r
.
By
a
ppl
yi
ng
t
he
Ba
c
kprop
a
g
a
t
i
on
a
n
d
P
S
O
m
e
t
hod
,
t
h
e
RM
S
E
a
c
c
ura
c
y
v
a
l
u
e
ob
t
a
i
ne
d
f
rom
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
1693
-
6930
T
E
L
K
O
M
N
IK
A
T
e
l
e
c
om
m
un
Co
m
put
E
l
Con
t
rol
,
V
ol
.
18
,
N
o.
2
,
A
pri
l
2
020:
7
7
6
-
7
8
2
782
t
he
pr
e
di
c
t
i
on
i
s
0
.
144
8.
T
hi
s
m
e
a
ns
t
h
a
t
t
he
m
e
t
hod
a
ppl
i
e
d
i
n
t
hi
s
s
t
ud
y
i
s
good
e
noug
h
t
o
pr
e
di
c
t
f
i
s
h
produc
t
i
o
n
a
ga
i
ns
t
c
l
i
m
a
t
e
c
h
a
ng
e
(
ra
i
nfa
l
l
,
w
i
nd
s
pe
e
d,
a
nd
t
e
m
pe
r
a
t
u
re
)
R
EF
ER
EN
C
ES
[
1]
F
or
t
y
-
ni
n
t
h
M
e
e
t
i
ng
of
t
he
C
oun
c
i
l
S
out
he
a
s
t
A
s
i
a
n
F
i
s
h
e
r
i
e
s
D
e
ve
l
op
m
e
n
t
C
e
n
t
e
r
,
“
M
e
e
t
i
ng
a
g
r
e
e
d
t
h
a
t
t
he
S
E
A
S
O
F
I
A
c
o
ul
d
be
p
ub
l
i
s
he
d
on
a
r
e
g
ul
a
r
b
a
s
i
s
,
”
S
out
he
a
s
t
A
s
i
an
S
t
a
t
e
O
f
F
i
s
he
r
i
e
s
A
nd
A
qu
ac
u
l
t
ur
e
(
SE
A
SO
F
I
A
)
,
no
.
A
pr
i
l
.
2017
.
[
2]
F
ood
a
nd
A
g
r
i
c
ul
t
u
r
e
O
r
ga
n
i
z
a
t
i
on
of
t
he
U
n
i
t
e
d
N
a
t
i
on
s
,
“
T
he
S
t
a
t
e
O
f
W
or
l
d
F
i
s
h
e
r
i
e
s
A
n
d
A
q
ua
c
ul
t
u
r
e
,
”
2018
.
[
3]
A
s
i
a
n
D
e
ve
l
op
m
e
nt
B
a
n
k,
"
I
ndone
s
i
a
:
E
n
e
r
gy
S
e
c
t
o
r
A
s
s
e
s
s
m
e
nt
,
S
t
r
a
t
e
gy
,
a
nd
R
oa
d
M
a
p
,"
201
6.
[
4]
M
.
A
z
h
a
r
,
E
.
R
i
ks
a
ko
m
a
r
a
,
A
.
P
.
T
e
r
ka
i
t
,
“
F
o
r
e
c
a
s
t
i
n
g
t
he
A
m
oun
t
of
F
i
s
h
P
r
od
uc
t
i
on
b
y
U
s
i
ng
a
B
a
c
kpr
opa
g
a
t
i
on
N
e
ur
a
l
N
e
t
w
or
k
(
C
a
s
e
S
t
udy
:
U
P
T
D
B
a
n
j
a
r
m
a
s
i
n
F
i
s
he
r
y
P
o
r
t
)
(
i
n
B
a
ha
s
a
:
P
e
r
a
m
a
l
a
n
J
u
m
l
a
h
P
r
odu
ks
i
I
k
a
n
de
n
ga
n
M
e
n
ggun
a
ka
n
B
a
c
k
pr
o
pa
g
a
t
i
on
N
e
ur
a
l
N
e
t
w
or
k
(
S
t
ud
i
K
a
s
u
s
:
U
P
T
D
P
e
l
a
b
uha
n
P
e
r
i
k
a
na
n
B
a
nj
a
r
m
a
s
i
n)
)
,
”
J
.
T
e
k
.
I
T
S
,
vo
l
.
6
,
no
.
1,
p
p.
14
2
–
14
8,
20
17.
[
5]
P
.
K
.
S
a
r
i
,
A
.
M
.
