I
n
t
e
r
n
at
ion
al
Jou
r
n
al
of
E
lec
t
r
ical
an
d
Com
p
u
t
e
r
E
n
gin
e
e
r
in
g
(
I
JE
CE
)
Vol.
12
,
No.
1
,
F
e
br
ua
r
y
20
22
,
pp.
639
~
648
I
S
S
N:
2088
-
8708
,
DO
I
:
10
.
11591/
ij
e
c
e
.
v
12
i
1
.
pp
6
39
-
648
639
Jou
r
n
al
h
omepage
:
ht
tp:
//
ij
e
c
e
.
iaes
c
or
e
.
c
om
A
n
ove
l
w
e
at
h
e
r
p
a
r
am
e
t
e
r
s
p
r
e
d
ic
t
io
n
sc
h
e
m
e
a
n
d
t
h
e
i
r
e
f
f
e
c
t
s
on
c
r
o
p
s
L
in
gar
aj
u
Nave
e
n
,
H
o
s
a
a
g
r
a
ha
r
a
Sa
v
a
l
e
g
o
w
da
M
o
ha
n
D
e
pa
r
tm
e
nt
of
I
nf
or
ma
ti
on S
c
ie
nc
e
a
nd E
ngi
ne
e
r
in
g,
S
J
B
I
ns
ti
t
ut
e
of
T
e
c
hnol
ogy
,
V
is
ve
s
va
r
a
ya
T
e
c
hnol
ogi
c
a
l
U
ni
ve
r
s
it
y
,
B
e
lg
a
um
, I
ndi
a
Ar
t
icle
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
ived
F
e
b
11,
2021
R
e
vis
e
d
J
ul
16,
2021
Ac
c
e
pted
Aug
4
,
2021
W
eat
h
er
fo
reca
s
t
i
s
s
i
g
n
i
f
i
can
t
l
y
i
mp
era
t
i
v
e
i
n
t
o
d
a
y
’s
s
mart
t
ec
h
n
o
l
o
g
i
cal
w
o
r
l
d
.
A
p
rec
i
s
e
fo
reca
s
t
mo
d
el
e
n
t
a
i
l
s
a
p
l
en
t
i
f
u
l
d
at
a
i
n
o
r
d
er
t
o
at
t
ai
n
t
h
e
mo
s
t
accu
ra
t
e
p
red
i
ct
i
o
n
s
.
H
o
w
ev
er,
a
fo
recas
t
o
f
fu
t
u
re
rai
n
fal
l
fro
m
h
i
s
t
o
r
i
ca
l
d
a
t
a
s
am
p
l
e
s
h
a
s
a
l
w
ay
s
b
ee
n
ch
a
l
l
e
n
g
i
n
g
an
d
k
e
y
area
o
f
res
earch
.
H
en
ce,
i
n
mo
d
er
n
w
ea
t
h
er
fo
reca
s
t
i
n
g
a
co
mb
o
o
f
co
mp
u
t
er
mo
d
e
l
s
,
o
b
s
er
v
at
i
o
n
,
an
d
k
n
o
w
l
ed
g
e
o
f
t
re
n
d
s
an
d
p
a
t
t
ern
s
are
i
n
t
ro
d
u
c
ed
.
T
h
i
s
res
earch
w
o
r
k
h
a
s
p
re
s
en
t
ed
a
f
i
t
n
es
s
fu
n
ct
i
o
n
b
a
s
ed
ad
a
p
t
i
v
e
ar
t
i
f
i
ci
a
l
n
eu
ra
l
n
et
w
o
r
k
s
ch
eme
i
n
o
rd
er
t
o
fo
recas
t
rai
n
fa
l
l
an
d
t
em
p
erat
u
re
fo
r
u
p
c
o
mi
n
g
d
eca
d
e
(2
0
2
1
-
2
0
3
0
)
u
s
i
n
g
h
i
s
t
o
ri
ca
l
w
eat
h
er
d
at
a
o
f
2
0
d
i
ffere
n
t
d
i
s
t
ri
ct
s
o
f
K
ar
n
at
a
k
a
s
t
at
e.
F
u
rt
h
ermo
re,
effect
s
o
f
t
h
e
s
e
fo
recas
t
ed
w
e
at
h
er
p
aramet
er
s
are
real
i
zed
o
v
er
fi
v
e
maj
o
r
cro
p
s
o
f
K
ar
n
at
a
k
a
n
ame
l
y
r
i
ce,
w
h
ea
t
,
j
o
w
ar,
mai
ze,
an
d
rag
i
w
i
t
h
t
h
e
i
n
t
e
n
t
i
o
n
o
f
ev
al
u
a
t
i
o
n
fo
r
effi
c
i
en
t
cro
p
ma
n
ag
eme
n
t
i
n
t
erms
o
f
t
h
e
p
as
s
i
n
g
re
l
ev
a
n
t
me
s
s
ag
e
s
t
o
t
h
e
farmers
an
d
al
t
er
n
at
e
meas
u
res
s
u
ch
as
s
u
g
g
es
t
i
n
g
o
t
h
er
g
eo
g
rap
h
i
cal
l
o
c
at
i
o
n
s
t
o
g
ro
w
t
h
e
s
ame
cro
p
o
r
g
ro
w
i
n
g
o
t
h
er
s
u
i
t
ab
l
e
cro
p
s
at
s
ame
g
eo
g
ra
p
h
i
cal
l
o
ca
t
i
o
n
.
A
g
rap
h
i
ca
l
u
s
er
i
n
t
erface
(G
U
I)
ap
p
l
i
cat
i
o
n
h
as
b
een
d
ev
e
l
o
p
ed
fo
r
t
h
e
p
ro
p
o
s
ed
w
o
rk
i
n
o
r
d
er
t
o
eas
e
o
u
t
t
h
e
fl
o
w
o
f
w
o
rk
.
K
e
y
w
o
r
d
s
:
Ada
ptation
C
r
op
mana
ge
ment
Hidde
n
laye
r
M
a
c
hine
lea
r
ning
Ne
ur
on
Th
i
s
i
s
a
n
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
t
h
e
CC
B
Y
-
SA
l
i
ce
n
s
e.
C
or
r
e
s
pon
din
g
A
u
th
or
:
L
ingar
a
ju
Na
ve
e
n
De
pa
r
tm
e
nt
of
I
nf
or
mat
ion
S
c
ienc
e
a
nd
E
nginee
r
i
ng
,
Vis
ve
s
va
r
a
ya
T
e
c
hnologi
c
a
l
Unive
r
s
it
y
B
e
lgaum
,
I
ndia
E
mail:
na
ve
e
nli
nga
r
a
ju@gm
a
il
.
c
om
1.
