T
E
L
K
O
M
NIKA
T
elec
o
mm
un
ica
t
io
n Co
m
pu
t
i
ng
E
lect
ro
nics
a
nd
Co
ntr
o
l
Vo
l.
2
3
,
No
.
2
,
A
p
r
il
2
0
2
5
,
p
p
.
402
~
415
I
SS
N:
1
6
9
3
-
6
9
3
0
,
DOI
: 1
0
.
1
2
9
2
8
/
T
E
L
KOM
NI
K
A
.
v
2
3
i
2
.
26621
402
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//telko
mn
ika
.
u
a
d
.
a
c.
i
d
Adv
a
nced
crop
yi
eld predi
ction usi
ng
m
a
chin
e learni
ng
and
deep learning
:
a
c
o
m
preh
ensiv
e re
v
iew
Ay
us
h Ana
n
d,
K
a
v
it
a
J
ha
j
ha
ria
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
mat
i
o
n
T
e
c
h
n
o
l
o
g
y
,
F
a
c
u
l
t
y
o
f
En
g
i
n
e
e
r
i
n
g
,
M
a
n
i
p
a
l
U
n
i
v
e
r
si
t
y
Jai
p
u
r
,
R
a
j
a
s
t
h
a
n
,
I
n
d
i
a
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Au
g
2
8
,
2024
R
ev
i
s
ed
Dec
2
8
,
2
0
2
4
A
cc
ep
ted
J
an
2
2
,
2
0
2
5
T
h
e
a
d
v
a
n
c
e
m
e
n
t
o
f
m
a
c
h
in
e
lea
rn
in
g
(M
L
)
a
n
d
d
e
e
p
lea
rn
i
n
g
(DL
)
tec
h
n
iq
u
e
s
h
a
s
sig
n
if
ica
n
tl
y
i
m
p
ro
v
e
d
c
ro
p
y
ield
p
re
d
ictio
n
,
m
a
k
i
n
g
it
m
o
re
a
c
c
u
ra
te
a
n
d
re
li
a
b
le.
In
th
is
re
v
ie
w
,
th
e
i
m
p
le
m
e
n
tatio
n
o
f
M
L
a
n
d
DL
a
lg
o
rit
h
m
s
f
o
r
c
ro
p
y
ield
p
re
d
icti
o
n
is
th
o
ro
u
g
h
ly
in
v
e
stig
a
ted
,
f
o
c
u
sin
g
o
n
th
e
ir
c
ru
c
ial
ro
le
in
e
n
h
a
n
c
in
g
c
ro
p
p
r
o
d
u
c
ti
v
it
y
.
A
lo
n
g
w
it
h
M
L
a
n
d
DL
a
lg
o
rit
h
m
s
e
x
a
m
in
e
,
th
e
re
v
ie
w
a
n
a
l
y
se
s
th
e
u
se
o
f
r
e
m
o
te
se
n
sin
g
tec
h
n
o
l
o
g
ies
,
su
c
h
a
s
sa
telli
te
a
n
d
d
ro
n
e
d
a
ta,
i
n
p
r
o
v
id
in
g
h
ig
h
-
re
so
lu
ti
o
n
in
p
u
ts
e
ss
e
n
ti
a
l
f
o
r
a
c
c
u
ra
te
y
iel
d
p
re
d
ictio
n
s.
T
h
e
stu
d
y
id
e
n
ti
f
ies
th
e
sta
te
o
f
a
rt
a
lg
o
rit
h
m
s,
m
o
st
u
se
d
f
e
a
t
u
re
s,
d
a
ta
so
u
rc
e
s
a
n
d
e
v
a
lu
a
ti
o
n
m
e
tri
c
s,
p
ro
v
id
i
n
g
a
c
o
m
p
a
riso
n
o
f
M
L
a
n
d
DL
.
T
h
e
f
in
d
in
g
s
in
d
ica
te
th
a
t
DL
m
o
d
e
ls
a
re
m
o
re
e
ff
e
c
ti
v
e
w
it
h
larg
e
d
a
tas
e
ts,
w
h
il
e
M
L
m
o
d
e
ls
re
m
a
in
ro
b
u
st
f
o
r
sm
a
ll
e
r
d
a
tas
e
ts.
T
h
e
fu
t
u
re
d
irec
ti
o
n
s
a
re
p
r
o
p
o
se
d
to
d
e
v
e
lo
p
th
e
g
e
n
e
ra
li
se
d
m
o
d
e
ls
f
o
r
d
iff
e
re
n
t
c
ro
p
s
a
n
d
re
g
io
n
s.
T
h
e
re
v
ie
w
a
i
m
s
to
a
ss
ist
re
se
a
rc
h
e
rs
b
y
su
m
m
a
risin
g
st
a
te
o
f
a
rt
te
c
h
n
iq
u
e
s
a
n
d
id
e
n
ti
fy
in
g
th
e
p
re
se
n
t.
K
ey
w
o
r
d
s
:
C
r
o
p
y
ield
p
r
ed
ictio
n
Dee
p
lear
n
in
g
Ma
ch
i
n
e
lear
n
i
n
g
R
e
m
o
te
s
e
n
s
in
g
S
y
s
te
m
a
tic
liter
at
u
r
e
r
ev
ie
w
Veg
etatio
n
in
d
ice
s
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Kav
ita
J
h
aj
h
ar
ia
Dep
ar
t
m
en
t o
f
I
n
f
o
r
m
a
tio
n
T
e
ch
n
o
lo
g
y
,
Fac
u
lt
y
o
f
E
n
g
i
n
ee
r
in
g
,
Ma
n
ip
al
U
n
iv
er
s
it
y
J
aip
u
r
R
aj
asth
a
n
3
0
3
0
0
7
,
I
n
d
ia
E
m
ail:
Ka
v
ita.
j
h
aj
h
ar
ia@
j
aip
u
r
.
m
a
n
ip
al.
ed
u
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
f
ield
o
f
co
m
p
u
ter
s
cie
n
c
e
is
co
n
s
tan
t
l
y
ad
v
an
ci
n
g
an
d
ev
er
-
e
v
o
lv
in
g
,
d
r
iv
e
n
b
y
t
h
e
p
u
r
s
u
i
t
o
f
ev
en
m
o
r
e
s
o
p
h
i
s
ticated
s
o
l
u
t
io
n
s
to
co
m
p
le
x
p
r
o
b
lem
s
.
M
ac
h
in
e
lear
n
i
n
g
(
ML
)
h
as
e
m
er
g
ed
as
a
p
o
w
er
f
u
l
p
ar
ad
ig
m
w
it
h
i
n
th
i
s
d
o
m
ai
n
,
en
ab
lin
g
co
m
p
u
ter
s
to
lear
n
an
d
ad
ap
t
w
it
h
o
u
t
e
x
p
licit
p
r
o
g
r
a
m
m
i
n
g
[
1
]
.
ML
en
co
m
p
as
s
es
a
d
iv
er
s
e
s
et
o
f
tech
n
iq
u
es,
ea
ch
w
it
h
its
u
n
iq
u
e
s
tr
en
g
th
s
an
d
ap
p
licatio
n
s
[
2
]
.
So
m
e
co
m
m
o
n
ap
p
r
o
ac
h
es
in
clu
d
e
s
u
p
er
v
i
s
ed
lear
n
in
g
,
w
h
ich
i
n
v
o
l
v
es
tr
ai
n
in
g
alg
o
r
it
h
m
s
o
n
lab
elled
d
ata
to
p
er
f
o
r
m
tas
k
s
lik
e
clas
s
i
f
icatio
n
a
n
d
r
eg
r
ess
io
n
.
Un
s
u
p
er
v
i
s
ed
lear
n
i
n
g
,
o
n
th
e
o
t
h
er
h
an
d
,
f
o
c
u
s
e
s
o
n
u
n
co
v
er
in
g
h
id
d
en
s
tr
u
ct
u
r
es
w
i
th
in
u
n
lab
elled
d
ata,
allo
w
i
n
g
f
o
r
task
s
li
k
e
d
ata
clu
s
ter
in
g
a
n
d
d
i
m
e
n
s
io
n
alit
y
r
ed
u
ctio
n
.
A
d
d
itio
n
al
l
y
,
r
ei
n
f
o
r
ce
m
en
t
l
ea
r
n
in
g
e
n
ab
les
s
y
s
te
m
s
to
l
ea
r
n
th
r
o
u
g
h
tr
ial
an
d
er
r
o
r
,
in
ter
ac
ti
n
g
w
i
th
a
n
en
v
ir
o
n
m
e
n
t.
ML
is
al
s
o
m
ak
in
g
s
i
g
n
if
ican
t
co
n
tr
ib
u
t
io
n
in
t
h
e
ag
r
ic
u
l
tu
r
e
in
d
u
s
tr
y
,
p
ar
ticu
lar
l
y
i
n
th
e
ar
ea
o
f
cr
o
p
y
ield
p
r
ed
ictio
n
[
3
]
.
ML
ca
n
h
e
lp
f
ar
m
er
s
a
n
d
p
o
lic
y
m
ak
er
s
m
iti
g
ate
f
o
o
d
in
s
ec
u
r
itie
s
.
I
t
is
b
ased
o
n
th
e
co
n
ce
p
t
o
f
s
tatis
tic
s
an
d
ML
in
w
h
ic
h
cr
o
p
y
ield
is
p
r
ed
icte
d
u
s
in
g
h
is
to
r
ical
d
ata
ass
o
ciate
d
w
ith
t
h
e
cr
o
p
s
lik
e
cli
m
ate,
s
o
il,
an
d
r
eg
io
n
.
Mo
d
er
n
to
o
ls
s
u
ch
as
s
ate
l
lite
s
,
d
r
o
n
es
an
d
s
en
s
o
r
s
ar
e
also
u
s
ed
to
o
b
tain
d
ata
an
d
m
o
n
ito
r
cr
o
p
s
.
On
e
o
f
th
e
k
e
y
d
r
iv
er
s
o
f
th
is
p
r
o
g
r
ess
is
th
e
in
teg
r
atio
n
o
f
r
em
o
te
s
e
n
s
in
g
tec
h
n
o
lo
g
y
[
4
]
.
Satellite
s
an
d
d
r
o
n
es
eq
u
ip
p
ed
w
it
h
v
ar
io
u
s
s
e
n
s
o
r
s
ca
n
g
a
th
er
d
ata
o
n
f
ac
to
r
s
li
k
e
s
o
il
m
o
is
t
u
r
e,
v
e
g
etatio
n
h
ea
lt
h
,
an
d
w
ea
t
h
er
p
atter
n
s
f
r
o
m
a
d
is
ta
n
ce
[
5
]
.
T
h
ese
m
o
d
els
th
e
n
id
e
n
ti
f
y
in
tr
ica
te
r
el
atio
n
s
h
ip
s
b
et
w
ee
n
th
ese
d
i
v
er
s
e
f
ac
to
r
s
a
n
d
h
i
s
to
r
ical
cr
o
p
y
ield
s
,
allo
w
i
n
g
f
o
r
ac
cu
r
ate
p
r
ed
ictio
n
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
A
d
va
n
ce
d
cro
p
yield
p
r
ed
ictio
n
u
s
in
g
m
a
ch
in
e
lea
r
n
in
g
a
n
d
d
ee
p
le
a
r
n
in
g
:
…
(
A
yu
s
h
A
n
a
n
d
)
403
W
ith
s
u
r
g
e
in
d
e
m
a
n
d
o
f
f
o
o
d
w
it
h
i
n
cr
ea
s
i
n
g
p
o
p
u
lat
io
n
,
ML
in
a
g
r
ic
u
lt
u
r
e
h
a
s
p
r
o
p
elled
to
th
e
f
o
r
ef
r
o
n
t
o
f
r
esear
ch
ai
m
ed
at
ad
v
an
ci
n
g
t
h
e
s
ec
to
r
.
Ho
w
e
v
er
,
n
av
ig
ati
n
g
th
e
co
m
p
l
ex
itie
s
o
f
ch
o
o
s
in
g
s
u
itab
le
d
atase
ts
,
al
g
o
r
ith
m
s
,
an
d
m
et
h
o
d
o
lo
g
ies
ca
n
b
e
ch
alle
n
g
i
n
g
f
o
r
r
esear
ch
er
s
as
th
ese
v
ar
y
g
r
ea
tl
y
d
ep
e
n
d
in
g
o
n
th
e
ar
ea
o
f
s
tu
d
y
a
n
d
t
y
p
e
o
f
cr
o
p
.
T
h
is
r
ev
iew
p
ap
er
ad
d
r
ess
es
q
u
esti
o
n
s
s
u
ch
as
th
e
m
o
s
t
u
s
ed
f
ea
t
u
r
es,
d
ata
s
o
u
r
ce
s
,
ty
p
es
o
f
ev
alu
at
io
n
m
etr
ic
s
,
th
e
alg
o
r
ith
m
s
an
d
m
o
d
els
u
s
ed
an
d
th
e
ty
p
e
o
f
r
e
m
o
te
s
en
s
in
g
tech
n
iq
u
es
u
s
ed
in
r
ec
en
t
s
t
u
d
ies.
O
u
r
r
ev
ie
w
ad
d
r
ess
es
t
h
ese
g
ap
s
b
y
co
m
p
ar
in
g
an
d
s
u
m
m
ar
izi
n
g
th
e
m
o
s
t
r
ec
en
t
ad
v
an
ce
s
b
ase
d
o
n
th
e
liter
at
u
r
e
av
ailab
le
to
an
s
w
er
o
u
r
p
r
ep
ar
ed
r
esear
ch
q
u
esti
o
n
t
h
at
ai
m
s
to
cr
ea
te
a
m
o
r
e
g
en
er
ali
s
ed
ap
p
r
o
ac
h
f
o
r
r
esear
ch
er
s
th
at
ca
n
b
e
u
s
ed
f
o
r
m
o
s
t
cr
o
p
s
an
d
ar
ea
s
.
T
h
e
aim
is
to
eq
u
ip
r
esear
ch
er
s
w
it
h
th
e
i
n
s
i
g
h
t
s
n
ee
d
ed
to
m
a
k
e
in
f
o
r
m
ed
d
ec
is
io
n
s
b
y
a
n
s
w
er
i
n
g
th
e
f
o
llo
w
i
n
g
q
u
est
io
n
s
:
R
Q1
:
w
h
at
ar
e
t
h
e
s
tate
-
of
-
t
h
e
-
ar
t te
ch
n
iq
u
e
s
u
s
ed
?
R
Q2
:
w
h
ic
h
a
m
o
n
g
M
L
an
d
d
ee
p
lear
n
in
g
(
DL
)
is
b
etter
f
o
r
m
ak
in
g
y
ield
p
r
ed
ictio
n
s
?
R
Q3
:
w
h
at
ar
e
m
atr
ices
u
s
ed
f
o
r
m
o
d
el
ev
al
u
atio
n
?
R
Q4
:
w
h
at
ar
e
t
h
e
d
ata
s
o
u
r
ce
s
?
R
Q5
:
w
h
at
ar
e
t
h
e
m
o
s
t u
s
ed
f
ea
tu
r
es?
R
Q6
:
w
h
ic
h
a
m
o
n
g
en
s
e
m
b
le
d
m
o
d
els a
n
d
tr
ad
itio
n
al
M
L
an
d
DL
p
er
f
o
r
m
b
etter
?
R
Q7
:
w
h
at
ar
e
t
h
e
li
m
itatio
n
s
an
d
f
u
tu
r
e
d
ir
ec
ti
o
n
s
?
Si
m
i
lar
r
ev
ie
w
s
ar
e
co
n
d
u
cte
d
b
y
r
esear
ch
er
s
b
u
t
ea
ch
v
ar
y
w
i
th
o
n
e
a
n
o
th
er
b
ased
o
n
th
e
t
y
p
e
o
f
cr
o
p
o
r
a
r
ea
b
ein
g
s
tu
d
ied
.
I
t
i
s
cr
u
cial
to
an
al
y
s
e
th
e
r
ec
en
t
r
ev
ie
w
s
to
g
et
in
s
i
g
h
t
s
o
n
th
e
r
ec
en
t
p
r
ac
tices
in
cr
o
p
y
ield
p
r
ed
ictio
n
.