A
ba
di
,
“
P
r
e
d
i
c
t
i
on
of
C
a
t
f
i
s
h
P
r
o
duc
t
i
on
i
n
S
l
e
m
a
n
R
e
ge
n
c
y
w
i
t
h
N
e
u
r
o
F
uz
z
y
M
ode
l
(
i
n
B
a
ha
s
a
:
P
r
e
d
i
k
s
i
P
r
o
duk
s
i
I
ka
n
L
e
l
e
D
i
K
a
bup
a
t
e
n
S
l
e
m
a
n
D
e
ng
a
n
M
o
de
l
N
e
ur
o
F
u
z
z
y
)
,
”
P
r
os
.
Se
m
i
n.
N
a
s
.
P
e
n
e
l
i
t
i
a
n,
P
e
nd
i
di
k
.
da
n
P
e
ne
r
ap
an
M
I
P
A
,
p
p.
1
–
7,
20
12
.
[
6]
Y
.
E
.
L
u,
L
.
I
.
Y
up
i
ng
,
L
.
W
e
i
hong
,
S
.
Q
i
da
o
,
L
.
I
.
U
.
Y
a
nq
un
,
a
n
d
Q
.
I
.
N
.
X
i
a
ol
i
,
“
V
e
g
e
t
a
b
l
e
P
r
i
c
e
P
r
e
d
i
c
t
i
on
B
a
s
e
d
on
P
S
O
-
B
P
N
e
ur
a
l
N
e
t
w
or
k,
”
201
5
8t
h
I
n
t
e
r
na
t
i
o
nal
C
onf
e
r
e
nc
e
o
n
I
n
t
e
l
l
i
g
e
nt
C
om
pu
t
at
i
on
T
e
c
h
no
l
ogy
and
A
ut
om
a
t
i
on
(
I
C
I
C
T
A
)
,
2
015
.
[
7]
C
.
S
a
r
a
nya
a
nd
G
.
M
a
n
i
ka
n
da
n
,
“
A
S
t
udy
on
N
or
m
a
l
i
z
a
t
i
on
T
e
c
h
ni
que
s
f
o
r
P
r
i
va
c
y
P
r
e
s
e
r
vi
n
g
D
a
t
a
M
i
n
i
ng
,
”
I
n
t
.
J
.
E
ng.
T
e
c
hno
l
.
,
vo
l
.
5,
no
.
3,
p
p.
27
01
–
2
704
,
2013
.
[
8]
A
.
A
.
N
a
e
i
n
i
,
M
.
B
a
ba
d
i
,
a
nd
S
.
H
o
m
a
you
ni
,
“
A
s
s
e
s
s
m
e
nt
O
f
N
or
m
a
l
i
z
a
t
i
o
n
T
e
c
hn
i
qu
e
s
O
n
T
he
A
c
c
ur
a
c
y
O
f
H
ype
r
s
pe
c
t
r
a
l
D
a
t
a
C
l
us
t
e
r
i
n
g,
”
I
n
t
.
A
r
c
h
.
P
h
ot
o
gr
am
m
.
R
e
m
o
t
e
S
e
ns
.
Spa
t
.
I
n
f
.
Sc
i
.
,
vo
l
.
X
L
I
I
,
n
o.
O
c
t
ob
e
r
,
pp.
7
–
10,
2
017
.
[
9]
B
.
D
u
a
n,
L
.
H
a
n
,
Z
.
G
ou
,
Y
.
Y
a
n
g,
a
nd
S
.
C
h
e
n,
“
C
l
us
t
e
r
i
ng
M
i
xe
d
D
a
t
a
B
a
s
e
d
on
D
e
ns
i
t
y
P
e
a
k
s
a
nd
S
t
a
c
ke
d
D
e
noi
s
i
ng
A
ut
oe
nc
ode
r
s
,
”
Sy
m
m
e
t
r
y
(
B
as
e
l
)
.
,
vol
.
163
,
no
.
11
,
p
p.
1
–
23
,
2019
.
[
10]
J
.
K
.
I
p
i
n
j
ol
u,
I
.
M
a
g
a
w
a
t
a
,
a
nd
B
.
A
.
S
hi
n
ka
f
i
,
“
P
o
t
e
nt
i
a
l
i
m
p
a
c
t
o
f
c
l
i
m
a
t
e
c
ha
ng
e
on
f
i
s
he
r
i
e
s
a
nd
a
q
ua
c
u
l
t
ur
e
i
n
N
i
ge
r
i
a
,
”
J
.
F
i
s
h
.
A
qua
t
.
S
c
i
.
,
v
ol
.
9,
n
o.
5
,
pp.
3
38
–
3
44,
2
014
.
[
11]
D
.
S
.
k
hos
hne
v
i
s
Y
a
z
di
a
n
d
D
.
B
.