I
NT
RODU
C
T
I
ON
W
e
a
ther
f
or
e
c
a
s
ti
ng
in
the
domain
of
mete
or
ology
is
one
of
the
ke
y
a
ppli
c
a
ti
ons
o
f
moder
n
s
c
ienc
e
a
nd
tec
hnology
in
or
de
r
to
e
nvis
a
ge
the
s
tate
of
t
he
E
a
r
th’
s
a
tm
os
phe
r
e
f
or
the
pa
r
ti
c
ular
ge
ogr
a
ph
ica
l
a
r
e
a
.
M
or
e
ove
r
,
the
mos
t
a
c
c
ur
a
te
we
a
ther
pr
e
dictions
a
r
e
f
a
c
il
it
a
ted
by
making
the
e
xtens
ive
us
a
ge
of
wide
r
a
nge
his
tor
ica
l
we
a
ther
da
ta
a
nd
then
e
volvi
ng
the
e
f
f
icie
nt
a
lgor
it
hms
in
o
r
de
r
to
pr
ojec
t
the
we
a
ther
pa
r
a
mete
r
s
in
c
ur
r
e
nt
a
nd
f
utu
r
e
s
tate
of
ti
me
.
He
nc
e
,
the
f
o
r
e
c
a
s
t
models
de
r
ived
f
r
om
mac
hine
lea
r
ning,
s
of
t
c
omput
ing
a
nd
da
ta
m
ini
ng
a
r
e
qua
li
tat
ively
e
mpl
oye
d
he
r
e
i
n
c
ur
r
e
nt
s
c
e
na
r
ios
[
1
]
-
[
4]
.
As
a
r
e
s
ult
,
we
a
ther
f
o
r
e
c
a
s
ti
ng
ha
s
tur
ne
d
out
to
be
a
n
im
pe
r
a
ti
ve
a
nd
vit
a
l
a
r
e
a
of
r
e
s
e
a
r
c
h
in
the
today’
s
s
mar
t
wor
ld
of
int
e
r
ne
t
of
thi
ngs
(
I
oT
s
)
.
W
e
a
ther
is
e
mbr
a
c
e
d
of
va
r
ious
pa
r
a
mete
r
s
s
u
c
h
a
s
tempe
r
a
tur
e
,
r
a
inf
a
ll
,
r
e
lative
hu
mi
dit
y,
wind
s
pe
e
d,
a
nd
moi
s
tur
e
.
He
nc
e
,
a
n
a
c
c
ur
a
te
we
a
ther
p
r
e
diction
model
r
e
quir
e
s
a
n
a
mpl
e
a
mount
of
da
ta
on
thes
e
we
a
ther
pa
r
a
mete
r
s
to
a
c
hieve
the
highe
r
or
de
r
pr
e
c
is
ions
in
f
o
r
e
c
a
s
ts
.
Although
a
s
a
r
e
s
ult
of
mete
or
ologi
c
a
l
da
ta
im
pr
e
c
is
ion,
the
c
oll
e
c
ted
da
taba
s
e
is
e
xa
gge
r
a
ted
by
va
r
ious
types
of
im
pr
oba
bil
it
ies
,
unc
e
r
tai
nti
e
s
a
nd
pe
r
tur
ba
ti
ons
.
He
nc
e
,
with
the
int
e
ns
ion
of
r
e
s
o
lut
ion
thes
e
dil
e
mm
a
s
a
nd
nuis
a
nc
e
s
,
e
xtr
a
r
e
f
i
ne
d
a
nd
s
pe
c
if
ic
s
tocha
s
ti
c
models
a
r
e
e
nha
n
c
e
d
f
ur
ther
.
F
ur
ther
mor
e
,
the
f
r
e
nz
ied
a
nd
c
ha
oti
c
c
ha
r
a
c
ter
of
the
e
nvir
onment,
the
im
mens
e
a
nd
c
ons
ider
a
ble
a
mo
unt
of
c
omput
a
ti
ona
l
a
lgor
it
hms
a
r
e
r
e
quir
e
d
in
or
de
r
to
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
12
,
No.
1
,
F
e
br
ua
r
y
20
22
:
639
-
648
640
de
c
ipher
a
nd
he
nc
e
wor
k
out
the
mathe
matica
l
e
qua
ti
ons
por
tr
a
ying
the
a
tm
os
phe
r
e
in
c
onjunction
with
the
inac
c
ur
a
c
y
im
pli
c
a
ted
dur
ing
me
a
s
ur
e
ment
o
f
the
ini
ti
a
l
s
it
ua
ti
ons
,
a
n
i
mper
f
e
c
t
a
nd
c
ur
taine
d
pe
r
c
e
pti
ve
of
the
a
tm
os
phe
r
ic
pr
oc
e
s
s
e
s
r
e
s
ult
ing
in
f
or
e
c
a
s
ts
w
it
h
de
gr
a
de
d
a
c
c
ur
a
c
y.
Va
r
ious
types
of
mac
hine
lea
r
ning
ba
s
e
d
a
lgor
it
hms
a
r
e
e
mpl
oye
d
in
or
de
r
to
a
va
il
the
pr
e
dictions
with
hi
ghe
r
a
c
c
ur
a
c
y
s
uc
h
a
s
R
e
gr
e
s
s
ion
models
,
s
uppor
t
ve
c
tor
mac
hine
(
S
VM
)
,
a
r
ti
f
icia
l
ne
ur
a
l
ne
twor
k
(
AN
N)
,
de
e
p
ne
ur
a
l
ne
twor
k
(
D
NN
)
,
a
nd
S
e
gmenta
ti
on
a
nd
c
lus
ter
ing
[5
]
-
[
7]
.
I
n
thi
s
r
e
s
e
a
r
c
h
wor
k,
a
uthor
ha
s
wor
ke
d
on
the
p
r
e
diction
o
f
the
ke
y
we
a
ther
pa
r
a
m
e
ter
s
tempe
r
a
tur
e
a
nd
r
a
inf
a
ll
us
ing
a
da
pti
ve
a
r
ti
f
icia
l
ne
ur
a
l
ne
twor
k
ba
s
e
d
on
f
it
ne
s
s
f
unc
ti
on
e
va
lu
a
ti
on
f
or
upc
omi
ng
ten
ye
a
r
s
us
ing
pa
s
t
ye
a
r
s
w
e
a
ther
da
ta
of
va
r
ious
dis
tr
icts
a
c
r
os
s
Ka
r
na
taka
s
tate
.