A
cc
o
r
d
in
g
to
th
e
s
tu
d
y
[
6
]
w
h
ic
h
w
as
ca
r
r
ied
o
u
t
o
n
d
if
f
er
en
t
cr
o
p
s
,
g
eo
g
r
ap
h
ical
p
o
s
itio
n
s
an
d
v
ar
io
u
s
f
ea
t
u
r
e
s
.
I
t
w
as
f
o
u
n
d
th
a
t
DL
p
er
f
o
r
m
s
b
etter
th
an
M
L
f
o
r
m
a
k
in
g
p
r
ed
ictio
n
s
o
f
w
h
ic
h
co
n
v
o
lu
t
io
n
al
n
eu
r
al
n
et
w
o
r
k
(
C
NN
)
an
d
lo
n
g
s
h
o
r
t
-
ter
m
m
e
m
o
r
y
(
L
ST
M
)
-
b
ased
m
o
d
els
w
er
e
id
en
ti
f
ied
to
b
e
m
o
s
t
e
f
f
ec
tiv
e.
I
t
w
a
s
al
s
o
co
n
cl
u
d
ed
th
at
m
eteo
r
o
lo
g
ical
d
ata
a
n
d
Ve
g
e
tatio
n
ar
e
t
h
e
m
o
s
t
u
s
ed
f
ea
t
u
r
es.
Si
m
i
lar
l
y
,
t
h
e
r
ev
ie
w
[
7
]
also
in
clu
d
ed
p
ap
er
s
th
at
co
n
d
u
c
ted
s
tu
d
ie
s
p
er
f
o
r
m
ed
i
n
d
if
f
er
en
t
en
v
ir
o
n
m
e
n
t
s
s
ta
ted
th
at
th
er
e
w
er
e
n
o
s
i
n
g
le
o
r
m
u
l
tip
le
s
p
ec
if
ic
m
o
d
els
f
o
u
n
d
t
h
at
w
er
e
ab
le
to
o
u
tp
er
f
o
r
m
o
th
er
s
a
n
d
also
s
tated
th
a
t
i
n
c
lu
d
in
g
m
o
r
e
f
ea
tu
r
es
in
t
h
e
d
ataset
d
o
esn
’
t
n
ec
e
s
s
ar
il
y
m
ea
n
t
h
at
t
h
e
y
p
er
f
o
r
m
b
etter
.
Ho
w
ev
er
,
it
co
n
c
lu
d
ed
th
at
t
h
er
e
w
er
e
a
f
e
w
p
o
p
u
lar
m
o
d
els
t
h
at
ar
e
u
s
ed
v
er
y
o
f
t
en
s
u
ch
a
s
r
an
d
o
m
f
o
r
est
,
n
e
u
r
al
n
et
w
o
r
k
,
li
n
ea
r
r
eg
r
ess
io
n
,
an
d
g
r
ad
ien
t
b
o
o
s
ti
n
g
tr
e
e.
Fu
r
t
h
er
,
th
e
r
ev
ie
w
co
n
clu
d
ed
t
h
at
o
u
t
o
f
th
e
n
e
u
r
al
n
e
t
w
o
r
k
,
th
e
m
o
s
t
u
s
ed
m
o
d
els
w
er
e
C
NN,
L
ST
M
,
an
d
d
ee
p
n
eu
r
al
n
et
w
o
r
k
(
D
NN)
.
A
cc
o
r
d
in
g
to
th
e
s
tu
d
y
[
8
]
w
h
ich
w
a
s
co
n
d
u
cted
o
n
P
alm
o
il
p
r
ed
ictio
n
s
tated
th
at
w
h
i
l
e
th
er
e
w
as
n
o
p
ar
ticu
lar
alg
o
r
ith
m
t
h
at
c
o
u
ld
b
e
co
n
clu
d
ed
as
th
e
b
est
b
u
t
f
e
w
m
o
s
t
p
r
o
m
is
i
n
g
ML
alg
o
r
it
h
m
s
w
er
e
lin
ea
r
r
eg
r
ess
io
n
,
r
an
d
o
m
f
o
r
est
an
d
n
eu
r
al
n
et
w
o
r
k
.
Ou
t
o
f
th
e
DL
alg
o
r
it
h
m
s
,
th
e
p
o
p
u
lar
alg
o
r
ith
m
s
w
er
e
DNN,
C
NN
,
a
n
d
L
ST
M.
T
h
e
r
ev
ie
w
also
co
n
c
lu
d
ed
th
a
t
th
e
r
e
ar
e
v
er
y
f
e
w
s
t
u
d
ies
o
n
P
al
m
o
il
w
it
h
v
er
s
atil
e
f
ea
t
u
r
es
w
h
ic
h
m
ak
e
s
it
d
if
f
ic
u
lt
to
d
eter
m
in
e
w
h
ich
alg
o
r
it
h
m
o
r
f
ea
t
u
r
es
ar
e
b
es
t
s
i
n
ce
i
t’
s
s
til
l
i
n
t
h
e
ea
r
l
y
s
tag
e
s
.
A
cc
o
r
d
in
g
to
an
o
t
h
er
r
ev
ie
w
w
it
h
e
m
p
h
asi
s
s
p
ec
if
i
ca
ll
y
o
n
D
L
al
g
o
r
ith
m
s
f
o
r
y
ield
p
r
ed
ictio
n
[
9
]
.
C
r
o
p
y
ield
w
it
h
D
L
d
ep
en
d
s
m
aj
o
r
ly
o
n
t
h
e
t
y
p
e
o
f
d
ata
an
d
cr
o
p
s
.
I
t
w
a
s
also
n
o
ted
th
at
i
m
a
g
e
w
as
t
h
e
m
o
s
t
d
em
a
n
d
ed
s
o
u
r
ce
o
f
d
ata
w
it
h
th
e
m
aj
o
r
ity
o
f
p
u
b
licat
io
n
s
f
o
cu
s
i
n
g
o
n
s
u
p
er
v
i
s
ed
lear
n
i
n
g
.
C
N
N
w
as
w
id
el
y
u
s
ed
f
o
r
m
a
k
in
g
p
r
ed
ictio
n
s
w
h
ic
h
also
o
u
tp
er
f
o
r
m
ed
o
th
er
DL
al
g
o
r
ith
m
s
s
u
ch
a
s
D
NN,
L
ST
M,
Fas
ter
R
-
C
NN
an
d
h
y
b
r
id
m
o
d
els.
T
h
e
m
o
s
t
u
s
ed
ev
alu
at
io
n
m
etr
i
c
w
as
r
o
o
t
m
ea
n
s
q
u
ar
e
er
r
o
r
(
R
MSE
)
f
o
llo
w
e
d
b
y
R
^2
,
m
ea
n
ab
s
o
lu
te
p
er
ce
n
tag
e
er
r
o
r
(
MA
P
E
)
,
m
ea
n
ab
s
o
lu
te
er
r
o
r
(
MA
E
)
,
an
d
m
ea
n
s
q
u
ar
e
er
r
o
r
(
MSE
)
.
Si
m
i
lar
l
y
[
1
0
]
also
co
n
clu
d
ed
th
at
D
L
p
r
o
v
id
es a
p
r
o
m
i
s
i
n
g
s
o
lu
tio
n
f
o
r
cr
o
p
y
ield
esti
m
at
io
n
.
Ho
w
ev
er
,
t
h
e
y
ar
e
lar
g
el
y
d
ep
en
d
o
n
m
a
n
y
f
a
cto
r
s
in
cl
u
d
in
g
s
ca
lab
ilit
y
,
a
v
a
ilab
ilit
y
o
f
t
h
e
d
ata
s
et,
a
n
d
lo
ca
tio
n
o
f
s
t
u
d
y
.
W
e
s
till
ar
e
v
er
y
f
ar
f
r
o
m
f
i
n
d
i
n
g
a
g
en
er
alis
ed
ap
p
r
o
ac
h
to
p
r
e
d
ict
cr
o
p
y
ield
in
all
t
y
p
es
o
f
e
n
v
ir
o
n
m
e
n
t
s
.
Ou
r
s
tu
d
y
ai
m
s
at
f
i
n
d
in
g
th
e
m
o
s
t
r
elev
a
n
t
an
d
co
m
m
o
n
f
ea
tu
r
es
f
o
r
cr
o
p
y
ield
p
r
ed
ictio
n
in
v
ar
io
u
s
en
v
ir
o
n
m
e
n
t
s
w
it
h
s
tate
o
f
t
h
e
ar
t
tech
n
iq
u
e
s
an
d
d
ata
s
o
u
r
ce
s
to
g
iv
e
a
b
etter
an
d
clea
r
ed
id
ea
to
r
esear
ch
er
s
to
s
tar
t
w
i
th
cr
o
p
y
ield
p
r
ed
i
ctio
n
w
it
h
th
e
u
p
d
ated
tech
n
i
q
u
es
th
at
ca
n
b
e
b
e
ap
p
lied
o
v
er
m
o
s
t
cr
o
p
s
an
d
en
v
ir
o
n
m
e
n
t
s
.
T
ab
le
1
s
u
m
m
ar
izes
th
e
g
ap
s
i
n
t
h
e
co
n
s
id
er
ed
s
tu
d
ies
f
o
r
co
m
p
ar
is
o
n
w
it
h
o
u
t
s
t
u
d
y
.
‘
Y’
r
ep
r
esen
ts
YE
S a
n
d
‘
N
’
r
ep
r
esen
ts
NO
.
T
ab
le
1
.
T
est
m
o
d
el
s
p
ec
if
ica
t
io
n
s
an
d
te
s
t c
o
n
d
itio
n
s
C
o
mp
a
r
i
so
n
po
i
n
t
s
[
6
]
[
7
]
[
8
]
[
9
]
[
1
0
]
O
u
r
r
e
v
i
e
w
S
t
a
t
e
o
f
t
h
e
a
r
t
t
e
c
h
n
i
q
u
e
s
d
i
s
c
u
sse
d
Y
Y
Y
Y
N
Y
C
o
mp
a
r
i
so
n
b
e
t
w
e
e
n
M
L
a
n
d
DL
N
N
N
N
N
Y
Ev
a
l
u
a
t
i
o
n
me
t
r
i
c
N
Y
N
Y
N
Y
D
a
t
a
so
u
r
c
e
s
N
N
N
Y
N
Y
En
se
mb
l
e
d
v
s
c
l
a
ssi
c
M
L
,
D
L
mo
d
e
l
s
N
N
N
N
N
Y
M
o
st
u
se
d
f
e
a
t
u
r
e
s
N
Y
Y
Y
Y
Y
L
i
mi
t
a
t
i
o
n
s a
n
d
f
u
t
u
r
e
w
o
r
k
Y
Y
Y
Y
N
Y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
2
3
,
No
.
2
,
A
p
r
il
20
2
5
:
4
0
2
-
415
404
2.
M
E
T
H
O
D
T
h
is
r
ev
ie
w
ad
d
r
ess
es
u
n
a
n
s
w
er
ed
q
u
esti
o
n
s
o
u
tli
n
ed
in
T
ab
le
1
an
d
u
p
d
ates
th
e
ex
is
ti
n
g
liter
atu
r
e
w
it
h
th
e
late
s
t
f
i
n
d
in
g
s
.
I
t
ex
a
m
i
n
es
th
e
co
m
p
ar
is
o
n
b
et
w
e
en
en
s
e
m
b
le
an
d
tr
ad
itio
n
al
ML
/D
L
m
o
d
els,
as
w
ell
a
s
b
et
w
ee
n
M
L
an
d
DL
m
o
d
els
s
p
ec
i
f
icall
y
f
o
r
cr
o
p
y
ield
p
r
ed
ictio
n
,
ar
ea
s
n
o
t
clea
r
l
y
co
v
er
ed
in
p
r
ev
io
u
s
r
ev
ie
w
s
.
B
y
s
y
n
t
h
es
i
zin
g
r
ec
en
t
f
i
n
d
in
g
s
,
t
h
i
s
r
ev
ie
w
p
r
o
v
id
es r
esear
c
h
er
s
w
i
th
u
p
d
ated
in
s
ig
h
t
s
in
to
b
est
p
r
ac
tices,
d
ata
s
o
u
r
ce
s
,
a
n
d
m
eth
o
d
o
lo
g
ie
s
in
th
e
f
ield
.
T
h
is
co
n
tr
ib
u
tio
n
ai
m
s
to
s
u
p
p
o
r
t
r
esear
ch
er
s
in
b
u
ild
in
g
u
p
o
n
c
u
r
r
en
t
w
o
r
k
a
n
d
ad
v
an
cin
g
f
u
t
u
r
e
r
esear
ch
i
n
cr
o
p
y
ield
p
r
ed
ictio
n
.
2
.
1
.
L
it
er
a
t
ure
re
v
iew
A
d
etailed
s
y
s
te
m
atic
r
ev
ie
w
is
ca
r
r
ied
o
u
t
in
th
i
s
s
t
u
d
y
to
an
s
w
er
o
u
r
s
p
ec
i
f
ic
r
esear
ch
q
u
esti
o
n
s
.
T
h
is
in
clu
d
es
t
h
e
s
elec
tio
n
cr
ite
r
ia
o
f
all
th
e
liter
atu
r
e
in
cl
u
d
ed
an
d
r
ev
ie
w
ed
in
th
i
s
s
t
u
d
y
.
T
h
e
s
elec
tio
n
o
f
all
th
e
liter
atu
r
e
w
as
d
o
n
e
u
s
i
n
g
a
b
u
n
ch
o
f
r
ele
v
an
t
k
e
y
w
o
r
d
s
f
o
r
o
u
r
s
t
u
d
y
.
T
h
e
liter
at
u
r
e
in
cl
u
d
ed
in
t
h
is
r
ev
ie
w
w
er
e
p
u
lled
f
r
o
m
Go
o
g
le
Sch
o
lar
in
a
y
ea
r
-
w
i
s
e
m
an
n
er
.
T
h
e
k
e
y
w
o
r
d
s
u
s
ed
ar
e
m
e
n
tio
n
ed
i
n
T
ab
le
2
.
T
h
e
y
ea
r
l
y
f
il
ter
w
as
u
s
ed
o
n
Go
o
g
le
Sch
o
lar
to
d
o
w
n
lo
ad
p
ap
er
s
th
at
w
er
e
o
f
r
e
lev
an
ce
.
Af
ter
t
h
is
,
ea
ch
p
ap
er
w
a
s
r
ev
ie
w
ed
f
o
r
r
elev
an
ce
b
ased
o
n
t
h
e
ab
s
tr
ac
t,
in
tr
o
d
u
ctio
n
an
d
tec
h
n
o
lo
g
ie
s
u
s
ed
.
T
h
e
p
ap
er
s
f
u
r
t
h
er
d
is
ca
r
d
ed
w
er
e
d
u
e
to
th
e
r
ea
s
o
n
s
t
h
at
t
h
e
y
w
er
e
a
s
s
o
ciate
d
w
i
th
p
la
n
t
d
is
ea
s
e
d
etec
tio
n
,
tr
ad
itio
n
al
p
h
en
o
lo
g
y
w
i
th
o
u
t
u
s
in
g
ML
o
r
DL
,
s
p
ec
if
ic
to
d
ata
m
i
n
i
n
g
,
in
ter
n
et
o
f
th
i
n
g
(
I
o
T
)
an
d
s
o
il
m
an
a
g
e
m
en
t.
Fin
all
y
,
w
e
w
er
e
lef
t
w
it
h
8
0
q
u
alit
y
l
iter
atu
r
e
to
r
ev
ie
w
i
n
t
h
is
s
tu
d
y
w
h
ich
w
a
s
in
cl
u
d
ed
.
T
ab
le
2
.
T
h
e
k
e
y
w
o
r
d
s
u
s
ed
t
o
s
ea
r
ch
p
ap
er
s
S
r
.
N
o
.