S
h
a
ko
ur
i
,
“
T
h
e
E
f
f
e
c
t
s
of
C
l
i
m
a
t
e
C
ha
ng
e
on
A
q
ua
c
ul
t
u
r
e
,
”
I
n
t
.
J
.
E
n
v
i
r
on
.
Sc
i
.
D
e
v
.
,
vo
l
.
1
,
no.
5
,
pp
.
378
–
382
,
2010
.
[
12]
K
.
C
o
c
hr
a
ne
,
C
.
D
e
Y
ou
ng,
D
.
S
o
t
o
,
a
nd
T
.
B
a
hr
i
,
“
C
l
i
m
a
t
e
c
ha
nge
i
m
pl
i
c
a
t
i
on
s
f
o
r
fi
s
he
r
i
e
s
a
nd
a
qu
a
c
u
l
t
ur
e
:
O
ve
r
v
i
e
w
o
f
c
ur
r
e
n
t
s
c
i
e
nt
i
fi
c
kn
ow
l
e
dge
,
”
F
A
O
F
i
s
h
.
A
qua
c
.
T
e
c
h
.
P
ap
.
,
n
o.
53
0,
p
p
.
221
,
200
9.
[
13]
M
.
D
i
a
l
l
o
,
B
.
L
e
g
r
a
s
,
E
.
R
a
y
,
A
.
E
nge
l
,
a
nd
J
.
A
.
A
ñe
l
,
“
G
l
o
ba
l
di
s
t
r
i
b
ut
i
on
o
f
C
O
2
i
n
t
h
e
upp
e
r
t
r
opos
phe
r
e
a
nd
s
t
r
a
t
os
p
he
r
e
,
”
A
t
m
o
s
ph
r
i
c
C
h
e
m
.
P
h
y
s
.
,
vo
l
.
17,
p
p.
38
61
–
3
878
,
20
17.
[
14]
W
.
W
a
ng
a
nd
R
.
N
e
m
a
n
i
,
“
D
y
n
a
m
i
c
s
o
f
g
l
o
ba
l
a
t
m
os
p
he
r
i
c
C
O
2
c
o
nc
e
nt
r
a
t
i
on
f
r
o
m
1850
t
o
20
10
:
a
l
i
n
e
a
r
a
ppr
oxi
m
a
t
i
on
,
”
B
i
oge
os
c
i
e
nc
e
s
,
v
ol
.
11
,
2
014
.
[
15]
C
.
H
e
i
n
z
e
e
t
al
.
,
“
T
he
o
c
e
a
n
c
a
r
bon
s
i
nk
–
i
m
pa
c
t
s
,
vul
ne
r
a
b
i
l
i
t
i
e
s
a
nd,
”
E
ar
t
h
Sy
s
t
.
D
y
n
.
,
vo
l
.
2
,
pp
.
327
–
358
,
201
5.
[
16]
S.
R
ua
no
a
nd
A
.
M
i
l
a
n
,
“
C
l
i
m
a
t
e
C
ha
ng
e
,
R
a
i
nf
a
l
l
P
a
t
t
e
r
n
s
,
L
i
ve
l
i
hoo
ds
A
nd
M
i
g
r
a
t
i
o
n
I
n
C
a
b
r
i
c
a
n
,
G
ua
t
e
m
a
l
a
,
”
2
014
.
[
17]
C
.
M
.
D
unn
i
ng
,
E
.
C
.
L
.
B
l
a
c
k
,
a
nd
R
.
P
.
A
l
l
a
n
,
“
T
h
e
ons
e
t
a
nd
c
e
s
s
a
t
i
on
o
f
s
e
a
s
o
na
l
r
a
i
n
f
a
l
l
o
ve
r
A
f
r
i
c
a
,
”
J
.
G
e
oph
y
s
.
R
e
s
.
A
t
m
os
.
,
vol
.
121
,
no
.
19
,
p
p.
11
405
–
1142
4,
20
16
.
[
18]
C
.
D
e
H
u
a
y
l
a
s
,
W
.
G
ur
g
i
s
e
r
,
I
.
J
ue
n,
K
.
S
i
ng
e
r
,
M
.
N
e
ubur
ge
r
,
a
n
d
S
.
S
c
h
a
uw
e
c
ke
r
,
“
C
om
pa
r
i
n
g
pe
a
s
a
n
t
s
’
p
e
r
c
e
p
t
i
ons
o
f
pr
e
c
i
p
i
t
a
t
i
on
c
ha
ng
e
w
i
t
h
pr
e
c
i
p
i
t
a
t
i
on
r
e
c
or
d
s
i
n
t
he
t
r
o
pi
c
a
l
,
”
E
ar
t
h
Sy
s
t
.
D
y
n
.