F
ur
t
he
r
mor
e
,
ba
s
e
d
on
the
f
or
e
c
a
s
ted
pa
r
a
mete
r
s
,
a
uthor
ha
s
wor
ke
d
on
va
r
ious
c
r
op
models
in
or
de
r
to
e
va
l
ua
te
their
e
f
f
e
c
ts
on
c
r
op
yields
.
I
n
a
ddit
ion
to
th
is
if
f
o
r
e
c
a
s
ted
we
a
ther
pa
r
a
mete
r
s
a
r
e
not
met
with
idea
l
we
a
ther
c
ondit
ions
f
or
c
r
ops
,
a
uthor
ha
s
a
ls
o
wor
ke
d
on
a
lt
e
r
na
te
mea
s
ur
e
s
s
uc
h
a
s
s
ugge
s
ti
ng
other
ge
og
r
a
phica
l
loca
ti
ons
to
gr
ow
the
s
a
me
c
r
op
or
gr
owing
other
s
uit
a
ble
c
r
ops
a
t
s
a
me
ge
ogr
a
phic
a
l
loca
ti
on
in
or
de
r
to
boos
t
c
r
op
yield
a
nd
pr
oduc
ti
on
a
nd
he
nc
e
e
leva
te
e
c
onomy
in
upc
omi
ng
ye
a
r
s
de
c
a
de
.
T
he
r
e
s
t
of
the
pa
pe
r
is
planne
d
a
s
f
oll
o
ws
:
S
e
c
ti
on
2
pr
e
s
e
nts
the
br
ie
f
li
ter
a
tu
r
e
s
ur
ve
y
ove
r
the
va
r
ious
types
of
we
a
ther
pr
e
diction
models
.
S
e
c
ti
on
3
pr
e
s
e
nts
the
pr
opos
e
d
wor
k
o
f
we
a
ther
p
r
e
diction
model
a
nd
c
r
op
mana
ge
ment
s
ys
tem.
S
e
c
ti
on
4
pr
e
s
e
nts
de
tailed
r
e
s
ult
s
a
nd
a
na
lys
is
of
wor
k
di
s
c
us
s
e
d
in
s
e
c
ti
on
3.
S
e
c
ti
ons
5
p
r
e
s
e
nt
the
c
onc
lus
ion
of
the
pa
pe
r
.
2.
L
I
T
E
RA
T
UR
E
S
UR
VE
Y
Da
il
y
we
a
ther
f
or
e
c
a
s
t
is
int
e
ns
if
ied
with
the
he
lp
of
a
va
s
t
number
o
f
obs
e
r
ve
r
s
a
nd
mete
or
ologi
s
ts
univer
s
a
ll
y
r
ight
th
r
ough
the
whole
wor
ld
.
M
e
teor
ologi
s
ts
a
c
tually
make
us
e
of
a
me
r
ge
of
a
li
tt
le
in
im
it
a
ble
tec
hnique
in
or
de
r
to
c
ons
ider
the
c
oll
e
c
ted
da
i
ly
ba
s
is
we
a
ther
da
ta.
Va
r
ious
types
of
f
o
r
e
c
a
s
ti
ng
a
r
e
c
omput
e
r
f
or
e
c
a
s
ti
ng,
s
tatis
ti
c
a
l
f
or
e
c
a
s
ti
ng,
s
yno
pti
c
f
or
e
c
a
s
ti
ng,
a
nd
pe
r
s
is
tenc
e
f
or
e
c
a
s
ti
ng.
W
it
h
the
he
lp
of
thes
e
f
or
e
c
a
s
ti
ng
s
c
he
mes
,
a
mong
thes
e
f
or
e
c
a
s
ter
s
c
ons
ider
the
opti
mi
z
e
d
s
c
he
me
i
n
r
e
ga
r
d
of
the
we
a
ther
c
ondit
ions
in
or
de
r
to
be
pr
e
dicte
d
a
ll
t
he
wa
y
th
r
ough
e
ve
r
yda
y
da
ta.
F
u
r
ther
mor
e
,
mo
s
t
of
the
r
e
s
e
a
r
c
he
r
s
ha
ve
c
a
r
r
ied
out
the
wor
ks
while
s
e
tt
ing
up
a
li
ne
a
r
r
e
lations
hip
f
lanke
d
by
the
we
a
ther
da
ta
input
a
nd
the
s
ubs
e
que
nt
tar
ge
t
da
ta.
M
or
e
ove
r
,
due
to
the
innovation
of
nonli
ne
a
r
it
y
in
we
a
ther
da
ta,
t
he
pr
im
e
objec
ti
ve
the
r
e
s
e
a
r
c
he
s
ha
ve
budge
d
towa
r
ds
the
nonli
ne
a
r
we
a
ther
da
ta
f
or
e
c
a
s
ti
ng.
A
wor
ldwide
numer
ica
l
we
a
ther
pr
e
diction
ba
s
e
d
on
the
dif
f
e
r
e
nti
a
l
c
ondit
ions
wa
s
pr
opos
e
d
in
[
8]
.
T
he
polynom
ial
ne
ur
a
l
s
ys
tems
ba
s
e
d
on
the
pa
r
ti
a
l
dif
f
e
r
e
nti
a
l
s
c
he
mes
tol
e
r
a
ti
ng
the
dis
play
pr
og
r
e
s
s
ively
mor
e
c
ompos
it
e
a
uthentic
plan
c
a
pa
bil
it
y
a
s
of
dis
c
r
e
te
ti
me
obs
e
r
va
ti
ons
a
s
c
ompar
e
d
to
the
s
uppor
t
of
s
tanda
r
d
ins
ubs
tantial
r
e
gis
ter
i
ng
s
c
he
mes
.
A
hyb
r
id
c
a
lcula
ti
on
in
o
r
de
r
to
a
c
c
or
dingl
y
c
ons
tr
uc
t
a
s
pr
e
a
d
-
out
hypothes
is
wor
k
ba
s
e
d
on
ne
ur
a
l
s
ys
tem
wa
s
p
r
opos
e
d
in
[
9]
.
T
he
hybr
id
s
c
he
me
pr
opos
e
d
wa
s
b
a
s
e
d
on
ge
ne
ti
c
a
lgor
it
hm,
pa
r
ti
c
le
s
wa
r
m
opti
mi
z
a
ti
on
s
c
he
me
a
nd
wor
ldwide
im
p
r
ove
m
e
nt
e
xe
c
uti
on
s
c
he
me.
A
M
OS
s
c
he
duli
ng
s
c
he
me
ba
s
e
d
on
a
n
e
xtr
ove
r
ted
a
s
s
e
mbl
y
of
mete
or
ologi
c
a
l
f
e
a
tu
r
e
s
wa
s
pr
opos
e
d
in
[
10]
.