K
e
y
w
o
r
d
s
1
C
r
o
p
y
i
e
l
d
p
r
e
d
i
c
t
i
o
n
2
C
r
o
p
y
i
e
l
d
p
r
e
d
i
c
t
i
o
n
M
L
3
C
r
o
p
y
i
e
l
d
p
r
e
d
i
c
t
i
o
n
D
L
4
C
r
o
p
y
i
e
l
d
p
r
e
d
i
c
t
i
o
n
D
L
r
e
mo
t
e
se
n
si
n
g
5
C
r
o
p
y
i
e
l
d
p
r
e
d
i
c
t
i
o
n
M
L
r
e
m
o
t
e
se
n
si
n
g
Mo
r
e
p
r
ec
is
ely
as
s
h
o
w
n
in
Fig
u
r
e
1
,
th
e
to
tal
n
u
m
b
er
o
f
d
o
w
n
lo
ad
ed
p
ap
er
s
w
a
s
2
3
8
.
Fro
m
t
h
e
d
o
w
n
lo
ad
ed
p
ap
er
s
1
9
8
p
ap
e
r
s
w
er
e
s
elec
ted
b
ased
o
n
th
e
ti
tle
an
d
f
u
r
t
h
er
1
6
5
p
ap
er
s
w
er
e
s
elec
ted
f
r
o
m
t
h
e
th
en
s
elec
ted
p
ap
er
s
b
ased
o
n
th
e
ab
s
tr
ac
t o
f
th
e
p
ap
er
.
Am
o
n
g
th
e
1
6
5
p
ap
er
s
f
in
al
l
y
8
0
p
ap
er
s
w
er
e
s
elec
ted
th
at
ar
e
u
s
ed
f
o
r
o
u
r
s
tu
d
y
.
T
h
e
F
ig
u
r
e
1
g
iv
e
s
an
i
n
s
i
g
h
t o
n
th
e
s
elec
t
io
n
cr
iter
ia.
Fig
u
r
e
1
.
P
ap
er
s
elec
tio
n
cr
iter
ia
As
s
h
o
w
n
i
n
Fig
u
r
e
2
th
e
s
el
ec
ted
p
ap
er
s
r
an
g
e
f
r
o
m
th
e
y
ea
r
2
0
1
4
to
2
0
2
4
.
T
h
e
in
cr
em
en
t
in
t
h
e
p
ap
er
s
ca
n
b
e
s
ee
n
f
r
o
m
2
0
1
9
.
T
h
e
r
esear
ch
er
’
s
i
n
ter
es
t
i
n
th
e
d
o
m
a
in
h
a
s
g
r
o
w
n
w
i
th
t
h
e
ad
v
a
n
ce
m
en
t
i
n
s
atellite
tec
h
n
o
lo
g
y
a
n
d
en
h
an
ce
d
co
m
p
u
ta
tio
n
.
Fig
u
r
e
2
.
Yea
r
w
is
e
p
ap
er
d
is
tr
ib
u
tio
n
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
A
d
va
n
ce
d
cro
p
yield
p
r
ed
ictio
n
u
s
in
g
m
a
ch
in
e
lea
r
n
in
g
a
n
d
d
ee
p
le
a
r
n
in
g
:
…
(
A
yu
s
h
A
n
a
n
d
)
405
2
.
2
.
Alg
o
rit
h
m
s
T
h
er
e
ar
e
s
ev
er
al
alg
o
r
ith
m
s
av
ailab
le
f
o
r
m
a
k
i
n
g
cr
o
p
y
ie
ld
p
r
ed
ictio
n
,
an
d
th
e
s
elec
tio
n
o
f
m
o
s
t
s
u
itab
le
o
n
e
d
ep
en
d
s
o
n
m
u
lti
p
le
f
ac
to
r
s
,
in
cl
u
d
in
g
th
e
t
y
p
e
o
f
d
ata
av
ailab
le,
th
e
n
u
m
b
er
o
f
f
ea
tu
r
e
s
in
t
h
e
d
ataset,
an
d
t
h
e
n
a
tu
r
e
o
f
t
h
e
d
ata
—
w
h
et
h
er
it
is
s
tati
s
tical,
i
m
a
g
e
-
b
ased
,
o
r
a
co
m
b
in
at
io
n
o
f
b
o
th
.
A
d
d
itio
n
al
l
y
,
u
n
d
er
s
tan
d
i
n
g
t
h
e
lin
ea
r
it
y
o
r
n
o
n
-
lin
ea
r
it
y
o
f
t
h
e
d
ata
p
la
y
s
a
cr
u
cial
r
o
le
i
n
d
eter
m
in
i
n
g
w
h
ic
h
alg
o
r
ith
m
w
il
l
p
er
f
o
r
m
b
est
.
C
r
o
p
y
ield
p
r
ed
ictio
n
is
a
c
o
m
p
le
x
is
s
u
e
in
v
o
lv
in
g
v
ar
io
u
s
f
ac
to
r
s
s
u
ch
a
s
c
li
m
ate
co
n
d
itio
n
s
,
s
o
il
p
r
o
p
er
ties
,
r
ain
f
all,
te
m
p
er
at
u
r
e,
h
u
m
id
it
y
,
f
er
tili
ze
r
u
s
a
g
e,
an
d
cr
o
p
v
ar
iet
y
.
T
h
e
ac
cu
r
ac
y
o
f
p
r
ed
ictio
n
s
d
ep
en
d
s
o
n
s
elec
tin
g
an
ap
p
r
o
p
r
iate
ML
o
r
DL
m
o
d
el
th
at
ca
n
ef
f
ec
tiv
el
y
ca
p
t
u
r
e
th
e
r
elatio
n
s
h
ip
s
b
et
w
ee
n
t
h
ese
f
a
cto
r
s
.
2
.
2
.
1
.
Su
pp
o
rt
v
ec
t
o
r
m
a
chi
nes
T
h
ese
alg
o
r
ith
m
s
f
in
d
a
h
y
p
er
p
lan
e
in
th
e
f
ea
t
u
r
e
s
p
ac
e
th
at
b
est
s
ep
ar
ates
th
e
d
ata
p
o
in
ts
b
elo
n
g
i
n
g
to
d
if
f
er
e
n
t c
la
s
s
e
s
.
S
u
p
p
o
r
t v
ec
to
r
m
ac
h
i
n
e
(
SVM
)
f
o
cu
s
o
n
id
en
ti
f
y
i
n
g
a
s
m
all
s
u
b
s
et
o
f
tr
ain
i
n
g
d
ata
p
o
in
t
s
(
s
u
p
p
o
r
t
v
ec
to
r
s
)
th
at
d
ef
in
e
th
e
h
y
p
er
p
la
n
e
'
s
m
ar
g
i
n
s
.
T
h
is
ap
p
r
o
ac
h
m
ak
e
s
th
e
m
r
o
b
u
s
t
to
o
u
tlier
s
an
d
ef
f
icien
t
f
o
r
h
ig
h
-
d
i
m
e
n
s
io
n
al
d
ata.
SVM
h
as
g
i
v
e
n
p
r
o
m
is
i
n
g
r
e
s
u
l
ts
i
n
m
an
y
s
tu
d
ies
s
u
c
h
as
t
h
i
s
s
t
u
d
y
[
1
1
]
th
at
p
r
ed
icts
p
o
tato
y
ield
u
s
i
n
g
Sen
ti
n
el
2
d
ata
in
Seg
o
v
ia,
Sp
ain
w
ith
a
h
i
g
h
r
^2
v
alu
e
o
f
0
.
9
3
.
A
n
o
th
er
s
tu
d
y
[
1
2
]
c
o
n
d
u
cted
in
T
am
il
Nad
u
,
I
n
d
ia,
co
m
p
ar
ed
d
if
f
er
en
t
f
ea
tu
r
e
s
u
b
s
ets
f
o
r
cr
o
p
y
ield
p
r
ed
ictio
n
an
d
also
s
h
o
w
ed
th
at
SV
M
h
ad
a
h
ig
h
R
Sco
r
e
o
f
0
.
9
2
.
A
r
esear
ch
w
o
r
k
[
1
3
]
w
h
ic
h
also
u
s
ed
L
an
d
s
at
-
8
d
ata
s
h
o
w
ed
th
at
SVM
h
ad
ac
h
iev
ed
a
h
i
g
h
ac
c
u
r
ac
y
o
f
9
8
.
7
2
%.
A
r
ec
en
t
s
t
u
d
y
[
1
4
]
aim
ed
at
co
m
p
ar
i
n
g
v
ar
io
u
s
ML
m
o
d
el
s
f
o
r
So
y
b
ea
n
y
ield
p
r
ed
ictio
n
u
s
in
g
r
e
m
o
te
s
e
n
s
i
n
g
an
d
w
ea
th
er
d
ata
also
s
h
o
w
e
d
th
at
SVM
h
ad
a
d
ec
en
t R^2
s
co
r
e
o
f
0
.
7
2
2
.
A
s
tu
d
y
[
1
5
]
w
as
ea
c
h
o
n
t
h
e
p
r
ed
ictio
n
o
f
W
i
n
ter
W
h
ea
t o
n
m
u
lti
-
s
o
u
r
ce
d
d
ata
in
C
h
i
n
a
a
n
d
also
s
h
o
w
ed
t
h
at
S
VM
w
as
a
m
o
n
g
o
n
e
o
f
th
e
h
i
g
h
e
s
t
ac
c
u
r
ate
al
g
o
r
ith
m
s
f
o
r
m
ak
in
g
p
r
ed
ictio
n
s
.
T
h
is
b
ein
g
s
aid
SVM
th
o
u
g
h
n
o
t
th
e
b
est
in
all
ca
s
es
g
i
v
es
p
r
o
m
i
s
in
g
r
esu
l
ts
b
ec
au
s
e
o
f
it
s
in
ab
ilit
y
to
h
a
n
d
le
n
o
n
li
n
ea
r
an
d
v
er
y
lar
g
e
d
ata
s
ets.
SVM
i
s
also
co
m
p
u
tatio
n
all
y
v
er
y
ex
p
e
n
s
i
v
e.
2
.
2
.
2
.
Ra
nd
o
m
f
o
r
ests
I
t
is
b
u
ild
u
p
o
n
d
ec
is
io
n
tr
ee
s
b
y
cr
ea
ti
n
g
a
n
e
n
s
e
m
b
le
o
f
th
e
m
.
E
ac
h
tr
ee
is
tr
ain
ed
o
n
a
r
an
d
o
m
s
u
b
s
et
o
f
f
ea
t
u
r
es
a
n
d
d
ata
p
o
i
n
ts
,
e
n
h
an
c
in
g
ac
c
u
r
ac
y
a
n
d
r
ed
u
cin
g
o
v
er
f
itt
in
g
.
P
r
ed
ictio
n
s
f
r
o
m
all
tr
ee
s
ar
e
th
en
ag
g
r
eg
ated
f
o
r
a
f
i
n
al
o
u
t
p
u
t.
A
s
t
u
d
y
[
1
6
]
d
o
n
e
o
n
s
o
y
b
ea
n
an
d
co
r
n
d
ataset
s
4
ti
m
e
s
a
y
ea
r
f
o
r
3
y
ea
r
s
as
test
d
ata
co
n
cl
u
d
ed
th
at
r
a
n
d
o
m
f
o
r
est
g
iv
e
s
g
o
o
d
ac
cu
r
ac
y
w
it
h
R
M
SE
o
b
s
er
v
ed
5
.
6
2
b
u
s
h
els
p
er
ac
r
e
f
o
r
th
e
s
o
y
b
ea
n
d
ataset
f
o
r
A
u
g
u
s
t
2
0
1
7
.
Si
m
ilar
l
y
,
a
n
o
th
er
r
esear
ch
[
1
7
]
th
at
w
as
d
o
n
e
in
th
e
m
ai
n
w
h
ea
t
-
p
r
o
d
u
cin
g
r
eg
io
n
o
f
C
h
i
n
a
u
s
in
g
d
ata
f
r
o
m
v
ar
io
u
s
s
o
u
r
ce
s
s
u
ch
as
r
e
m
o
te
s
en
s
i
n
g
m
ete
o
r
o
lo
g
ical
d
ata
etc
also
r
ec
eiv
ed
th
e
s
ec
o
n
d
h
ig
h
est
R
^2
s
co
r
e
am
o
n
g
all
alg
o
r
i
th
m
s
s
t
u
d
ied
.
T
h
e
r
^
2
o
f
r
an
d
o
m
f
o
r
est
wa
s
0
.
7
2
.
A
s
tu
d
y
[
1
8
]
p
er
f
o
r
m
ed
o
n
wh
ea
t
cr
o
p
s
ai
m
ed
to
co
m
p
ar
e
r
an
d
o
m
f
o
r
est
a
n
d
th
r
ee
d
i
f
f
e
r
en
t
D
L
alg
o
r
it
h
m
s
an
d
co
n
clu
d
ed
t
h
at
r
an
d
o
m
f
o
r
est
h
ad
th
e
b
est
R
^2
s
co
r
e
o
f
0
.
8
9
.
A
n
o
th
er
r
esear
c
h
[
1
4
]
also
co
n
d
u
cted
o
n
s
o
y
b
ea
n
s
also
co
n
cl
u
d
ed
th
at
r
an
d
o
m
f
o
r
est
a
f
ter
t
u
n
i
n
g
it
s
h
y
p
er
p
ar
a
m
eter
s
s
p
ec
if
icall
y
g
av
e
a
p
r
o
m
is
in
g
an
d
th
e
h
i
g
h
e
s
t
R
^2
o
f
0
.
7
4
8
in
t
h
e
p
ar
ticu
lar
s
t
u
d
y
o
u
tp
er
f
o
r
m
in
g
s
u
p
p
o
r
t
v
ec
to
r
r
eg
r
ess
i
o
n
(
SVR
)
.
An
o
th
er
s
tu
d
y
[
1
5
]
th
at
w
as
d
o
n
e
to
p
r
ed
ict
w
i
n
ter
w
h
ea
t
y
ie
ld
co
n
clu
d
ed
th
at
r
an
d
o
m
f
o
r
est
d
e
m
o
n
s
tr
ates
t
h
e
b
es
t
g
en
er
aliza
tio
n
ab
ilit
y
a
m
o
n
g
o
th
er
p
o
p
u
lar
alg
o
r
ith
m
s
u
s
ed
.
A
ls
o
,
ac
co
r
d
in
g
to
Sar
r
an
d
Su
lta
n
[
1
9
]
r
an
d
o
m
f
o
r
est
p
er
f
o
r
m
ed
t
h
e
b
est
i
n
p
r
ed
ictin
g
Ma
ize
y
ield
w
i
th
a
n
R
^2
v
al
u
e
o
f
0
.
6
4
.
T
h
e
s
tu
d
y
a
ls
o
h
ad
p
ea
n
u
t
,
m
illet a
n
d
s
o
r
g
h
u
m
d
a
taset
s
in
w
h
ic
h
r
an
d
o
m
f
o
r
est als
o
p
er
f
o
r
m
ed
w
ell
a
n
d
w
a
s
b
eh
i
n
d
b
y
a
s
m
all
m
ar
g
i
n
.
2
.
2
.
3
.
Art
if
icia
l
neura
l net
wo
rks
T
h
ey
co
n
s
is
t
o
f
i
n
ter
co
n
n
ec
te
d
lay
er
s
o
f
p
r
o
ce
s
s
i
n
g
u
n
i
ts
(
n
eu
r
o
n
s
)
th
at
lear
n
p
atter
n
s
f
r
o
m
d
ata
th
r
o
u
g
h
an
iter
ati
v
e
p
r
o
ce
s
s
ca
lled
b
ac
k
p
r
o
p
ag
atio
n
.
Stre
n
g
t
h
s
in
c
lu
d
e
tack
l
in
g
co
m
p
lex
,
n
o
n
-
li
n
ea
r
p
r
o
b
lem
s
an
d
ex
ce
lli
n
g
at
f
ea
t
u
r
e
ex
tr
ac
tio
n
.
A
cc
o
r
d
in
g
to
a
s
tu
d
y
[
2
0
]
p
o
tato
y
ield
in
B
an
g
la
d
esh
u
s
i
n
g
r
e
m
o
te
s
en
s
in
g
s
a
te
llit
es
u
s
in
g
ar
tific
ial
n
eu
r
al
n
et
w
o
r
k
(
A
NN
)
a
n
d
th
e
er
r
o
r
o
f
p
r
e
d
ictio
n
w
as
v
er
y
s
m
all
a
n
d
less
th
an
1
0
%.