,
v
ol
.
7,
pp.
49
9
–
51
5,
20
16.
[
19]
P
.
S
.
P
a
t
ki
,
V
.
W
e
s
t
,
a
nd
V
.
V
K
e
l
ka
r
,
“
C
l
a
s
s
i
f
i
c
a
t
i
on
u
s
i
n
g
D
i
f
f
e
r
e
nt
N
or
m
a
l
i
z
a
t
i
on
T
e
c
hn
i
qu
e
s
i
n
S
up
por
t
V
e
c
t
o
r
M
a
c
hi
n
e
,
”
I
n
t
.
J
.
C
om
pu
t
.
A
ppl
.
,
p
p.
4
–
6,
2
013
.
[
20]
D
.
V
i
r
m
a
n
i
,
S
.
T
a
ne
j
a
,
a
nd
G
.
M
a
l
hot
r
a
,
“
N
or
m
a
l
i
z
a
t
i
on
b
a
s
e
d
K
m
e
a
ns
C
l
us
t
e
r
i
ng
A
l
go
r
i
t
h
m
,
”
I
nt
.
J
.
A
d
v
.
E
ng
.
R
e
s
.
Sc
i
.
,
p
p.
1
–
5,
20
15.
[
21]
G
.
A
i
r
l
a
n
gga
,
A
.
R
a
c
h
m
a
t
,
a
nd
D
.
L
a
pi
h
u,
“
C
o
m
pa
r
i
s
on
o
f
e
xpo
n
e
nt
i
a
l
s
m
oo
t
h
i
ng
a
n
d
ne
u
r
a
l
ne
t
w
or
k
m
e
t
h
od
t
o
f
or
e
c
a
s
t
r
i
c
e
p
r
od
uc
t
i
o
n
i
n
I
ndon
e
s
i
a
,
”
T
E
L
K
O
M
N
I
K
A
T
e
l
e
c
om
m
uni
c
at
i
on
C
om
p
ut
i
ng
E
l
e
c
t
r
o
n
i
c
s
and
C
on
t
r
ol
,
vol
.
17,
n
o.
3
,
pp.
1
367
–
1375
,
201
9.
[
22]
A
.
K
.
C
hou
dha
r
y
a
n
d
N
.
K
u
m
a
r
,
“
I
n
t
e
r
na
t
i
ona
l
J
o
ur
n
a
l
O
f
E
ngi
ne
e
r
i
ng
S
c
i
e
nc
e
s
&
R
e
s
e
a
r
c
h
T
e
c
hno
l
ogy
S
e
c
u
r
i
t
y
,
”
I
nt
.
J
.
E
ng.
S
c
i
.
R
e
s
.
T
e
c
hno
l
.
,
vo
l
.
6
,
no
.
9,
p
p
.
5,
2
017
.
[
23]
S
.
K
a
ul
,
R
.
S
ha
r
m
a
,
S
.
S
i
t
oo
t
,
a
nd
S
.
C
ha
udh
a
r
i
,
“
A
S
ur
ve
y
on
M
R
B
r
a
i
n
I
m
a
ge
S
e
g
m
e
nt
a
t
i
on
u
s
i
n
g
H
y
br
i
d
P
S
O
a
nd
A
F
S
A
,
”
p
p.
1
0883
–
108
90,
2
018
.
[
24]
E
.
H
a
da
v
a
nd
i
,
“
D
e
ve
l
o
pi
n
g
a
T
i
m
e
S
e
r
i
e
s
M
ode
l
B
a
s
e
d
O
n
P
a
r
t
i
c
l
e
,
”
201
0.
[
25]
J
.
H
.
Y
a
n
g
a
nd
W
.
D
ou,
“
A
n
e
w
p
r
e
d
i
c
t
i
on
m
e
t
hod
of
go
l
d
p
r
i
c
e
:
E
M
D
-
PSO
-
S
V
M
,
”
J
.
So
f
t
w
.
,
vo
l
.
9,
no
.
1
,
pp.
19
5
–
20
2,
20
14.
[
26]
M
.
K
a
ng
a
r
a
ni
f
a
r
a
h
a
n
i
a
nd
S
.
M
e
h
r
a
l
i
a
n
,
“
C
o
m
pa
r
i
s
on
B
e
t
w
e
e
n
A
r
t
i
f
i
c
i
a
l
N
e
ur
a
l
N
e
t
w
or
k
a
n
d
N
e
u
r
o
-
F
uz
z
y
f
o
r
G
o
l
d
P
r
i
c
e
P
r
e
di
c
t
i
on
,
”
2
013
.
Evaluation Warning : The document was created with Spire.PDF for Python.