T
his
s
c
he
me
wa
s
de
r
ived
f
r
o
m
a
s
tepw
is
e
dir
e
c
t
r
e
laps
e
c
omput
a
ti
on
mus
hy
a
r
e
laps
e
dis
plays
with
a
n
a
s
s
e
mbl
y
of
f
a
c
tor
s
.
T
he
c
on
jec
tur
e
a
ga
ins
t
E
T
0
judged
by
hou
r
ly
ba
s
is
c
oll
e
c
ted
da
ta
us
ing
the
40
va
r
ious
invo
lunt
a
r
y
we
a
ther
s
tations
a
ll
a
c
r
os
s
Aus
tr
a
li
a
c
ountr
y
w
a
s
pr
opos
e
d
in
[
11]
.
T
he
s
uppor
t
of
numer
ica
l
we
a
ther
pr
e
diction
(
NW
P
)
e
s
ti
mate
f
o
r
da
il
y
ba
s
is
E
T
0
wa
s
f
ound
to
be
of
be
tt
e
r
-
qua
li
ty
while
incor
por
a
ti
ng
the
a
va
il
a
ble
da
ta
to
mont
hly
a
nd
he
nc
e
ye
a
r
ly
ba
s
is
d
a
ta.
A
ge
r
m
-
gr
a
in
s
c
he
me
wa
s
pr
opos
e
d
in
[
12
]
.
I
n
thi
s
wor
k,
the
gr
a
ins
we
r
e
us
ua
ll
y
int
e
r
p
r
e
ted
mor
e
l
ikely
to
be
p
r
e
c
ipi
ta
ti
on
c
e
ll
s
.
M
or
e
ove
r
,
the
ge
r
m
-
gr
a
in
dis
pl
a
y
wa
s
a
bs
olut
e
ly
r
e
pr
e
s
e
nted
a
s
c
ombi
na
ti
on
of
the
c
los
e
by
r
e
lative
f
or
c
e
s
a
long
with
t
he
gr
a
in
mea
s
ur
e
.
A
nove
l
s
c
he
me
obs
e
r
ve
d
f
r
o
m
the
twi
s
t
e
s
ti
mations
a
r
is
ing
due
to
s
e
a
winds
jum
ble
m
e
ter
wa
s
p
r
o
p
o
s
e
d
i
n
[
1
3
]
.
A
n
e
x
p
e
r
i
m
e
n
t
a
l
a
s
s
e
s
s
m
e
n
t
d
e
r
i
v
e
d
f
r
o
m
t
h
e
m
e
a
n
s
q
u
a
r
e
e
r
r
o
r
(
M
S
E
)
,
r
o
o
t
m
e
a
n
s
q
u
a
r
e
e
r
r
o
r
(
R
M
S
E
)
a
n
d
t
h
e
a
s
t
o
n
i
s
h
i
n
g
a
n
d
s
t
a
n
d
a
r
d
r
e
l
i
a
b
l
e
c
o
r
r
e
l
a
t
i
o
n
w
a
s
s
o
p
h
i
s
t
i
c
a
t
e
d
f
r
o
m
s
p
a
t
i
a
l
,
m
o
m
e
n
t
a
r
y
a
n
d
dir
e
c
ti
ona
l
view
point
.
B
us
tami
e
t
al
.
[
14
]
pr
opos
e
d
a
f
ull
y
c
onne
c
ted,
f
e
e
d
f
or
wa
r
d
mu
lt
i
-
laye
r
pr
e
c
e
ptor
(
M
L
P
)
ne
twor
k
c
ons
is
ti
ng
of
thr
e
e
laye
r
s
-
ba
s
e
d
tempe
r
a
tur
e
pr
e
diction
s
c
he
mes
.
T
he
e
r
r
or
c
a
lcula
ted
he
r
e
wa
s
f
ound
to
be
mi
nim
a
l.
F
o
r
tr
a
ini
ng
ba
c
k
pr
opa
ga
t
ion
a
lgor
it
hm
wa
s
incor
por
a
ted.
F
u
r
ther
mor
e
,
the
p
r
e
dictions
a
r
e
c
ons
tr
a
ined
with
a
n
uppe
r
he
a
d,
whic
h
c
a
n
f
ur
the
r
be
r
e
ga
r
de
d
a
s
tum
bli
ng
down
the
di
f
f
e
r
e
nt
ge
o
gr
a
phica
l
loca
ti
ons
int
e
r
c
ha
nge
a
bil
it
ies
.
A
nove
l
tempe
r
a
tur
e
f
or
e
c
a
s
t
model
f
or
maximum
a
nd
mi
nim
um
tem
pe
r
a
tur
e
f
or
e
c
a
s
ti
ng
a
long
with
the
r
e
lative
hum
idi
ty
p
r
e
di
c
ti
on
wa
s
pe
r
f
or
med
b
y
making
the
e
xtens
ive
us
e
of
t
im
e
s
e
r
ies
a
na
lys
i
s
.
T
he
incor
po
r
a
ted
ne
twor
k
model
wa
s
mul
ti
laye
r
f
e
e
d
f
o
r
wa
r
d
a
r
ti
f
icia
l
ne
ur
a
l
ne
twor
k
with
tr
a
ini
ng
a
lgo
r
it
hm
of
ba
c
k
p
r
opa
ga
ti
on.
F
or
both
maximum
a
nd
mi
nim
um
tempe
r
a
tur
e
f
or
e
c
a
s
t
q
ua
r
ter
ly
ba
s
e
d
da
taba
s
e
wa
s
uti
li
z
e
d.