W
h
ic
h
i
n
d
icate
d
th
at
A
NN
i
s
h
ig
h
l
y
ac
c
u
r
ate
i
n
p
r
ed
ictin
g
p
o
tato
y
i
e
ld
in
B
a
n
g
lad
es
h
.
T
h
e
s
t
u
d
y
[
2
1
]
w
h
ic
h
w
as
d
o
n
e
o
n
R
ice
cr
o
p
s
in
th
e
I
n
d
ian
r
eg
io
n
f
o
u
n
d
th
a
t
th
e
ac
cu
r
ac
y
o
f
A
N
N
w
as
9
7
.
5
in
th
is
s
tu
d
y
w
it
h
a
s
e
n
s
iti
v
it
y
o
f
9
6
.
3
.
A
n
o
th
er
r
esear
c
h
[
1
9
]
w
as
p
er
f
o
r
m
ed
in
Sen
e
g
al
lo
ca
ted
in
t
h
e
Af
r
ican
co
n
tin
e
n
t
an
d
s
tat
is
tica
l
an
d
s
a
tellite
d
ata
w
er
e
u
s
ed
A
NN
o
u
tp
er
f
o
r
m
ed
all
o
th
er
m
o
d
els
in
p
r
ed
ictin
g
P
ea
n
u
t
an
d
So
r
g
h
u
m
y
ield
w
ith
a
n
R
^2
s
co
r
e
o
f
0
.
6
6
an
d
0
.
5
7
r
esp
ec
tiv
el
y
.
Si
m
ilar
l
y
[
2
2
]
w
as
ca
r
r
ied
to
p
r
e
d
ict
Mu
s
tar
d
cr
o
p
y
ield
w
h
ich
co
n
clu
d
ed
t
h
at
A
NN
h
ad
a
n
ac
cu
r
ac
y
o
f
9
9
.
9
4
%,
p
r
ec
is
io
n
o
f
9
9
.
9
4
%
an
d
an
F
-
Sco
r
e
o
f
0
.
9
9
7
6
.
A
n
o
t
h
er
s
t
u
d
y
[
1
2
]
w
h
ic
h
w
as
ex
ec
u
ted
to
p
r
ed
ict
p
ad
d
y
cr
o
p
s
in
th
e
s
tate
o
f
T
am
il
Nad
u
in
I
n
d
ia
w
h
ich
ac
h
iev
ed
an
R
^
2
s
co
r
e
o
f
0
.
9
2
f
o
r
A
NN.
Si
m
i
lar
l
y
[
2
3
]
aim
i
n
g
to
p
r
e
d
ict
r
ic
e
p
r
o
d
u
ce
d
a
h
ig
h
test
i
n
g
R
^2
s
co
r
e
o
f
0
.
9
7
8
.
2
.
2
.
4
.
E
x
t
re
m
e
g
ra
dient
bo
o
s
t
ing
I
t
lev
er
ag
es
e
n
s
e
m
b
le
lear
n
i
n
g
w
it
h
g
r
ad
ien
t
b
o
o
s
tin
g
.
Ne
w
m
o
d
el
s
ar
e
s
eq
u
en
t
iall
y
ad
d
ed
to
co
r
r
ec
t
th
e
er
r
o
r
s
o
f
p
r
ev
io
u
s
m
o
d
el
s
,
f
o
cu
s
i
n
g
o
n
m
i
n
i
m
is
in
g
t
h
e
l
o
s
s
f
u
n
ctio
n
w
h
ile
co
n
tr
o
llin
g
m
o
d
el
co
m
p
le
x
it
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
2
3
,
No
.
2
,
A
p
r
il
20
2
5
:
4
0
2
-
415
406
to
p
r
ev
en
t
o
v
er
f
it
tin
g
.
A
cc
o
r
d
in
g
to
r
esear
ch
[
2
4
]
ex
tr
e
m
e
g
r
ad
ien
t
b
o
o
s
ti
n
g
(
XGB
o
o
s
t
)
w
a
s
t
h
e
b
est
-
p
er
f
o
r
m
in
g
al
g
o
r
ith
m
f
o
r
cr
o
p
p
r
ed
ictio
n
w
it
h
th
e
h
i
g
h
e
s
t
R
^2
s
co
r
e
o
f
8
4
.
7
9
,
w
h
ich
o
u
tp
er
f
o
r
m
ed
all
o
th
er
alg
o
r
ith
m
s
i
n
t
h
e
s
t
u
d
y
.
T
h
e
h
ig
h
e
s
t
R
^2
s
co
r
e
b
y
XGB
o
o
s
t
w
as
0
.
9
2
in
th
e
m
o
n
t
h
o
f
A
p
r
il
2
0
2
2
.
A
s
tu
d
y
[
2
5
]
aim
ed
at
p
r
ed
ic
tin
g
m
a
iz
e
y
ield
an
d
Nitr
o
g
e
n
lo
s
s
f
r
o
m
s
o
il
u
s
i
n
g
d
ata
f
r
o
m
s
e
v
en
lo
ca
tio
n
s
in
t
h
e
US
Mid
w
e
s
t
o
v
er
5
-
7
y
ea
r
s
a
n
d
co
n
clu
d
ed
th
at
X
GB
o
o
s
t
h
ad
3
rd
h
ig
h
e
s
t
R
^2
s
co
r
e
a
m
o
n
g
all
alg
o
r
ith
m
s
u
s
ed
i
n
th
e
s
tu
d
y
.
Ho
w
ev
er
,
XGB
o
o
s
t
h
ad
th
e
h
i
g
h
est
R
-
R
MSE
i
n
p
r
ed
ictin
g
N
-
L
o
s
s
at
9
8
.
3
%.
Si
m
ilar
l
y
[
2
6
]
ai
m
ed
at
p
r
ed
ictin
g
cr
o
p
y
ield
u
s
i
n
g
m
eteo
r
o
lo
g
ical
d
ata
a
n
d
r
e
m
o
t
e
s
en
s
i
n
g
d
ata
f
r
o
m
m
o
d
er
ate
r
eso
lu
tio
n
i
m
a
g
i
n
g
s
p
ec
tr
o
r
ad
io
m
eter
(
MO
DI
S
)
a
n
d
it
w
a
s
co
n
c
lu
d
ed
th
at
XG
B
o
o
s
t
h
ad
th
e
b
est
ac
cu
r
ac
y
i
n
th
e
s
t
u
d
y
w
it
h
a
n
R
^2
s
co
r
e
o
f
0
.
8
4
5
.
Ho
w
e
v
er
,
a
r
esear
ch
[
2
7
]
aim
ed
at
p
r
e
d
ictin
g
co
r
n
y
ield
i
n
US
A
co
u
n
t
y
-
w
i
s
e
f
r
o
m
th
e
y
ea
r
2
0
0
0
-
2
0
1
8
XGB
o
o
s
t
d
id
n
o
t
p
e
r
f
o
r
m
w
ell
as
it
h
ad
o
n
e
o
f
th
e
h
ig
h
est
R
M
SE
s
co
r
es
am
o
n
g
all
th
e
alg
o
r
ith
m
s
u
s
ed
.
Si
m
ilar
l
y
[
2
8
]
w
as
co
n
d
u
c
ted
o
n
n
in
e
f
ea
t
u
r
es
f
r
o
m
r
e
m
o
te
s
e
n
s
i
n
g
s
atellites
an
d
M
L
alg
o
r
ith
m
s
w
er
e
ap
p
lied
m
o
n
th
w
i
s
e
o
u
t
o
f
w
h
ic
h
XGB
o
o
s
t’
s
p
er
f
o
r
m
a
n
ce
w
a
s
n
o
t
o
u
t
s
tan
d
i
n
g
w
it
h
v
er
y
h
ig
h
R
^2
.
2
.
2
.
5
.
L
o
ng
s
ho
rt
-
t
er
m
m
e
mo
ry
I
t
is
a
s
p
ec
if
ic
ty
p
e
o
f
r
ec
u
r
r
en
t
n
e
u
r
al
n
et
w
o
r
k
(
R
NN)
,
th
at
ex
ce
ls
at
h
an
d
li
n
g
s
eq
u
e
n
tia
l
d
ata
(
tim
e
s
er
ies)
b
y
lear
n
i
n
g
lo
n
g
-
t
er
m
d
ep
en
d
en
cies.
L
ST
Ms
u
til
ize
m
e
m
o
r
y
ce
ll
s
w
ith
g
ates
to
co
n
tr
o
l
in
f
o
r
m
at
io
n
f
lo
w
,
allo
w
i
n
g
t
h
e
n
et
w
o
r
k
to
r
etain
r
ele
v
an
t
i
n
f
o
r
m
atio
n
f
o
r
ex
ten
d
ed
p
er
io
d
s
.
A
s
t
u
d
y
[
2
9
]
w
as
ai
m
ed
a
t
p
r
ed
ictin
g
w
in
ter
w
h
ea
t
y
ield
u
s
ed
a
B
ay
e
s
ian
o
p
ti
m
iza
tio
n
-
b
a
s
ed
L
ST
M
m
o
d
el
w
h
ic
h
co
n
clu
d
ed
th
at
th
e
p
r
o
p
o
s
ed
m
o
d
el
p
er
f
o
r
m
ed
t
h
e
b
est
co
m
p
ar
ed
to
all
o
t
h
e
r
m
o
d
els
in
th
e
s
t
u
d
y
w
i
th
a
R
^2
s
co
r
e
o
f
0
.
8
2
.
An
o
th
er
s
t
u
d
y
[
3
0
]
p
er
f
o
r
m
ed
o
v
er
th
e
r
eg
io
n
o
f
P
u
n
j
ab
,
I
n
d
ia
to
p
r
e
d
ict
w
h
ea
t
cr
o
p
s
s
h
o
w
ed
th
a
t
R
NN
w
i
th
L
ST
M
o
u
tp
er
f
o
r
m
ed
all
o
th
er
alg
o
r
ith
m
s
i
n
t
h
e
s
t
u
d
y
b
y
a
co
n
s
id
er
ab
le
m
ar
g
i
n
.
Si
m
ilar
l
y
,
th
e
s
t
u
d
y
[
3
1
]
w
a
s
ex
ec
u
ted
o
v
er
a
d
ataset
co
n
s
is
ti
n
g
o
f
m
eteo
r
o
lo
g
ical
d
ata
an
d
s
o
il
an
d
cr
o
p
d
ata
an
d
co
m
p
ar
ed
d
if
f
er
en
t
m
o
d
el
s
i
n
t
h
e
s
tu
d
y
.
L
ST
M
p
er
f
o
r
m
ed
th
e
b
est
a
m
o
n
g
all
o
th
er
m
o
d
els
w
it
h
a
m
ar
g
i
n
al
d
if
f
er
en
ce
w
i
th
an
ac
cu
r
ac
y
o
f
8
6
%
in
p
r
ed
ictin
g
y
ie
ld
.
An
o
th
er
s
t
u
d
y
[
3
2
]
also
s
h
o
w
ed
th
a
t
Stack
ed
L
ST
M
p
er
f
o
r
m
ed
th
e
b
est
o
u
t
o
f
a
ll
al
g
o
r
ith
m
s
co
n
s
id
er
ed
f
o
r
th
e
s
t
u
d
y
w
i
th
w
ea
t
h
er
v
ar
iab
les
a
n
d
h
ad
an
R
^2
s
c
o
r
e
o
f
~0
.
7
3
2
.
T
h
is
s
h
o
w
s
t
h
at
L
ST
M
ca
n
p
r
o
d
u
ce
p
r
o
m
is
i
n
g
r
es
u
lts
f
o
r
cr
o
p
y
ie
ld
p
r
ed
ictio
n
.
2
.
2
.
6
.
Co
nv
o
lutio
na
l
neura
l
net
w
o
rk
s
C
NNs
ar
e
s
p
ec
ialized
ANN
ar
ch
itect
u
r
es
d
esig
n
ed
f
o
r
p
r
o
ce
s
s
i
n
g
g
r
id
-
li
k
e
d
ata,
p
ar
ticu
lar
l
y
i
m
a
g
e
s
.
C
NNs
e
f
f
icien
tl
y
e
x
tr
ac
t
s
p
a
tial
f
ea
t
u
r
es
th
r
o
u
g
h
co
n
v
o
l
u
tio
n
al
la
y
er
s
w
i
th
lear
n
ab
le
f
ilter
s
a
n
d
p
o
o
lin
g
l
a
y
er
s
f
o
r
d
i
m
e
n
s
io
n
alit
y
r
ed
u
ctio
n
.
A
r
esear
c
h
[
3
3
]
ai
m
ed
to
m
a
k
e
cr
o
p
y
ie
ld
p
r
ed
ictio
n
u
s
i
n
g
C
NN
-
R
N
N
p
er
f
o
r
m
ed
w
ell
o
u
tp
er
f
o
r
m
i
n
g
all
o
th
er
al
g
o
r
ith
m
s
u
s
ed
i
n
th
e
s
t
u
d
y
.
S
i
m
il
ar
l
y
,
a
n
o
t
h
er
s
tu
d
y
[
3
4
]
ai
m
ed
at
p
r
ed
ictin
g
s
o
y
b
ea
n
y
ield
s
als
o
s
h
o
w
ed
C
NN
p
er
f
o
r
m
ed
well
w
h
e
n
co
m
b
i
n
ed
w
ith
L
ST
M
th
e
C
NN
-
L
ST
M
m
o
d
el
p
r
o
d
u
ce
d
th
e
b
est
R
MSE
co
m
p
ar
ed
to
all
o
th
er
m
o
d
els
in
th
e
s
tu
d
y
.
An
o
th
er
s
t
u
d
y
[
3
5
]
aim
ed
at
p
r
ed
ictin
g
cr
o
p
y
ield
u
s
i
n
g
s
a
tellite
i
m
a
g
es
i
n
th
e
US
s
h
o
w
ed
th
at
th
e
C
NN
-
L
ST
M
m
o
d
e
l
h
ad
an
R
^2
s
co
r
e
o
f
0
.
9
1
.
Sim
i
lar
l
y
[
3
6
]
au
th
o
r
s
also
s
h
o
w
ed
th
a
t
C
NN+
R
NN+
2
FF
NN
p
r
o
d
u
ce
d
th
e
h
ig
h
e
s
t
co
r
r
elatio
n
co
ef
f
icie
n
t
0
.
9
1
8
3
.
T
h
is
s
h
o
w
s
th
at
i
f
th
e
r
ig
h
t
d
ata
is
o
b
tain
ed
C
NN
in
co
m
b
i
n
atio
n
w
ith
o
th
er
n
et
w
o
r
k
s
ca
n
b
e
v
er
y
ac
cu
r
ate
i
n
p
r
ed
ictin
g
cr
o
p
y
ield
.
3.
E
VA
L
UA
T
I
O
N
M
E
T
RIC
S
3
.
1
.
Reg
re
s
s
io
n
m
et
rics
R
eg
r
es
s
io
n
i
s
a
s
u
p
er
v
i
s
ed
ML
tech
n
iq
u
e
th
at
is
u
s
ed
to
p
r
e
d
ict
co
n
tin
u
o
u
s
v
al
u
es.
I
t
p
lo
ts
a
b
est
-
f
it
lin
e
p
ass
i
n
g
th
r
o
u
g
h
t
h
e
d
ata.
C
r
o
p
y
ield
p
r
ed
ictio
n
is
t
y
p
icall
y
a
r
eg
r
ess
io
n
tas
k
,
w
h
er
e
m
o
d
els
p
r
ed
ict
co
n
tin
u
o
u
s
v
al
u
es
(
y
ield
in
to
n
s
p
er
h
ec
tar
e)
.
No
m
o
d
el
is
p
er
f
ec
t
an
d
th
er
e
is
al
w
a
y
s
a
s
co
p
e
o
f
s
o
m
e
er
r
o
r
.