T
he
e
s
ti
mate
d
e
r
r
or
w
a
s
f
ound
to
be
les
s
than
3%
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
E
lec
&
C
omp
E
ng
I
S
S
N:
2088
-
8708
A
N
ov
e
l
w
e
ather
par
ame
ter
s
pr
e
diction
s
c
he
me
an
d
their
e
ff
e
c
ts
on
c
r
o
ps
(
L
ingar
aju
N
av
e
e
n
)
641
Ar
ti
f
icia
l
ne
u
r
a
l
ne
twor
k
ba
s
e
d
opti
mi
z
a
ti
on
ba
s
e
d
on
the
mul
ti
va
r
iate
f
o
r
e
c
a
s
ti
ng
[
15
]
,
global
s
olar
r
a
diation
e
s
ti
mation
ba
s
e
d
on
a
r
ti
f
icia
l
ne
ur
a
l
ne
tw
or
k
[
16
]
,
r
a
dial
ba
s
is
f
unc
ti
on
ne
ur
a
l
ne
two
r
k
de
r
iv
e
d
f
r
om
a
n
im
pr
ove
d
e
xpone
nti
a
l
de
c
r
e
a
s
ing
iner
ti
a
we
ight
-
pa
r
ti
c
le
s
wa
r
m
opti
mi
z
a
ti
on
a
l
g
o
r
i
t
h
m
[
1
7
]
,
p
a
r
t
i
c
l
e
s
w
a
r
m
o
p
t
i
m
i
z
a
t
i
o
n
(
P
S
O
)
b
a
s
e
d
G
a
u
s
s
i
a
n
r
a
d
i
a
l
b
a
s
i
s
f
u
n
c
t
i
o
n
n
e
t
w
o
r
k
s
[
1
8
]
,
n
e
u
r
o
-
f
u
z
z
y
i
n
f
e
r
e
n
c
e
s
y
s
t
e
m
a
n
d
s
u
b
s
e
q
u
e
n
t
l
e
a
r
n
i
n
g
m
e
c
h
a
n
i
s
m
s
[
1
9
]
a
r
e
f
e
w
o
t
h
e
r
f
o
r
e
c
a
s
t
i
n
g
s
c
h
e
m
e
s
p
r
o
p
o
s
e
d
in
li
ter
a
tur
e
s
.
3.
P
ROP
OS
E
D
WORK
T
his
wor
k
is
e
xtens
ion
of
the
wor
k
c
a
r
r
ied
out
f
or
r
a
inf
a
ll
pr
e
diction
a
nd
c
r
op
mana
ge
ment
c
a
r
r
ied
out
in
[
20]
,
[
21
]
mot
ivate
d
f
r
om
the
va
r
ious
l
it
e
r
a
tur
e
s
s
ur
ve
ye
d
in
[
22]
by
a
uthor
in
or
de
r
to
c
a
r
r
y
out
the
we
a
ther
f
or
e
c
a
s
t
(
ba
s
e
d
on
tempe
r
a
tur
e
a
nd
r
a
inf
a
ll
)
a
long
with
the
e
f
f
e
c
ts
of
f
o
r
e
c
a
s
ted
we
a
ther
pa
r
a
mete
r
s
on
va
r
ious
c
r
ops
f
o
r
va
r
ious
dis
tr
icts
a
c
r
os
s
Ka
r
na
taka
s
tate
f
or
upc
omi
ng
de
c
a
de
(
2021
-
2030)
.
He
nc
e
,
the
f
or
e
c
a
s
t
of
im
pe
nding
c
r
op
s
c
ulpt
s
de
r
ived
f
r
om
the
f
or
e
c
a
s
ted
we
a
ther
a
s
s
oc
iate
d
da
ta
e
na
bles
a
n
uppe
r
ha
nd
to
the
f
a
r
mer
s
in
o
r
de
r
to
ge
t
hold
o
f
e
ve
r
y
im
pe
r
a
ti
ve
s
tep
f
or
ne
c
e
s
s
a
r
y
pr
otec
ti
ons
or
pr
e
ve
nti
ve
/alter
na
te
mea
s
ur
e
s
f
or
s
pe
c
if
ied
c
r
ops
.
As
a
r
e
s
ult
,
the
pr
opos
e
d
wor
k
c
a
n
be
f
ur
ther
s
ubd
ivi
de
d
in
thr
e
e
major
s
tage
s
na
mely,
we
a
ther
(
tempe
r
a
tur
e
a
nd
r
a
inf
a
ll
)
a
nd
c
r
ops
r
e
late
d
da
ta
c
oll
e
c
ti
on,
we
a
ther
(
tempe
r
a
tur
e
a
nd
r
a
inf
a
ll
)
f
or
e
c
a
s
t,
e
va
luation
o
f
f
or
e
c
a
s
ted
we
a
ther
pa
r
a
mete
r
s
on
c
r
ops
(
e
f
f
e
c
ts
o
n
c
r
ops
)
a
nd
s
ugge
s
ti
ng
a
lt
e
r
na
ti
ve
mea
s
ur
e
s
.
T
he
de
tailed
f
low
diagr
a
m
f
or
p
r
opos
e
d
wor
k
is
p
r
e
s
e
nted
in
F
ig
ur
e
1.
F
igur
e
1.
F
low
diagr
a
m
o
f
p
r
opos
e
d
wor
k
3.
1.
Adap
t
ive
ar
t
i
f
icial
n
e
u
r
al
n
e
t
wor
k
(
AA
NN
)
Ar
ti
f
icia
l
ne
ur
a
l
ne
twor
ks
a
r
e
a
mongs
t
one
of
t
he
mos
t
pr
omi
s
ing
mac
hine
lea
r
ning
a
lgor
i
thm
s
mot
ivate
d
f
r
om
the
s
ubs
e
que
nt
biol
ogica
l
e
quivale
nc
ies
.
M
or
e
ove
r
,
a
da
ptation
is
a
nother
ke
y
f
e
a
tu
r
e
in
thes
e
ne
twor
ks
.
He
nc
e
,
thes
e
a
da
pti
ve
a
r
ti
f
icia
l
ne
ur
a
l
ne
tw
or
ks
a
r
e
ba
s
ica
ll
y
e
mpl
oye
d
in
dyna
mi
c
a
nd
vibr
a
nt
s
it
ua
ti
ons
.
T
he
y
a
r
e
us
ua
ll
y
e
xe
mpl
i
f
ied
a
nd
s
uppor
ted
with
the
he
lp
of
onli
ne
lea
r
ning
f
ur
ther
.
T
he
a
da
pti
ve
na
tur
e
of
ne
twor
ks
is
a
c
hieve
d
by
e
it
he
r
upda
ti
ng
w
e
ight
s
or
a
lt
e
r
a
ti
on
in
ne
ur
ons
c
ha
r
a
c
ter
is
t
ics
or
a
da
ptation
in
ne
twor
k
s
tr
uc
tur
e
.
He
nc
e
,
a
da
pti
ve
c
ha
r
a
c
ter
is
ti
c
is
a
n
indi
s
pe
ns
a
ble
a
nd
ke
y
e
leme
n
t
f
or
the
given
a
r
ti
f
icia
l
ne
ur
a
l
ne
two
r
k
in
or
de
r
to
s
howc
a
s
e
the
a
dde
d
c
us
hion
of
indepe
nde
nc
e
.
I
n
c
a
s
e
of
dyna
mi
c
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
12
,
No.
1
,
F
e
br
ua
r
y
20
22
:
639
-
648
642
s
ys
tem
e
nvir
onment,
f
utur
e
pr
e
dict
ions
o
r
f
or
e
c
a
s
ts
a
r
e
dif
f
icult
a
nd
he
nc
e
the
lea
r
ning
mec
ha
nis
m
f
or
the
c
or
r
e
s
ponding
e
nvir
onment
tur
ns
out
to
be
wor
thl
e
s
s
.