R
eg
r
es
s
io
n
m
etr
ic
s
h
e
lp
in
e
v
alu
ati
n
g
th
e
m
o
d
els.
Her
e
ar
e
th
e
k
e
y
m
etr
ic
s
f
o
r
ev
a
lu
at
in
g
r
eg
r
ess
io
n
m
o
d
els
:
R
MSE
m
ea
s
u
r
es
t
h
e
a
v
er
ag
e
d
if
f
er
e
n
ce
b
et
w
ee
n
p
r
ed
icted
an
d
ac
tu
al
y
ield
v
al
u
es.
L
o
we
r
R
MSE
i
n
d
icate
s
b
etter
m
o
d
el
p
er
f
o
r
m
a
n
ce
.
MA
E
ca
lcu
la
tes
th
e
av
er
a
g
e
ab
s
o
lu
te
d
if
f
er
en
ce
b
et
w
ee
n
p
r
ed
icted
an
d
ac
tu
al
y
ield
v
al
u
es.
R
-
s
q
u
ar
ed
(
R
²)
m
etr
ic
r
ep
r
esen
ts
t
h
e
p
r
o
p
o
r
ti
o
n
o
f
v
ar
ia
n
ce
in
t
h
e
ac
tu
a
l
y
i
eld
d
ata
ex
p
lain
ed
b
y
th
e
m
o
d
el
'
s
p
r
ed
ictio
n
s
.
A
v
al
u
e
clo
s
er
to
1
in
d
icate
s
a
b
etter
f
it.
3
.
2
.
Cla
s
s
if
ica
t
io
n
m
et
ric
s
C
las
s
i
f
icatio
n
tas
k
s
ar
e
also
p
ar
t
o
f
s
u
p
er
v
is
ed
ML
an
d
ar
e
ty
p
icall
y
u
s
ed
f
o
r
ca
teg
o
r
is
in
g
d
ata
b
y
p
r
ed
ictin
g
its
co
r
r
ec
t
lab
el.
I
n
s
o
m
e
ca
s
es,
m
o
d
els
m
i
g
h
t
p
r
ed
ict
y
ie
ld
ca
teg
o
r
ies
(
lo
w
,
m
e
d
iu
m
,
h
ig
h
)
in
s
tead
o
f
co
n
ti
n
u
o
u
s
v
a
lu
e
s
.
Her
e
ar
e
th
e
r
elev
an
t
ev
al
u
atio
n
m
e
tr
ics
u
s
ed
to
ev
al
u
ate
clas
s
i
f
i
ca
tio
n
alg
o
r
it
h
m
s
:
A
cc
u
r
ac
y
m
etr
ic
s
i
m
p
l
y
m
ea
s
u
r
es
t
h
e
p
er
ce
n
ta
g
e
o
f
co
r
r
e
ctl
y
cla
s
s
i
f
ied
y
ield
ca
teg
o
r
ie
s
.
F1
-
s
co
r
e
m
etr
ic
co
n
s
id
er
s
b
o
th
p
r
ec
is
io
n
(
p
r
o
p
o
r
tio
n
o
f
tr
u
e
p
o
s
iti
v
es
a
m
o
n
g
p
r
ed
icted
p
o
s
itiv
e
s
)
an
d
r
ec
all
(
p
r
o
p
o
r
tio
n
o
f
tr
u
e
p
o
s
itiv
e
s
id
en
ti
f
ied
b
y
th
e
m
o
d
el)
.
An
F1
-
s
co
r
e
clo
s
er
to
1
in
d
icate
s
b
etter
m
o
d
el
p
er
f
o
r
m
a
n
ce
,
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
A
d
va
n
ce
d
cro
p
yield
p
r
ed
ictio
n
u
s
in
g
m
a
ch
in
e
lea
r
n
in
g
a
n
d
d
ee
p
le
a
r
n
in
g
:
…
(
A
yu
s
h
A
n
a
n
d
)
407
esp
ec
iall
y
f
o
r
i
m
b
ala
n
ce
d
d
ata
s
ets
w
h
er
e
s
o
m
e
y
ie
ld
ca
teg
o
r
ies
m
i
g
h
t
b
e
less
f
r
eq
u
e
n
t.
Am
o
n
g
th
e
8
0
p
a
p
er
s
co
n
s
id
er
ed
in
o
u
r
r
ev
ie
w
f
o
u
r
class
i
f
icat
io
n
m
etr
ics
w
er
e
u
s
ed
F
-
Sco
r
e,
r
ec
all,
p
r
ec
is
io
n
,
ac
cu
r
ac
y
a
to
tal
2
0
ti
m
e
s
an
d
r
eg
r
ess
io
n
m
etr
ic
s
w
er
e
u
s
ed
1
3
5
tim
e
s
in
to
ta
l
w
h
ich
co
n
s
is
ted
o
f
MSE
,
R
MSE
,
M
A
E
,
R
^2
,
MA
P
E
,
r
elativ
e
r
o
o
t
m
ea
n
s
q
u
ar
ed
er
r
o
r
(
R
R
MSE
)
,
an
d
R
S
co
r
e.
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
Fo
r
th
e
p
r
esen
t
r
ev
ie
w
,
w
e
h
av
e
co
n
s
id
er
ed
b
o
th
co
n
f
er
en
ce
an
d
j
o
u
r
n
al
ar
ticles
to
en
s
u
r
e
a
co
m
p
r
e
h
en
s
iv
e
an
a
l
y
s
is
o
f
t
h
e
ex
i
s
ti
n
g
liter
at
u
r
e
.
T
o
m
ai
n
tain
t
h
e
q
u
alit
y
a
n
d
r
elev
a
n
ce
o
f
th
e
r
ev
ie
w
,
w
e
ap
p
lied
a
r
ig
o
r
o
u
s
i
n
clu
s
io
n
a
n
d
ex
cl
u
s
io
n
cr
iter
ia
,
b
ased
o
n
p
ar
a
m
eter
s
s
u
c
h
as
r
ele
v
a
n
ce
o
f
s
tu
d
ies
to
t
h
e
th
e
m
e
,
an
d
alig
n
m
e
n
t
w
it
h
th
e
o
b
j
ec
tiv
es
o
f
o
u
r
s
tu
d
y
.
W
e
h
av
e
f
i
n
ali
s
ed
6
7
jo
u
r
n
al
ar
ticles
an
d
1
3
a
r
ticles
f
r
o
m
co
n
f
er
en
ce
s
i.e
.
,
8
4
% f
r
o
m
j
o
u
r
n
als a
n
d
1
6
% f
r
o
m
co
n
f
er
en
ce
s
as r
ep
r
esen
ted
i
n
Fi
g
u
r
e
3
.
Fig
u
r
e
3
.
J
o
u
r
n
al
v
s
co
n
f
er
en
c
e
p
ap
er
s
4
.
1
.
Resea
rc
h
q
ues
t
io
ns
4
.
1
.
1
.
RQ
1
:
w
ha
t
a
re
t
he
s
t
a
t
e
-
of
-
t
he
-
a
rt
t
ec
hn
iqu
es us
e
d
?
State
-
of
-
th
e
-
ar
t
tec
h
n
iq
u
e
s
ca
n
b
e
j
u
d
g
ed
b
ased
o
n
t
h
e
m
o
s
t
u
s
ed
tec
h
n
iq
u
es
i
n
r
ec
e
n
t
p
u
b
licatio
n
s
an
d
th
e
tec
h
n
iq
u
es
o
r
m
o
d
els
t
h
at
te
n
d
to
p
er
f
o
r
m
th
e
b
est
i
n
v
ar
io
u
s
s
t
u
d
ies
to
p
r
ed
ict
cr
o
p
y
ield
p
r
o
d
u
ctio
n
.
I
n
th
e
s
tu
d
y
[
3
7
]
co
n
clu
d
ed
th
at
r
an
d
o
m
f
o
r
e
s
t
r
eg
r
es
s
o
r
o
u
tp
er
f
o
r
m
ed
all
o
th
er
s
u
p
er
v
i
s
ed
lear
n
in
g
m
o
d
el
s
in
cl
u
d
ed
in
t
h
e
s
t
u
d
y
.
An
o
th
er
s
tu
d
y
[
3
8
]
th
at
w
as d
o
n
e
o
n
S
o
y
ab
ea
n
cr
o
p
co
n
clu
d
ed
th
at
t
h
e
b
est
-
p
er
f
o
r
m
i
n
g
m
o
d
el
w
a
s
R
NN
i
n
th
e
s
t
u
d
y
.
A
s
i
m
ilar
s
t
u
d
y
[
3
9
]
s
h
o
w
ed
th
e
r
eliab
ilit
y
o
f
m
ak
i
n
g
s
i
g
n
if
ica
n
t
p
r
ed
ictio
n
s
.
T
h
e
s
tu
d
y
[
3
1
]
also
co
m
p
ar
ed
v
ar
io
u
s
m
o
d
els a
n
d
co
n
cl
u
d
ed
th
at
L
ST
M
o
u
tp
er
f
o
r
m
ed
all
o
th
er
m
o
d
els i
n
t
h
e
s
tu
d
y
.
An
o
th
er
s
t
u
d
y
[
4
0
]
u
s
ed
an
o
p
tim
is
ed
L
ST
M
ap
p
r
o
a
ch
to
r
ec
eiv
e
a
h
ig
h
ac
cu
r
ac
y
in
p
r
ed
ictio
n
.
T
h
e
s
tu
d
y
[
4
1
]
ca
r
r
ied
in
r
eg
io
n
o
f
C
h
i
n
a
o
n
w
i
n
ter
w
h
ea
t
also
s
h
o
w
ed
th
at
L
ST
M
p
er
f
o
r
m
ed
th
e
b
est
a
m
o
n
g
all
o
th
er
m
o
d
els
in
cl
u
d
ed
in
th
e
s
tu
d
y
.
T
h
e
r
esear
ch
er
s
[
2
9
]
th
at
u
s
ed
B
as
s
ei
n
o
p
ti
m
izer
w
i
t
h
L
ST
M
p
er
f
o
r
m
ed
th
e
b
est
a
m
o
n
g
all
m
o
d
els.
C
o
n
s
id
er
in
g
th
e
d
is
c
u
s
s
ed
s
t
u
d
i
es
an
d
o
th
er
d
etailed
s
tu
d
ies
th
at
w
e
co
m
p
ar
ed
in
th
i
s
r
e
v
ie
w
,
w
e
ca
n
i
n
f
er
t
h
a
t
s
tate
o
f
ar
t
alg
o
r
ith
m
s
u
s
ed
ar
e
en
s
e
m
b
led
tr
ee
m
o
d
els
lik
e
d
ec
is
io
n
tr
ee
,
r
an
d
o
m
f
o
r
est,
XGB
o
o
s
t
an
d
L
ST
M,
R
NN
,
an
d
C
NN
ar
e
also
u
s
ed
th
at
u
s
u
all
y
ten
d
to
p
er
f
o
r
m
b
etter
co
m
p
ar
ed
to
th
e
clas
s
ic
M
L
m
o
d
els.
Fig
u
r
e
4
s
h
o
w
s
t
h
e
n
u
m
b
er
o
f
ti
m
e
s
th
e
s
e
alg
o
r
it
h
m
s
wer
e
u
s
ed
in
all
t
h
e
co
n
s
id
er
e
d
s
tu
d
ies.
r
an
d
o
m
f
o
r
est
e
m
er
g
ed
as
t
h
e
m
o
s
t
f
r
eq
u
en
tl
y
e
m
p
lo
y
ed
alg
o
r
ith
m
ac
r
o
s
s
3
4
s
t
u
d
ies.
Fo
llo
w
i
n
g
clo
s
el
y
b
eh
in
d
w
er
e
SVM
,
A
NN,
L
S
T
M,
least
ab
s
o
lu
te
s
h
r
i
n
k
a
g
e
an
d
s
elec
tio
n
o
p
er
ato
r
(
L
A
S
SO
)
,
d
ec
is
io
n
tr
ee
r
eg
r
ess
io
n
,
lin
ea
r
r
eg
r
ess
io
n
,
C
NN,
an
d
g
r
ad
ien
t
b
o
o
s
tin
g
r
eg
r
ess
io
n
,
w
i
th
u
s
ag
e
co
u
n
ts
o
f
2
0
,
1
6
,
1
4
,
1
4
,
1
3
,
1
3
,
1
0
,
an
d
1
0
s
tu
d
ies,
r
esp
ec
tiv
el
y
.
T
h
e
d
iv
er
s
it
y
o
f
alg
o
r
i
th
m
s
i
n
d
icate
s
th
e
v
ar
iet
y
o
f
task
s
p
er
f
o
r
m
ed
i
n
th
e
r
esear
ch
.
T
h
e
o
t
h
er
s
in
th
e
b
elo
w
f
i
g
u
r
e
co
n
s
is
ted
o
f
B
a
y
esia
n
R
id
g
e,
R
-
C
NN,
A
C
N
N
-
B
D
L
ST
M,
lig
h
t
u
s
e
e
f
f
ic
ien
c
y
(
LU
E)
,
b
a
y
es
i
an
r
id
g
e
(
BR
)
,
Hu
b
er
g
r
eg
r
es
s
io
n
,
lo
n
g
s
h
o
r
t
-
ter
m
m
e
m
o
r
y
-
Gau
s
s
ia
n
p
r
o
ce
s
s
(
L
ST
M
-
GP
)
,
w
a
v
elet
co
n
v
o
l
u
tio
n
al
n
eu
r
al
n
et
w
o
r
k
(
W
-
C
N
N
)
,
r
ad
ial
b
asis
f
u
n
c
tio
n
n
e
u
r
al
n
et
w
o
r
k
(
R
B
F
-
NN
)
,
ca
t
b
o
o
s
t
r
eg
r
ess
io
n
,
R
es
-
NE
T
2
D,
3
D
,
A
B
R
,
DC
NN,
m
u
ltip
le
lo
g
i
s
tic
r
eg
r
ess
io
n
,
te
m
p
o
r
al
co
n
v
o
lu
tio
n
al
n
e
u
r
al
n
et
w
o
r
k
(
T
C
NN
)
,
C
NN
-
R
N
N,
co
n
v
o
l
u
tio
n
al
n
eu
r
al
n
et
w
o
r
k
-
Ga
u
s
s
ian
pr
o
ce
s
s
(
C
NN
-
GP
)
,
m
u
lti
-
v
ie
w
g
ated
F
u
s
io
n
(
M
V
GF
)
,
an
d
ad
ap
tiv
e
b
o
o
s
tin
g
(
A
D
A
B
o
o
s
t
)
ea
ch
b
ein
g
u
s
ed
a
s
i
n
g
le
ti
m
e
i
n
o
u
r
s
elec
ted
p
ap
er
s
an
d
co
llectiv
e
co
u
n
t b
ein
g
1
9
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
2
3
,
No
.
2
,
A
p
r
il
20
2
5
:
4
0
2
-
415
408
Fig
u
r
e
4
.
C
o
u
n
t o
f
al
g
o
r
it
h
m
s
4
.
1
.
2
.
RQ
2
:
w
hich a
m
o
ng
ma
chine le
a
rning
a
nd
deep
lea
rning
is
bet
t
er
f
o
r
m
a
k
i
ng
y
ield
predict
io
n
s
?
A
s
t
u
d
y
[
4
2
]
th
a
t
co
m
p
ar
ed
SVR
,
p
ar
tial
least
s
q
u
ar
es
(
P
L
S)
re
g
r
ess
io
n
,
r
an
d
o
m
f
o
r
est
r
eg
r
ess
io
n
(
R
FR
)
an
d
DNN
s
h
o
w
ed
th
at
DNN
o
u
tp
er
f
o
r
m
ed
all
o
th
er
m
o
d
el
s
in
th
e
s
t
u
d
y
.
An
o
th
er
s
tu
d
y
[
3
8
]
d
o
n
e
o
n
s
o
y
b
ea
n
cr
o
p
co
m
p
ar
ed
v
ar
io
u
s
M
L
a
n
d
DL
m
o
d
els
s
u
ch
a
s
A
D
A
B
o
o
s
t,
DNN
,
least
ab
s
o
lu
te
s
h
r
i
n
k
a
g
e
an
d
s
elec
tio
n
o
p
er
ato
r
,
r
an
d
o
m
f
o
r
est,
an
d
SVM
o
u
t
o
f
w
h
ich
D
NN
o
u
tp
er
f
o
r
m
ed
all
th
e
m
o
d
els.
T
h
e
s
tu
d
y
[
4
3
]
w
h
ic
h
i
n
cl
u
d
ed
DNN
co
m
p
a
r
ed
v
ar
io
u
s
m
o
d
els
o
u
t
o
f
w
h
i
ch
SV
R
a
n
d
KNN
o
u
tp
er
f
o
r
m
ed
all
o
th
er
m
o
d
el
s
.
T
h
is
h
ap
p
en
ed
d
u
e
to
th
e
s
m
a
ll
d
ataset
u
s
ed
to
tr
ain
t
h
e
m
o
d
el
s
in
ce
DNN
is
m
o
r
e
s
en
s
iti
v
e
to
th
e
a
m
o
u
n
t
o
f
d
ata
f
ed
in
to
it.
A
s
i
m
ilar
s
t
u
d
y
[
4
1
]
w
h
ich
i
n
cl
u
d
ed
L
ASSO,
ra
n
d
o
m
f
o
r
est,
an
d
L
S
T
M
co
n
clu
d
ed
th
at
L
ST
M
p
er
f
o
r
m
ed
th
e
b
est
a
m
o
n
g
all
s
tu
d
ied
m
o
d
els
i
n
p
r
ed
ictin
g
w
i
n
ter
w
h
ea
t
y
ield
in
C
h
in
a.
T
h
e
s
tu
d
y
[
1
8
]
th
at
co
m
p
ar
ed
DL
m
o
d
els
s
u
c
h
as
DNN,
C
N
N,
L
ST
M
,
an
d
r
an
d
o
m
f
o
r
est
s
h
o
w
ed
t
h
at
DNN
p
er
f
o
r
m
ed
b
est
a
m
o
n
g
th
e
co
m
p
ar
ed
m
o
d
els.
T
h
e
r
esear
ch
w
o
r
k
[
4
4
]
s
h
o
w
ed
th
at
XGB
o
o
s
t
p
er
f
o
r
m
ed
b
etter
th
an
C
NN
an
d
L
ST
M
d
u
e
to
s
m
al
l
d
ata
s
et
a
g
ain
s
h
o
w
in
g
th
at
D
L
m
o
d
els
r
eq
u
ir
e
a
lar
g
e
d
ataset.
An
o
th
er
s
t
u
d
y
[
3
0
]
co
n
d
u
cted
o
v
er
th
e
r
e
g
io
n
o
f
P
u
n
j
ab
,
I
n
d
ia
o
n
W
h
ea
t
cr
o
p
co
m
p
ar
ed
R
NN
a
n
d
L
ST
M
w
ith
A
N
N,
r
an
d
o
m
f
o
r
est
an
d
m
u
lti
v
ar
iate
L
i
n
ea
r
r
eg
r
ess
io
n
.
R
N
N
an
d
L
ST
M
o
u
tp
er
f
o
r
m
ed
th
e
clas
s
ical
M
L
alg
o
r
it
h
m
s
w
it
h
a
lar
g
e
m
ar
g
i
n
.
An
o
t
h
er
s
tu
d
y
[
3
2
]
th
at
also
s
h
o
w
ed
th
at
s
tack
ed
L
ST
M
o
u
tp
er
f
o
r
m
ed
o
th
er
ML
m
o
d
el
s
in
cl
u
d
ed
in
th
e
s
t
u
d
y
s
u
c
h
as
L
A
S
SO
an
d
S
VR
.
An
o
th
er
s
tu
d
y
[
3
6
]
th
at
u
s
ed
en
s
e
m
b
l
ed
DL
m
o
d
els
al
s
o
s
h
o
w
ed
t
h
at
C
NN
-
R
N
N+
2
Fe
ed
f
o
r
w
ar
d
n
e
u
r
al
n
et
w
o
r
k
o
u
tp
er
f
o
r
m
ed
lin
ea
r
r
e
g
r
ess
io
n
,
XGB
o
o
s
t,
r
an
d
o
m
f
o
r
est
w
i
th
co
n
s
id
er
ab
le
d
if
f
er
en
ce
in
test
i
n
g
ac
cu
r
ac
y
.
W
h
il
e
it
ca
n
b
e
s
aid
th
at
th
er
e
is
n
o
d
ef
in
itiv
e
an
s
w
er
as
to
w
h
ic
h
ML
p
er
f
o
r
m
s
b
est.
I
t
ca
n
b
e
co
n
clu
d
ed
f
r
o
m
t
h
e
p
r
esen
t
liter
atu
r
e
th
a
t
th
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
m
o
d
el
s
m
aj
o
r
ly
d
ep
en
d
s
o
n
th
e
d
ataset
th
at
i
s
b
ein
g
u
s
ed
an
d
m
o
r
e
o
v
er
th
e
s
ize
o
f
d
ataset.
I
n
g
en
er
al
w
h
e
n
th
er
e
is
lar
g
e
d
ataset
av
a
ilab
le
DL
alg
o
r
it
h
m
s
ten
d
to
p
er
f
o
r
m
b
etter
an
d
w
h
e
n
th
er
e
i
s
co
m
p
ar
ativ
el
y
a
s
m
al
ler
d
ataset
ML
te
n
d
to
p
er
f
o
r
m
b
etter
f
o
r
m
a
k
i
n
g
cr
o
p
y
i
eld
p
r
ed
ictio
n
s
.
4
.
1
.
3
.
RQ
3
:
w
ha
t
a
re
m
a
t
ric
es us
ed
f
o
r
m
o
del e
v
a
lua
t
io
n?
E
v
alu
a
tio
n
m
etr
ic
s
ar
e
an
im
p
o
r
ta
n
t
asp
ec
t
o
f
an
y
s
t
u
d
y
a
s
th
ese
s
er
v
e
as
th
e
p
ar
am
eter
s
f
o
r
ev
alu
a
tin
g
h
o
w
w
ell
a
m
o
d
el
p
er
f
o
r
m
s
.
T
h
e
s
elec
tio
n
o
f
ev
alu
atio
n
m
etr
ics
d
ep
en
d
s
o
n
th
e
o
b
j
ec
tiv
e
o
f
th
e
s
tu
d
y
h
o
w
e
v
er
,
th
e
f
e
w
m
o
s
t
u
s
ed
m
etr
ics
f
o
r
cr
o
p
y
ield
p
r
ed
ictio
n
ar
e
R
MSE
,
R
^2
,
a
n
d
MA
E
.
T
h
ese
ar
e
u
s
ed
m
o
s
tl
y
i
n
r
eg
r
es
s
io
n
ta
s
k
s
.
W
h
ile
R
MSE
i
s
g
o
o
d
to
co
m
p
ar
e
m
o
d
el
s
o
n
th
e
s
a
m
e
d
a
taset
th
i
s
is
n
o
t
th
e
p
er
f
ec
t
p
ar
am
eter
to
co
m
p
ar
e
m
o
d
els
tr
ain
ed
o
n
d
if
f
er
en
t
d
a
tasets
as
R
MSE
ca
n
v
ar
y
a
n
d
d
o
esn
’
t
h
a
v
e
a
f
i
x
ed
r
an
g
e.
Fo
r
co
m
p
ar
is
o
n
o
f
m
o
d
els
f
r
o
m
d
i
f
f
er
en
t
s
tu
d
ies,
R
^2
v
a
lu
e
i
s
u
s
ed
b
ec
au
s
e
th
e
v
a
lu
e
r
a
n
g
e
s
f
r
o
m
0
to
1
,
1
b
ein
g
a
p
er
f
ec
t
m
o
d
el.
St
u
d
ies
o
f
ten
u
s
e
m
o
r
e
th
an
o
n
e
e
v
alu
a
tio
n
m
etr
ic
f
o
r
m
o
d
el
ev
al
u
atio
n
to
g
iv
e
a
b
etter
id
ea
f
o
r
c
o
m
p
ar
is
o
n
s
u
c
h
as
[4
5]
co
m
p
ar
in
g
4
m
o
d
els
o
n
all
t
h
r
ee
p
ar
a
m
eter
s
to
p
r
ed
ict
th
e
y
ield
.
A
u
th
o
r
s
h
av
e
s
i
m
ilar
l
y
[
1
8
]
m
ad
e
u
s
e
o
f
R
^2
an
d
R
MSE
to
ev
al
u
ate
p
r
ed
ictio
n
s
m
ad
e
o
n
w
h
ea
t
cr
o
p
s
.
T
h
e
Fig
u
r
e
5
s
h
o
w
s
th
e
f
r
eq
u
en
c
y
o
f
ev
al
u
atio
n
m
etr
ics
u
s
ed
i
n
t
h
e
p
ap
er
.
T
h
e
R
MSE
is
t
h
e
m
o
s
t
u
s
ed
ev
alu
a
tio
n
m
etr
ic
f
o
llo
w
ed
b
y
R
^2
an
d
M
A
E
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
A
d
va
n
ce
d
cro
p
yield
p
r
ed
ictio
n
u
s
in
g
m
a
ch
in
e
lea
r
n
in
g
a
n
d
d
ee
p
le
a
r
n
in
g
:
…
(
A
yu
s
h
A
n
a
n
d
)
409
Fig
u
r
e
5
.
E
v
alu
atio
n
m
e
tr
ics u
s
ed
in
p
ap
er
s
4
.
1
.
4
.
RQ
4
:
w
ha
t
a
re
t
he
da
t
a
s
o
urce
s
?
Data
s
o
u
r
ce
s
p
r
i
m
ar
il
y
d
ep
en
d
u
p
o
n
th
e
r
eg
io
n
o
f
s
tu
d
y
a
n
d
th
e
t
y
p
e
o
f
cr
o
p
b
ein
g
s
t
u
d
ied
.
C
r
o
p
lar
g
el
y
d
ep
en
d
s
u
p
o
n
m
eteo
r
o
lo
g
ical
d
ata
s
u
c
h
as
r
ai
n
f
all,
t
e
m
p
er
atu
r
e,
an
d
h
u
m
id
it
y
.
Na
tio
n
al
Oce
a
n
ic
a
n
d
A
t
m
o
s
p
h
er
ic
A
d
m
i
n
is
tr
atio
n
(
NOA
A
)
w
h
ic
h
is
a
p
ar
t
o
f
t
h
e
U.
S.
Dep
ar
t
m
e
n
t
o
f
C
o
m
m
er
ce
s
to
r
es
a
w
id
e
r
an
g
e
o
f
m
e
teo
r
o
lo
g
ical
d
atasets
m
aj
o
r
ly
f
r
o
m
th
e
U.
S.
T
er
r
ito
r
ies
an
d
w
a
ter
s
.
St
u
d
ies
s
u
c
h
as
[
2
0
]
,
[
4
6
]
co
n
d
u
cted
o
v
er
th
e
U.
S
m
a
k
e
u
s
e
o
f
th
i
s
s
er
v
ice.
Si
m
i
lar
to
NOAA
co
u
n
tr
ies
h
a
v
e
t
h
eir
o
w
n
d
atac
en
tr
e
to
m
o
n
ito
r
an
d
co
llect
d
ata
f
o
r
r
esear
ch
s
u
ch
as
t
h
e
s
t
u
d
y
[
4
7
]
co
n
d
u
cted
o
n
C
o
lu
m
b
ia
u
s
ed
a
d
ataset
o
b
tain
ed
f
r
o
m
t
h
e
C
o
n
s
u
l
tatio
n
an
d
D
o
w
n
lo
ad
o
f
H
y
d
r
o
m
eteo
r
o
lo
g
ical
Data
s
y
s
te
m
o
f
th
e
I
n
s
tit
u
te
o
f
H
y
d
r
o
lo
g
y
,
Me
teo
r
o
lo
g
y
an
d
E
n
v
ir
o
n
m
e
n
tal
S
tu
d
y
o
f
C
o
lu
m
b
ia,
Mi
n
is
tr
y
o
f
Ag
r
ic
u
lt
u
r
al
a
n
d
R
u
r
al
De
v
elo
p
m
en
t.
An
o
t
h
er
s
tu
d
y
[
4
8
]
co
n
d
u
cte
d
o
v
er
T
am
il
Nad
u
,
I
n
d
ia
u
s
ed
th
e
Dep
ar
t
m
e
n
t
o
f
E
co
n
o
m
ics
an
d
Sta
tis
tic
s
,
Go
v
er
n
m
en
t
o
f
T
a
m
il
Nad
u
.
Si
m
i
lar
s
t
u
d
ies
[
4
9
]
,
[
5
0
]
also
u
s
ed
I
n
d
ia
n
g
o
v
er
n
m
en
t
s
o
u
r
ce
s
to
o
b
tain
d
ata.
T
h
e
y
ie
ld
d
ata
o
f
cr
o
p
p
ar
ticu
lar
l
y
is
o
b
tai
n
ed
f
r
o
m
t
h
e
U
n
i
ted
States
Dep
ar
t
m
e
n
t
o
f
Ag
r
icu
lt
u
r
e
s
u
ch
a
s
i
n
th
e
s
t
u
d
ies
[
3
4
]
,
[
5
1
]
-
[
5
6
]
.
R
e
m
o
te
s
e
n
s
i
n
g
d
ata
is
ac
tiv
el
y
u
s
ed
to
m
o
n
ito
r
cr
o
p
s
.
Am
o
n
g
o
u
r
s
elec
ted
s
t
u
d
ies
s
atell
it
e
d
ata
w
as
th
e
m
aj
o
r
s
o
u
r
ce
o
f
r
em
o
te
s
e
n
s
i
n
g
d
ata
w
it
h
s
o
m
e
s
tu
d
ie
s
also
u
s
i
n
g
u
n
m
a
n
n
ed
ae
r
ial
v
e
h
icle
(
U
A
V)
.
Fo
r
th
e
s
atel
lites
,
MO
DI
S
w
as
u
s
e
d
m
aj
o
r
l
y
f
o
r
th
e
ca
lcu
latio
n
o
f
Veg
eta
tio
n
I
n
d
ice
s
in
5
8
%
o
f
th
e
s
tu
d
ie
s
s
u
c
h
as
[
1
4
]
,
[
1
9
]
,
[
2
6
]
,
[
3
5
]
,
[
4
4
]
,
[
5
1
]
,
[
5
7
]
-
[
6
1
]
.
Fo
llo
w
ed
b
y
L
a
n
d
s
at
8
s
atellite
u
s
ed
in
th
e
s
t
u
d
ies
[
1
3
]
,
[
5
7
]
,
[
6
2
]
,
[
6
3
]
an
d
L
an
d
s
at
7
u
s
ed
i
n
th
e
s
t
u
d
ies
[
3
5
]
,
[
6
2
]
.
T
h
is
w
as
f
o
llo
w
ed
b
y
t
h
e
u
s
e
o
f
o
th
e
r
s
atellite
s
s
u
ch
a
s
Se
n
ti
n
e
l
2
[
2
3
]
,
[
3
5
]
,
[
5
2
]
,
[
5
7
]
,
[
6
2
]
,
Sen
tin
el
2
B
an
d
Sen
t
in
el
2
B
[
6
4
]
,
W
o
r
ld
View
-
3
an
d
UAV
[
4
2
]
,
L
an
d
s
at
2
,
L
a
n
d
s
a
t
2
(
L
2
A
)
[
6
5
]
,
an
d
Sen
ti
n
el
2
L
1
C
[
1
1
]
.
T
h
e
Fig
u
r
e
6
r
ep
r
esen
t
s
th
e
s
atellite
s
o
u
r
ce
s
u
s
ed
in
t
h
e
s
t
u
d
ies.
Fig
u
r
e
6
.
T
y
p
e
o
f
s
atellite
s
u
s
ed
in
th
e
s
t
u
d
y
4
.
1
.
5
.
RQ
5
:
w
ha
t
a
re
t
he
m
o
s
t
us
ed
f
ea
t
ures?