As
a
r
e
s
ult
,
f
or
mul
a
ti
on
of
a
methodology
in
or
de
r
to
a
da
pt
the
c
ha
nge
s
o
r
dyna
mi
c
s
of
s
ur
r
ounding
in
r
e
a
l
ti
me
be
c
omes
e
s
s
e
nti
a
l
whe
r
e
the
ne
w
a
da
pti
ve
s
ys
tem
r
e
a
c
ts
to
e
a
c
h
a
nd
indi
vidual
input
s
e
pa
r
a
tely.
I
n
the
p
r
opos
e
d
we
a
ther
f
or
e
c
a
s
t
a
nd
c
r
op
mana
g
e
ment
model
dur
ing
the
tr
a
ini
ng
of
c
onve
nti
ona
l
a
r
ti
f
icia
l
ne
ur
a
l
ne
twor
k,
the
ne
ur
a
l
we
ight
s
a
r
e
r
e
vived
dur
ing
e
a
c
h
e
poc
h
ti
ll
th
e
de
s
ir
e
d
a
c
c
ur
a
c
y
leve
l
is
a
c
hieve
d.
T
his
is
quit
e
it
e
r
a
ti
ve
,
tedious
a
nd
c
o
s
t
inef
f
e
c
ti
ve
.
He
nc
e
,
int
r
oduc
ti
on
o
f
a
da
pti
ve
modeling
withi
n
a
r
ti
f
icia
l
ne
ur
a
l
ne
twor
k
to
a
va
il
we
a
ther
f
or
e
c
a
s
t
f
or
upc
omi
ng
de
c
a
de
ba
s
e
d
on
his
tor
ica
l
we
a
ther
da
ta
is
f
ound
t
o
be
ha
ndf
ul.
An
int
a
ngibl
e
mod
e
l
f
or
s
oli
tar
y
leve
l
of
a
da
pti
ve
a
r
ti
f
icia
l
ne
ur
a
l
ne
twor
k
pr
oc
e
s
s
ing
s
ys
tem
is
s
hown
in
F
igur
e
2.
T
he
mode
l
take
s
his
tor
ica
l
da
ta
a
s
input
a
nd
p
r
ovides
pr
e
dict
ion
da
ta
a
s
output
.
L
e
t’
s
f
or
input
va
lue
of
x(
n
)
a
nd
with
t
he
s
e
t
of
va
lues
of
x(
n
-
1)
,
the
e
xpe
c
ted
pr
e
diction
is
x*(
n
)
.
I
n
a
ddit
ion,
thi
s
e
xpe
c
ted
pr
e
diction
x*(
n
)
is
c
omp
a
r
e
d
with
the
a
c
tual
va
lue
x(
n
)
a
nd
c
or
r
e
c
ti
on
va
lu
e
c
(
n
)
is
c
a
lcula
ted.
Onc
e
the
z
e
r
o
c
or
r
e
c
ti
on
is
a
c
hieve
d
mea
ns
pr
e
diction
is
a
c
hieve
d
with
s
up
e
r
ior
a
c
c
ur
a
c
y.
W
hil
e
the
c
or
r
e
c
ti
on
va
lue
is
non
-
z
e
r
o,
model
is
s
ti
ll
in
pr
ogr
e
s
s
to
pr
e
dict
the
c
or
r
e
c
t
va
lue
a
nd
da
ta
upd
a
te
is
s
ti
ll
in
pr
oc
e
s
s
.
T
h
is
c
or
r
e
c
ti
on
s
ignal
is
us
e
f
ul
in
o
r
d
e
r
to
r
e
gulate
the
ne
twor
k
model
in
s
uc
h
a
manne
r
that
it
e
na
bles
th
e
pe
r
f
e
c
t
a
nd
p
r
e
c
is
e
pr
e
diction
f
or
t
he
given
s
it
ua
ti
on.
He
nc
e
,
it
a
da
pts
the
e
nvi
r
o
nmenta
l
dyna
mi
c
s
ha
ppe
ning
in
r
e
a
l
ti
me
a
ppli
c
a
ti
ons
.
F
igur
e
2
.
S
t
r
uc
tur
e
o
f
a
da
pti
ve
model
3.
2.
P
r
op
os
e
d
ad
ap
t
ive
ar
t
if
icial
n
e
u
r
al
n
e
t
wor
k
(
AA
NN
)
wit
h
we
igh
t
u
p
d
at
e
an
d
f
it
n
e
s
s
f
u
n
c
t
ion
T
he
p
r
opos
e
d
a
r
ti
f
icia
l
ne
ur
a
l
ne
twor
k
us
e
s
we
i
ght
upda
te
s
c
he
me
in
or
de
r
to
e
na
ble
a
da
pti
ve
f
e
a
tur
e
.
T
h
is
is
a
c
hieve
d
by
s
tabili
ty
i.
e
.
the
a
bi
li
ty
of
the
de
s
igned
ne
twor
k
f
or
s
uc
c
e
s
s
f
ul
uti
li
z
a
ti
on
of
his
tor
ica
l
da
ta
a
nd
plas
ti
c
it
y
i
.
e
.
the
a
bil
it
y
of
th
e
de
s
igned
ne
twor
k
to
a
da
pt
the
upda
ted
or
ne
wly
a
dde
d
f
e
a
tur
e
s
.
T
he
r
e
a
r
e
two
s
e
t
of
ne
u
r
on
laye
r
s
na
me
ly
input
laye
r
ne
u
r
on
or
f
e
a
tur
e
s
e
t
laye
r
ne
ur
on
(
N1)
a
nd
output
laye
r
ne
ur
on
or
c
las
s
if
ied
s
e
t
la
ye
r
ne
ur
on
(
N2)
.
Ne
ur
ons
withi
n
both
the
laye
r
s
a
r
e
f
ull
y
int
e
r
c
onne
c
ted.
Ne
ur
ons
of
N2
laye
r
a
r
e
c
a
pa
ble
t
o
hold
c
or
r
e
s
ponding
input
ne
ur
ons
f
r
om
N1
laye
r
thr
ough
de
ve
loped
downw
a
r
d
a
s
s
oc
iate
d
we
ight
s
.
He
nc
e
,
dur
ing
lea
r
ning
a
nd
r
e
c
a
ll
,
the
we
ight
s
a
r
e
up
da
ted
a
c
c
or
dingl
y
with
the
wa
y
of
s
ignal
f
low.
A
c
o
r
r
e
c
t
pr
e
diction
is
given
by
pr
ope
r
s
ignal
f
low
f
r
om
N1
laye
r
to
N2
laye
r
,
whic
h
r
e
s
ult
s
in
c
or
r
e
c
tnes
s
of
N2
laye
r
.
N2
f
u
r
ther
r
e
tur
ns
the
pr
e
diction
r
e
late
d
e
xpe
c
tati
on
ba
c
k
to
N1.
T
h
is
pr
oc
e
s
s
will
c
ont
inue
ti
ll
the
c
or
r
e
c
tnes
s
is
non
-
z
e
r
o.
He
nc
e
a
da
ptation
c
onti
nue
ti
ll
the
u
pda
te
in
N1
laye
r
ne
ur
ons
a
nd
f
ur
ther
N2
laye
r
ne
u
r
ons
ge
tt
ing
a
s
s
igned
to
a
f
or
mer
ly
une
xploi
ted
c
las
s
to
the
ne
w
c
las
s
a
nd
the
c
or
r
e
s
ponding
we
ight
s
a
r
e
a
c
c
us
tom
e
d
in
or
de
r
to
e
xp
li
c
it
ly
dis
ti
nguis
h
the
ne
w
c
las
s
.
T
he
c
umul
a
ti
ve
model
f
or
the
pr
opos
e
d
a
da
pti
ve
a
r
ti
f
icia
l
ne
ur
a
l
ne
twor
k
p
r
e
s
e
nted
in
F
ig
ur
e
3
is
given
by
‘
N’
f
or
r
e
s
ult
a
nt
outcome
o
f
the
de
s
ig
ne
d
ne
twor
k
f
or
the
r
e
s
pe
c
ti
ve
hidden
laye
r
.
I
n
the
pr
opos
e
d
ne
twor
k,
e
a
c
h
input
da
ta
gr
ows
by
f
a
c
t
or
of
upda
ted
r
e
s
pe
c
ti
ve
we
ight
in
c
or
r
e
s
ponding
hidden
laye
r
da
ta.
He
nc
e
,
the
s
ubs
e
que
nt
we
ight
e
d
in
f
e
r
e
nc
e
s
f
or
the
c
or
r
e
s
ponding
input
da
ta
ge
t
s
umm
e
d
up
with
the
int
e
nti
on
o
f
the
lea
ning
of
the
r
e
s
pe
c
ti
ve
ne
ur
o
n
a
s
s
hown
in
(
1
)
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
E
lec
&
C
omp
E
ng
I
S
S
N:
2088
-
8708
A
N
ov
e
l
w
e
ather
par
ame
ter
s
pr
e
diction
s
c
he
me
an
d
their
e
ff
e
c
ts
on
c
r
o
ps
(
L
ingar
aju
N
av
e
e
n
)
643
′
(
)
=
[
∑
∗
=
1
]
+
(
=
1
,
2
,
…
,
)
(
1)
W
he
r
e
,
is
s
tea
dy
s
tate
b
ias
va
lue
a
nd
is
upda
t
e
d
we
ight
ge
tt
ing
mul
ti
pli
e
d
with
the
c
or
r
e
s
ponding
input
.
I
n
a
ddit
ion,
K
r
e
pr
e
s
e
nts
number
of
inpu
t
node
s
a
nd
L
r
e
pr
e
s
e
nts
number
of
hidden
nod
e
s
in
c
or
r
e
s
ponding
laye
r
.
F
ur
the
r
mor
e
,
the
a
c
tual
r
e
s
ult
a
nt
ne
twor
k
output
is
e
va
luate
d
with
r
e
s
pe
c
t
to
the
pr
e
dicte
d
ne
twor
k
output
in
or
de
r
to
c
a
lcula
te
the
e
r
r
or
a
nd
he
nc
e
to
e
va
luate
the
pe
r
f
or
manc
e
o
f
a
da
pti
ve
s
ys
tem
a
s
s
hown
in
(
2)
.
T
h
is
e
r
r
o
r
s
hould
be
r
e
s
tr
icte
d
in
or
de
r
to
a
tt
a
in
the
opti
mal
ne
twor
k
s
tr
uc
tur
e
.
He
nc
e
,
the
we
ight
upda
te
s
hould
be
c
onti
nue
d
ti
l
l
t
his
e
r
r
or
ge
ts
s
e
tt
led
withi
n
s
pe
c
if
ied
tol
e
r
a
nc
e
li
m
it
s
.
=
1
∗
∑
(
(
)
−
(
)
)
2
=
1
(
2)
F
ur
ther
mor
e
,
the
ba
s
e
d
thi
s
e
r
r
or
f
it
ne
s
s
f
unc
ti
on
dur
ing
e
a
c
h
c
yc
le
is
e
va
luate
d
in
or
de
r
to
a
c
hieve
the
be
s
t
f
or
e
c
a
s
t
r
e
s
ult
s
whic
h
is
obtaine
d
us
ing
(
3
)
.
Onc
e
the
be
s
t
f
it
r
e
s
ult
s
a
r
e
a
tt
a
ined
with
mi
nim
um
e
r
r
or
,
the
upda
ted
we
ight
s
c
a
n
be
us
e
d
in
o
r
de
r
to
tr
a
in
a
nd
he
nc
e
s
im
ulate
d
the
de
s
igned
ne
twor
k
to
ob
tain
the
f
or
e
c
a
s
ted
we
a
ther
pa
r
a
mete
r
s
.
[
]
=
m
in
(
)
(
3)
F
igur
e
3
.
Ar
c
hit
e
c
tur
e
of
p
r
opos
e
d
a
da
pti
ve
a
r
ti
f
ici
a
l
ne
ur
a
l
ne
twor
k
model
3.
3.
P
r
op
os
e
d
wor
k
f
low
As
s
hown
in
f
low
diagr
a
m
in
F
ig
ur
e
1
ba
s
e
d
on
th
e
his
tor
ica
l
we
a
ther
da
ta
a
nd
c
r
op
da
ta
of
va
r
ious
ge
ogr
a
phica
l
loca
ti
ons
,
the
de
s
igned
ne
twor
k
model
is
pr
e
pa
r
e
d
with
a
ppr
opr
iate
ne
twor
k
tr
a
ini
ng
.
Ne
xt,
f
or
the
s
e
lec
ted
de
mogr
a
phic
loca
ti
on
a
nd
r
e
s
pe
c
ti
ve
ye
a
r
the
we
a
ther
pa
r
a
mete
r
s
(
r
a
inf
a
ll
,
m
a
xim
um
tempe
r
a
tur
e
,
a
ve
r
a
ge
tempe
r
a
tu
r
e
a
nd
mi
n
im
um
t
e
mper
a
tur
e
)
a
r
e
f
or
e
c
a
s
ted.