Me
teo
r
o
lo
g
ical
d
ata
is
cr
u
cial
f
o
r
m
a
k
i
n
g
ac
cu
r
ate
p
r
ed
ictio
n
s
.
T
h
e
ty
p
e
o
f
m
eteo
r
o
lo
g
i
ca
l
f
ea
tu
r
e
d
ep
en
d
s
o
n
th
e
cr
o
p
b
ein
g
s
tu
d
ied
.
Ho
w
e
v
er
,
th
er
e
ar
e
a
ce
r
tain
g
r
o
u
p
o
f
m
o
s
t
co
m
m
o
n
l
y
u
s
ed
f
ea
tu
r
e
s
w
e
h
av
e
e
n
co
u
n
ter
ed
in
o
u
r
s
e
lec
ted
p
ap
er
s
s
u
ch
as
te
m
p
e
r
atu
r
e.
T
h
is
w
as
t
h
e
m
o
s
t
u
s
ed
f
ea
tu
r
e
a
m
o
n
g
all
t
h
e
s
elec
ted
s
tu
d
ie
s
s
u
ch
a
s
[
1
4
]
,
[
1
9
]
,
[
4
1
]
,
[
4
7
]
,
[
5
2
]
,
[
5
9
]
,
[
6
5
]
-
[
7
5
]
f
o
llo
w
ed
b
y
p
r
ec
ip
itatio
n
in
t
h
e
s
t
u
d
ie
s
[
1
5
]
,
[
1
7
]
,
[
1
8
]
,
[
2
1
]
,
[
3
3
]
,
[
5
2
]
,
[
5
4
]
,
[
5
9
]
,
[
7
2
]
,
[
7
4
]
-
[
7
6
]
,
r
ain
f
all
[
1
2
]
,
[
1
4
]
,
[
4
8
]
,
[
6
7
]
,
[
6
9
]
,
[
7
0
]
,
[
7
7
]
-
[
7
9
]
,
v
ap
o
u
r
p
r
ess
u
r
e
[
1
9
]
,
[
8
0
]
a
n
d
o
th
er
m
e
teo
r
o
lo
g
ical
d
ata
s
u
c
h
as
h
u
m
id
it
y
,
w
i
n
d
s
p
e
ed
,
h
u
m
id
it
y
.
So
il
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
2
3
,
No
.
2
,
A
p
r
il
20
2
5
:
4
0
2
-
415
410
ch
ar
ac
ter
is
tic
s
s
u
c
h
as
Ni
tr
o
g
en
,
P
h
o
s
p
h
o
r
u
s
a
n
d
p
o
tass
iu
m
w
er
e
m
o
s
t
u
s
ed
s
o
il
f
ea
tu
r
es
[
2
2
]
,
[
3
3
]
,
[
4
1
]
,
[
6
5
]
,
[
6
8
]
,
[
6
9
]
,
[
8
1
]
-
[
8
3
]
f
o
l
lo
w
ed
b
y
So
il
P
h
[
2
2
]
,
[
4
1
]
,
[
5
0
]
,
[
5
4
]
,
[
7
6
]
,
[
8
2
]
,
[
8
4
]
.
A
m
o
n
g
t
h
e
v
e
g
etatio
n
in
d
ices
o
b
tain
ed
f
r
o
m
s
atellite
i
m
ag
e
s
n
o
r
m
a
lized
d
if
f
er
e
n
c
e
v
eg
etatio
n
in
d
e
x
(
NDVI
)
[
1
1
]
,
[
1
3
]
,
[
1
4
]
,
[
1
9
]
,
[
2
3
]
,
[
2
6
]
,
[
2
9
]
,
[
5
7
]
,
[
5
8
]
,
[
6
5
]
,
[
6
8
]
,
[
8
5
]
-
[
8
7
]
w
a
s
t
h
e
m
o
s
t
u
s
ed
v
e
g
etatio
n
i
n
d
ex
f
o
ll
o
w
ed
b
y
E
n
h
an
ce
d
Veg
etatio
n
I
n
d
ex
(
E
VI
)
[
1
3
]
-
[
1
5
]
,
[
1
8
]
,
[
5
7
]
,
[
5
8
]
,
[
6
0
]
,
[
7
8
]
,
[
8
8
]
,
an
d
L
A
I
[
2
3
]
,
[
2
9
]
,
[
5
6
]
,
[
6
2
]
,
[
7
5
]
.
Fig
u
r
e
7
s
h
o
w
s
t
h
e
to
tal
co
u
n
t
o
f
th
e
m
o
s
t u
s
ed
f
ea
t
u
r
es u
s
e
d
in
th
e
s
elec
ted
p
ap
er
s
.
Fig
u
r
e
7
.
C
o
u
n
t
of
f
ea
t
u
r
es
4
.
1
.
6
.
RQ
6
:
w
hich
a
m
o
ng
ens
e
m
b
led
m
o
dels
a
nd
t
ra
ditio
na
l
m
a
c
hin
e
lea
rn
ing
a
nd
deep
lea
rni
ng
perf
o
r
m
bet
t
er
?
E
n
s
em
b
led
m
o
d
els
a
r
e
a
w
a
y
o
f
in
teg
r
at
in
g
m
o
r
e
th
an
o
n
e
c
lass
ic
al
m
o
d
el
to
g
et
ad
v
an
t
ag
es
o
f
b
o
th
an
d
in
cr
ea
s
e
th
e
a
cc
u
r
ac
y
o
f
th
e
p
r
ed
ict
io
n
.
T
h
e
p
a
p
e
r
[
5
2
]
d
em
o
n
s
tr
at
es
th
e
ef
f
ec
t
iv
en
ess
o
f
en
s
em
b
le
m
o
d
els
th
at
c
o
m
b
in
e
C
N
N
an
d
DNN
in
p
r
ed
ict
in
g
c
o
r
n
y
iel
d
s
.
T
h
e
en
s
em
b
le
m
o
d
e
ls
o
u
t
p
e
r
f
o
r
m
ed
in
d
iv
id
u
al
ML
m
o
d
els,
s
u
g
g
esti
n
g
th
at
th
e
co
m
b
in
ati
o
n
o
f
d
if
f
er
en
t
ty
p
es
o
f
n
eu
r
al
n
etw
o
r
k
s
ca
n
i
m
p
r
o
v
e
p
r
e
d
i
cti
o
n
ac
cu
r
a
cy
.
Sim
ilar
ly
,
th
e
r
esea
r
ch
[
2
7
]
als
o
ad
v
o
c
ates
f
o
r
e
n
s
em
b
le
m
o
d
els
in
c
o
r
n
y
ield
f
o
r
e
ca
s
t
in
g
.
T
h
e
o
p
tim
ized
w
eig
h
ted
en
s
em
b
le
an
d
th
e
av
e
r
ag
e
en
s
em
b
le
w
e
r
e
f
o
u
n
d
to
b
e
th
e
m
o
s
t
p
r
ec
is
e
m
o
d
e
ls
.
A
n
o
th
er
s
tu
d
y
[
8
7
]
u
s
e
d
en
s
em
b
le
t
r
e
e
m
eth
o
d
s
,
s
p
ec
if
i
ca
l
ly
b
o
o
s
te
d
r
eg
r
ess
i
o
n
t
r
ee
s
(
B
R
T
)
an
d
r
an
d
o
m
f
o
r
ests
,
f
o
r
ea
r
ly
p
r
e
d
i
cti
o
n
o
f
w
in
ter
w
h
e
at
y
ield
.
T
h
e
r
esu
lts
s
u
g
g
est
t
h
at
en
s
em
b
le
t
r
e
e
m
eth
o
d
s
ca
n
ef
f
ec
tiv
ely
h
an
d
le
co
m
p
lex
in
t
er
ac
ti
o
n
s
b
etw
ee
n
v
ar
ia
b
l
es
an
d
im
p
r
o
v
e
p
r
ed
i
cti
o
n
a
cc
u
r
a
cy
.
T
h
e
s
tu
d
y
[
8
9
]
f
o
cu
s
e
d
o
n
p
r
e
d
ic
tin
g
s
u
g
a
r
ca
n
e
y
iel
d
in
B
r
az
il
u
s
in
g
NDV
I
tim
e
s
e
r
i
es
an
d
n
eu
r
al
n
e
tw
o
r
k
s
en
s
em
b
l
e
an
d
als
o
s
u
g
g
este
d
th
at
en
s
em
b
le
m
eth
o
d
s
ca
n
b
e
ef
f
ec
tiv
e
in
d
if
f
e
r
en
t
g
e
o
g
r
ap
h
ical
l
o
c
ati
o
n
s
an
d
f
o
r
d
if
f
er
en
t
cr
o
p
s
.
Fro
m
th
e
s
e
s
t
u
d
ies,
it
s
ee
m
s
th
at
en
s
e
m
b
le
m
et
h
o
d
s
,
w
h
eth
er
th
e
y
ar
e
b
ased
o
n
class
ical
M
L
alg
o
r
ith
m
s
o
r
DL
alg
o
r
it
h
m
s
,
s
h
o
w
s
u
p
er
io
r
p
er
f
o
r
m
a
n
ce
i
n
cr
o
p
y
ield
p
r
ed
ictio
n
.
T
h
e
en
s
e
m
b
le
m
eth
o
d
s
ca
n
ef
f
ec
tiv
e
l
y
co
m
b
i
n
e
th
e
s
tr
en
g
th
s
o
f
m
u
lt
ip
le
m
o
d
els
to
im
p
r
o
v
e
p
r
ed
ictio
n
ac
cu
r
ac
y
an
d
p
r
ec
is
io
n
.
Ho
w
e
v
er
,
it’
s
i
m
p
o
r
tan
t
to
n
o
te
t
h
at
th
e
ch
o
ice
b
et
w
ee
n
en
s
e
m
b
le
an
d
class
ical
ML
o
r
DL
al
g
o
r
ith
m
s
m
a
y
d
ep
en
d
o
n
th
e
s
p
ec
i
f
ic
p
r
o
b
le
m
an
d
d
at
a
at
h
an
d
.
W
h
ile
e
n
s
e
m
b
le
m
et
h
o
d
s
h
a
v
e
s
h
o
w
n
p
r
o
m
i
s
i
n
g
r
es
u
lt
s
i
n
th
e
s
e
s
tu
d
ie
s
,
class
ical
M
L
o
r
DL
al
g
o
r
ith
m
s
m
i
g
h
t p
er
f
o
r
m
b
etter
in
o
th
er
s
ce
n
ar
io
s
.
4
.
1
.
7
.
RQ
7
:
w
ha
t
a
re
t
he
li
m
it
a
t
io
ns
a
nd
f
uture
direct
io
ns
?
W
ith
ad
v
an
c
in
g
tech
n
o
lo
g
y
o
f
s
atell
ites
,
it
h
a
s
b
ec
o
m
e
ea
s
ier
to
m
o
n
i
to
r
cr
o
p
s
an
d
o
b
tain
d
ata
f
o
r
m
ak
in
g
p
r
ed
ictio
n
s
.
Ho
w
e
v
er
,
a
lar
g
e
m
o
d
el
t
h
at
h
ea
v
il
y
r
elies
o
n
s
atellite
d
ata
is
s
till
v
er
y
d
i
f
f
icu
l
t
to
r
u
n
b
ec
au
s
e
o
f
th
e
v
er
y
h
i
g
h
-
r
e
s
o
lu
tio
n
s
a
tellite
i
m
ag
e
s
th
at
ar
e
t
y
p
icall
y
a
co
u
p
le
o
f
g
i
g
ab
y
te
s
,
an
d
th
e
y
h
av
e
to
b
e
co
llected
f
o
r
a
ce
r
tain
p
er
i
o
d
to
m
ak
e
a
h
i
s
to
r
ic
d
ataset
f
o
r
p
r
ed
ictio
n
s
m
ak
in
g
it
co
s
t
l
y
to
r
u
n
.
T
h
is
h
a
s
h
o
w
ev
er
b
ee
n
m
ad
e
a
litt
le
e
asier
b
y
Go
o
g
le
E
ar
th
E
n
g
i
n
e
w
h
ic
h
u
s
es
clo
u
d
co
m
p
u
ti
n
g
f
o
r
r
u
n
n
in
g
s
u
ch
h
ea
v
y
co
m
p
u
tat
io
n
s
.
B
u
t
w
it
h
lar
g
e
s
tu
d
y
ar
ea
s
,
th
is
is
s
till
a
p
r
o
b
lem
as
p
latf
o
r
m
s
lik
e
E
ar
th
E
n
g
i
n
e
h
a
v
e
th
eir
li
m
itat
io
n
s
s
u
ch
a
s
ti
m
e
o
u
t
li
m
it,
a
n
d
li
m
ited
s
ize
o
f
p
ar
ticu
lar
s
atell
ite
I
m
a
g
es
f
o
r
co
m
p
u
tatio
n
at
a
ti
m
e.
Fo
r
v
er
y
lar
g
e
s
t
u
d
y
ar
e
as,
it
’
s
n
ec
es
s
ar
y
to
b
u
y
h
ig
h
clo
u
d
s
to
r
ag
e
o
n
p
lat
f
o
r
m
s
lik
e
E
ar
th
E
n
g
i
n
e
f
o
r
m
ak
in
g
ca
lc
u
latio
n
s
o
r
w
o
r
k
o
n
v
er
y
h
i
g
h
-
e
n
d
m
ac
h
in
e
s
wh
ich
ar
e
t
y
p
icall
y
o
n
l
y
i
n
L
a
b
s
.
Fo
r
r
esear
ch
er
s
lo
o
k
in
g
to
u
s
e
th
ese
Sate
llit
e
i
m
ag
e
s
d
ir
ec
tl
y
,
th
e
y
h
av
e
to
b
e
d
o
w
n
lo
ad
ed
to
p
er
f
o
r
m
o
p
er
atio
n
s
u
s
i
n
g
v
ar
io
u
s
f
r
a
m
e
w
o
r
k
s
s
u
c
h
as
Geo
s
p
atial
L
ib
r
ar
ies
an
d
Op
en
C
V.
His
to
r
ic
i
m
a
g
es
o
f
a
n
a
r
ea
f
o
r
a
co
u
p
le
o
f
y
ea
r
s
ca
n
ea
s
il
y
g
o
o
v
er
a
f
e
w
h
u
n
d
r
ed
g
i
g
ab
y
te
s
an
d
m
o
r
e.
T
h
is
r
eq
u
ir
es v
er
y
h
ig
h
s
to
r
a
g
e
ca
p
ac
it
y
an
d
v
er
y
ad
v
an
ce
d
an
d
ca
p
ab
le
GP
Us
f
o
r
i
m
ag
e
f
ea
t
u
r
e
e
x
tr
ac
tio
n
s
.
T
h
e
co
s
t
r
ed
u
ctio
n
ca
n
b
e
lo
o
k
ed
as
a
j
o
in
t
e
f
f
o
r
t
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
A
d
va
n
ce
d
cro
p
yield
p
r
ed
ictio
n
u
s
in
g
m
a
ch
in
e
lea
r
n
in
g
a
n
d
d
ee
p
le
a
r
n
in
g
:
…
(
A
yu
s
h
A
n
a
n
d
)
411
o
f
r
esear
ch
er
s
in
t
h
e
f
ield
t
h
at
ca
n
s
et
u
p
lab
s
d
ed
icate
d
to
cr
o
p
p
r
ed
ictio
n
s
an
d
o
th
er
r
elate
d
w
o
r
k
to
cr
o
p
s
,
to
g
r
o
w
t
h
e
co
m
m
u
n
it
y
a
n
d
h
elp
r
esear
ch
er
s
co
llab
o
r
ate
w
h
ich
w
o
u
ld
in
t
u
r
n
r
ed
u
ce
t
h
e
co
s
t.
I
n
d
ev
elo
p
ed
co
u
n
tr
ies
li
k
e
T
h
e
U
n
ited
Sta
tes
a
n
d
C
a
n
ad
a
th
e
Ag
r
icu
ltu
r
e
s
ec
to
r
is
v
er
y
o
r
g
an
i
s
ed
th
at
t
h
e
g
o
v
er
n
m
e
n
t
is
ab
le
t
o
r
elea
s
e
a
h
ig
h
-
r
eso
lu
tio
n
cr
o
p
d
ata
lay
er
.
T
h
e
cr
o
p
ca
n
b
e
id
en
ti
f
ied
th
r
o
u
g
h
C
D
L
w
i
th
o
u
t
an
y
o
th
er
p
r
ep
r
o
c
ess
in
g
.
T
h
is
is
v
er
y
u
s
ef
u
l
as
it
allo
w
s
r
esear
ch
er
s
to
elim
i
n
ate
th
e
u
n
n
ec
es
s
ar
y
f
ield
s
o
r
r
o
a
d
s
a
n
d
b
u
ild
in
g
s
b
ein
g
co
n
s
id
er
ed
in
th
e
d
ataset.
Ho
w
e
v
er
,
th
er
e
ar
e
o
n
l
y
a
f
e
w
co
u
n
tr
ies
th
a
t h
a
v
e
b
ee
n
ab
le
t
o
d
o
th
is
.
I
n
d
e
v
elo
p
in
g
co
u
n
t
r
ies cr
o
p
d
ata
la
y
er
s
ar
e
s
ti
ll
u
n
attai
n
ab
le
b
ec
au
s
e
o
f
m
i
x
ed
ag
r
icu
lt
u
r
al
p
r
ac
tices
an
d
lack
o
f
m
o
n
ito
r
in
g
b
y
t
h
e
g
o
v
er
n
m
en
t
w
h
ic
h
m
a
k
es
it
d
if
f
ic
u
lt
f
o
r
r
esear
ch
er
s
to
tr
ac
k
d
o
w
n
t
h
e
ex
ac
t
ar
ea
o
f
t
h
e
cr
o
p
lan
d
t
h
at
is
to
b
e
s
t
u
d
ied
.
T
h
is
m
a
k
e
s
cr
o
p
d
ata
lay
er
s
a
v
er
y
cr
u
cia
l
p
ar
t
o
f
s
tu
d
ie
s
r
elate
d
to
cr
o
p
s
as
th
e
y
ar
e
r
esp
o
n
s
ib
le
f
o
r
th
e
p
r
ec
is
io
n
o
f
th
e
s
tu
d
y
as
s
ate
llit
e
i
m
a
g
es
b
ein
g
u
s
ed
d
ir
ec
tl
y
ca
n
n
o
t
p
r
o
v
id
e
as
a
s
ate
llit
e
i
m
a
g
e
co
n
tai
n
s
i
n
f
o
r
m
atio
n
o
f
n
o
t
o
n
l
y
t
h
e
cr
o
p
lan
d
b
u
t
also
o
f
s
u
r
r
o
u
n
d
in
g
u
n
cr
o
p
p
ed
ar
ea
th
at
is
n
o
t
n
ee
d
ed
f
o
r
s
tu
d
y
.
W
h
ile
i
t
is
d
i
f
f
icu
lt
to
m
ak
e
cr
o
p
d
ata
la
y
er
s
i
n
d
ev
elo
p
in
g
co
u
n
tr
ies,
ef
f
o
r
ts
ca
n
b
e
m
ad
e
b
y
s
tar
tin
g
to
w
o
r
k
w
it
h
s
m
all
ar
ea
s
th
at
co
v
er
a
p
ar
ticu
lar
s
tate,
d
is
tr
ict
o
r
co
u
n
t
y
an
d
f
u
r
t
h
er
b
e
ex
p
an
d
ed
to
o
th
er
ar
ea
s
o
f
t
h
e
co
u
n
tr
y
.
T
h
is
,
h
o
w
e
v
er
,
w
o
u
ld
r
eq
u
ir
e
a
lo
t
o
f
g
r
o
u
n
d
w
o
r
k
as
w
el
l
to
o
r
g
an
is
e
t
h
e
ag
r
ic
u
lt
u
r
al
p
r
ac
tices
s
o
tr
ac
k
in
g
cr
o
p
f
ield
s
ca
n
b
e
m
ad
e
ea
s
ier
f
o
r
a
p
ar
ticu
lar
cr
o
p
.
Mu
lti
-
s
o
u
r
ce
d
ata
ca
n
also
b
e
ex
p
er
i
m
e
n
ted
in
cl
u
d
in
g
d
if
f
er
en
t
s
o
il
f
ea
t
u
r
es,
w
ater
,
a
n
d
cli
m
ate
f
o
r
ea
ch
c
r
o
p
s
o
th
e
r
esear
ch
er
ca
n
m
a
k
e
in
f
o
r
m
ed
d
ec
is
io
n
s
a
s
to
w
h
at
e
x
ac
t
f
ea
tu
r
e
s
i
m
p
ac
t
th
e
cr
o
p
th
at
i
s
b
ein
g
s
tu
d
ied
.
T
h
e
m
o
r
e
r
elev
an
t
V
I
s
s
u
c
h
as
p
er
p
en
d
icu
lar
v
e
g
e
tatio
n
in
d
e
x
(
P
VI
)
,
s
o
il
-
ad
j
u
s
ted
v
e
g
etatio
n
in
d
e
x
(
S
A
VI
)
,
at
m
o
s
p
h
er
icall
y
r
esis
ta
n
t
v
e
g
etatio
n
in
d
e
x
(
AR
VI
)
,
s
o
lar
-
in
d
u
ce
d
f
l
u
o
r
escen
ce
(
SIF
)
,
a
n
d
d
if
f
er
en
ce
v
e
g
etat
io
n
i
n
d
ex
(
D
VI
)
ca
n
also
b
e
u
s
ed
to
f
u
r
t
h
er
f
o
r
m
ak
in
g
ac
c
u
r
ate
s
tu
d
ie
s
.
C
o
m
p
ar
in
g
t
h
i
s
s
t
u
d
y
w
it
h
s
i
m
ila
r
r
ev
ie
w
s
,
ac
co
r
d
in
g
to
a
r
ev
ie
w
[
7
]
,
th
e
m
o
s
t
f
r
eq
u
en
tl
y
u
s
ed
f
ea
t
u
r
es
i
n
cl
u
d
e
te
m
p
er
at
u
r
e,
s
o
il
t
y
p
e,
an
d
r
ai
n
f
all,
w
i
th
n
eu
r
al
n
et
w
o
r
k
s
a
n
d
li
n
ea
r
r
e
g
r
ess
io
n
b
ei
n
g
th
e
m
o
s
t
co
m
m
o
n
al
g
o
r
it
h
m
s
,
f
o
llo
w
ed
b
y
r
an
d
o
m
f
o
r
est
a
n
d
SVM.
C
o
m
m
o
n
e
v
al
u
atio
n
m
etr
ics
w
er
e
R
MSE
an
d
R
R
^2
.
A
n
o
t
h
er
s
t
u
d
y
[
6
]
n
o
ted
L
ST
M
an
d
C
NN
-
b
ased
ap
p
r
o
ac
h
es
as
p
r
ev
alen
t
D
L
tech
n
iq
u
es,
p
r
i
m
ar
il
y
u
s
i
n
g
th
e
M
ODI
S
s
ate
llit
e,
f
o
llo
w
ed
b
y
L
a
n
d
s
at
8
an
d
L
an
d
s
at
7
,
w
it
h
VI
s
,
m
eteo
r
o
lo
g
i
ca
l
d
ata,
an
d
y
ield
in
f
o
r
m
atio
n
a
s
k
e
y
f
ea
t
u
r
es.
A
d
d
itio
n
al
l
y
,
[
8
]
id
en
tif
ied
V
I
s
an
d
s
atellite
d
ata,
alo
n
g
s
id
e
h
is
to
r
ical
y
ield
an
d
cli
m
ate
d
ata,
as
f
r
eq
u
en
t
l
y
u
s
ed
in
p
u
ts
,
w
it
h
r
an
d
o
m
f
o
r
est
an
d
A
NN
a
s
to
p
alg
o
r
ith
m
s
,
f
o
llo
w
ed
b
y
C
NN,
u
s
i
n
g
R
MSE
,
A
cc
u
r
ac
y
,
an
d
R
2
R
^2
R
2
f
o
r
ev
al
u
atio
n
.
Si
m
ilar
l
y
,
[
9
]
o
b
s
er
v
ed
im
ag
e
s
,
p
r
ec
ip
itatio
n
,
an
d
ac
tu
al
y
ie
ld
as
m
aj
o
r
f
ea
tu
r
es
,
w
it
h
C
NN,
L
ST
M,
A
NN,
an
d
DNN
as
co
m
m
o
n
alg
o
r
it
h
m
s
,
u
s
i
n
g
R
MSE
as
th
e
p
r
i
m
ar
y
ev
al
u
atio
n
m
e
tr
ic.
An
o
th
er
r
ev
ie
w
[
1
0
]
s
h
o
w
ed
C
NN
a
n
d
R
NN
a
s
th
e
m
o
s
t
-
u
s
ed
al
g
o
r
ith
m
s
.
I
n
co
m
p
ar
is
o
n
,
o
u
r
s
t
u
d
y
h
ig
h
li
g
h
ts
te
m
p
er
atu
r
e
a
n
d
p
r
ec
ip
itatio
n
as
p
r
i
m
ar
y
f
ea
t
u
r
es,
MO
DI
S
an
d
L
a
n
d
s
at
8
as
p
r
im
ar
y
s
atellite
s
,
w
ith
R
MSE
an
d
R
^2
as
m
ai
n
ev
al
u
atio
n
m
etr
ics,
an
d
r
an
d
o
m
f
o
r
est
an
d
SVM
as
p
r
ev
alen
t
m
o
d
el
s
,
w
it
h
C
NN
a
n
d
L
ST
M
as
s
tate
-
of
-
t
h
e
-
ar
t
tec
h
n
iq
u
es.
T
ab
le
3
r
e
p
r
esen
ts
th
e
co
m
p
ar
is
o
n
a
m
o
n
g
r
an
d
o
m
l
y
s
elec
ted
p
ap
er
s
.
T
ab
le
3
.
C
o
m
p
ar
is
o
n
w
i
th
s
i
m
ilar
w
o
r
k
R
e
f
P
a
p
e
r
f
o
c
u
s
F
i
n
d
i
n
g
s
[
4
6
]
I
mp
r
o
v
i
n
g
c
r
o
p
y
i
e
l
d
p
r
e
d
i
c
t
i
o
n
i
n
M
o
r
o
c
c
o
.
M
L
mo
d
e
l
s
o
u
t
p
e
r
f
o
r
me
d
st
a
t
i
s
t
i
c
a
l
mo
d
e
l
s
i
n
m
a
k
i
n
g
p
r
e
d
i
c
t
i
o
n
s
.
M
L
mo
d
e
l
s a
c
h
i
e
v
e
d
R
^
2
r
a
n
g
i
n
g
f
r
o
m 0
.
7
6
t
o
0
.
8
4
.
[
5
2
]
C
o
u
n
t
y
l
e
v
e
l
c
o
r
n
y
i
e
l
d
p
r
e
d
i
c
t
i
o
n
u
si
n
g
C
N
N
-
D
N
N
i
n
U
S
c
o
r
n
b
e
l
t
.
T
h
e
mo
d
e
l
mad
e
2
0
1
9
p
r
e
d
i
c
t
i
o
n
w
i
t
h
R
M
S
E
o
f
8
6
6
k
g
/
h
a
.
[
7
6
]
I
mp
r
o
v
i
n
g
c
r
o
p
y
i
e
l
d
p
r
e
d
i
c
t
i
o
n
i
n
C
h
i
n
a
.
P
r
o
p
o
se
d
a
mo
d
e
l
t
h
a
t
p
r
e
d
i
c
t
s
p
r
e
se
a
so
n
a
n
d
i
n
se
a
so
n
p
r
e
d
i
c
t
i
o
n
f
o
r
5
c
r
o
p
s.
[
6
3
]
S
i
l
a
g
e
mai
z
e
y
i
e
l
d
p
r
e
d
i
c
t
i
o
n
u
s
i
n
g
t
i
me
se
r
i
e
s d
a
t
a
se
t
f
r
o
m N
D
V
I
.
B
R
T
h
a
d
h
i
g
h
e
st
R
v
a
l
u
e
o
f
0
.
8
7
.
[
2
8
]
P
r
o
p
o
se
d
a
f
r
a
me
w
o
r
k
f
o
r
w
h
e
a
t
y
i
e
l
d
p
r
e
d
i
c
t
i
o
n
.
L
A
S
S
O
r
e
c
e
i
v
e
d
h
i
g
h
e
st
p
e
r
f
o
r
man
c
e
w
i
t
h
R
^
2
o
f
0
.
9
3
.
T
h
e
f
in
d
in
g
s
o
f
t
h
e
r
esear
ch
q
u
esti
o
n
s
i
n
th
is
s
t
u
d
y
h
ig
h
lig
h
t
th
at,
lead
i
n
g
cr
o
p
y
ie
ld
p
r
ed
ictio
n
m
o
d
el
s
in
cl
u
d
e
en
s
e
m
b
le
tec
h
n
iq
u
es
li
k
e
r
an
d
o
m
f
o
r
est
a
n
d
XGB
o
o
s
t,
alo
n
g
s
id
e
D
L
ap
p
r
o
ac
h
es
s
u
ch
a
s
L
ST
M
an
d
C
NN.
E
ac
h
m
o
d
el
t
y
p
e
o
f
f
er
s
d
is
tin
c
t
s
tr
en
g
th
s
d
e
p
en
d
in
g
o
n
d
ataset
s
iz
e
an
d
co
m
p
le
x
it
y
.
E
n
s
e
m
b
le
an
d
n
eu
r
al
n
et
w
o
r
k
m
o
d
els
w
o
r
k
esp
ec
iall
y
w
el
l
w
it
h
lar
g
er
d
atasets
d
u
e
to
th
e
ir
ab
ilit
y
to
ca
p
tu
r
e
co
m
p
le
x
p
atter
n
s
,
w
h
ile
tr
ad
i
tio
n
al
ML
m
o
d
els
ca
n
p
er
f
o
r
m
ef
f
ec
ti
v
el
y
w
it
h
s
m
aller
d
atasets
.
Me
tr
ics
li
k
e
R
MSE
,
R
²,
a
n
d
M
A
E
ar
e
co
m
m
o
n
l
y
u
s
ed
to
ev
al
u
ate
m
o
d
el
ac
cu
r
ac
y
an
d
r
eliab
ilit
y
.
Data
s
o
u
r
ce
s
s
p
an
m
eteo
r
o
lo
g
ical
i
n
f
o
r
m
atio
n
f
r
o
m
o
r
g
a
n
i
s
atio
n
s
li
k
e
NO
AA
,
w
h
er
e
te
m
p
er
atu
r
e
i
s
a
f
r
eq
u
en
tl
y
u
s
ed
f
ea
t
u
r
e,
to
s
atellite
-
b
ased
r
e
m
o
te
s
en
s
in
g
d
ata
w
it
h
v
e
g
etatio
n
i
n
d
i
ce
s
lik
e
NDVI
a
n
d
E
VI
,
w
h
i
ch
ar
e
ess
e
n
tia
l
f
o
r
m
o
n
ito
r
i
n
g
cr
o
p
h
ea
lt
h
.
Ho
wev
er
,
ch
alle
n
g
e
s
p
er
s
is
t,
s
u
c
h
as
th
e
h
i
g
h
co
m
p
u
tat
io
n
al
c
o
s
t
o
f
p
r
o
ce
s
s
in
g
s
atellite
i
m
ag
er
y
an
d
li
m
ited
cr
o
p
d
ata
av
ailab
ilit
y
i
n
d
ev
elo
p
in
g
r
eg
io
n
s
.
Fu
tu
r
e
r
esear
ch
s
h
o
u
ld
f
o
cu
s
o
n
i
m
p
r
o
v
i
n
g
m
o
d
el
ef
f
icie
n
c
y
,
c
r
ea
tin
g
cr
o
p
-
s
p
ec
i
f
ic
d
ata
r
eso
u
r
ce
s
in
r
u
r
al
ar
ea
s
,
an
d
f
o
s
ter
in
g
co
llab
o
r
atio
n
s
Evaluation Warning : The document was created with Spire.PDF for Python.