Ne
xt,
ba
s
e
d
on
the
f
o
r
e
c
a
s
ted
we
a
the
r
pa
r
a
mete
r
s
a
nd
s
e
lec
ted
c
r
ops
,
c
r
ops
a
r
e
e
va
luate
d
a
ga
ins
t
the
f
or
e
c
a
s
ted
we
a
ther
pa
r
a
mete
r
s
.
Us
ing
thes
e
da
ta,
e
f
f
e
c
ts
on
c
r
ops
a
r
e
e
va
luate
d
a
nd
r
e
s
p
e
c
ti
ve
mes
s
a
ge
s
a
r
e
c
onve
ye
d
to
f
a
r
mer
s
.
F
ur
ther
mor
e
,
the
a
lt
e
r
na
te
mea
s
ur
e
s
uc
h
a
s
s
ugge
s
ti
ons
f
or
dif
f
e
r
e
n
t
ge
ogr
a
phica
l
loca
ti
ons
in
o
r
de
r
to
g
r
ow
s
a
me
c
r
op
or
f
or
s
a
me
ge
ogr
a
phica
l
loca
ti
on
in
or
de
r
to
gr
ow
other
c
r
ops
a
r
e
pa
s
s
e
d
on
to
f
a
r
mer
s
in
or
de
r
to
incr
e
a
s
e
the
c
r
op
yield
a
nd
he
nc
e
to
boos
t
c
ountr
y
e
c
onomy
in
upc
o
mi
ng
de
c
a
de
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
12
,
No.
1
,
F
e
br
ua
r
y
20
22
:
639
-
648
644
4.
RE
S
UL
T
S
AN
D
AN
AL
YSI
S
Author
ha
s
c
oll
e
c
ted
a
his
tor
ica
l
we
a
ther
da
ta
o
f
20
dis
tr
icts
a
c
r
os
s
Ka
r
na
taka
s
tate
f
or
r
a
inf
a
ll
(
mont
hly
a
ve
r
a
ge
da
ta
f
or
23
ye
a
r
s
)
a
nd
tempe
r
a
t
ur
e
(
mont
hly
maximu
m,
a
ve
r
a
ge
a
nd
mi
n
im
um
da
ta
f
or
10
ye
a
r
s
)
f
r
om
Ka
r
na
taka
S
tate
Na
tu
r
a
l
Dis
a
s
ter
M
onit
or
ing
C
e
ntr
e
thr
ough
De
pa
r
tm
e
nt
of
Agr
icultur
e
,
B
a
nga
lor
e
.
I
n
a
ddit
ion
to
thi
s
,
a
utho
r
a
ls
o
c
ol
lec
ted
a
s
a
li
e
nt
ge
ogr
a
phica
l
C
r
op
da
ta
(
tempe
r
a
tu
r
e
,
r
a
inf
a
ll
,
s
oil
type,
d
is
tr
icts
pr
oduc
ing
a
nd
s
e
a
s
ons
)
f
or
5
major
c
r
ops
(
whe
a
t,
r
ice
,
r
a
gi,
jowa
r
a
nd
maiz
e
)
a
c
r
os
s
Ka
r
na
taka
s
tate
.
T
he
p
r
opos
e
d
r
e
s
e
a
r
c
h
wo
r
k
ha
s
be
e
n
s
olely
im
pl
e
mente
d
u
s
i
n
g
m
a
t
r
i
x
l
a
b
o
r
a
t
o
r
y
(
M
A
T
L
A
B
)
t
o
o
l
.
F
u
r
t
h
e
r
m
o
r
e
,
a
g
r
a
p
h
i
c
a
l
u
s
e
r
i
n
t
e
r
f
a
c
e
(
G
U
I
)
a
p
p
l
i
c
a
t
i
o
n
h
a
s
b
e
e
n
d
e
s
i
g
n
e
d
i
n
M
A
T
L
A
B
i
n
o
r
d
e
r
t
o
f
o
r
e
c
a
s
t
t
h
e
w
e
a
t
h
e
r
p
a
r
a
m
e
t
e
r
s
a
n
d
h
e
n
c
e
e
v
a
l
u
a
t
e
t
h
e
m
f
o
r
c
r
o
p
m
a
n
a
g
e
m
e
n
t
a
s
s
h
o
w
n
i
n
F
i
g
u
r
e
4
.
F
i
g
u
r
e
s
5
a
nd
6
s
how
the
f
o
r
e
c
a
s
ted
r
a
inf
a
ll
pr
o
f
il
e
s
,
f
o
r
e
c
a
s
ted
tempe
r
a
te
pr
o
f
il
e
s
a
nd
c
or
r
e
s
ponding
c
r
op
de
tails
a
long
with
r
e
leva
nt
mes
s
a
ge
s
to
f
a
r
me
r
s
a
nd
a
lt
e
r
na
ti
ve
me
a
s
ur
e
f
or
‘
Ha
s
s
a
n’
c
it
y
f
o
r
ye
a
r
2026
f
o
r
‘
M
a
ize
’
c
r
op
.
F
ur
ther
mor
e
,
in
a
ddit
ion
to
thi
s
,
F
ig
ur
e
s
7
a
nd
8
s
how
the
f
or
e
c
a
s
ted
r
a
inf
a
ll
pr
of
il
e
s
,
f
o
r
e
c
a
s
ted
tempe
r
a
te
pr
of
il
e
s
a
nd
c
or
r
e
s
ponding
c
r
op
de
tails
a
long
with
r
e
leva
nt
mes
s
a
ge
s
to
f
a
r
mer
s
a
nd
a
lt
e
r
na
ti
ve
mea
s
ur
e
f
or
‘
M
ys
or
e
’
c
it
y
f
o
r
ye
a
r
2021
f
o
r
‘
R
a
gi’
c
r
op
.
F
igur
e
4
.
De
s
igned
gr
a
phica
l
us
e
r
in
ter
f
a
c
e
(
GU
I
)
a
ppli
c
a
ti
on
f
or
pr
opos
e
d
wor
k
F
igur
e
5
.
F
or
e
c
a
s
ted
r
a
inf
a
l
l
a
nd
tempe
r
a
tur
e
pr
o
f
i
le
f
or
‘
Ha
s
s
a
n’
c
it
y
f
o
r
2026
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
E
lec
&
C
omp
E
ng
I
S
S
N:
2088
-
8708
A
N
ov
e
l
w
e
ather
par
ame
ter
s
pr
e
diction
s
c
he
me
an
d
their
e
ff
e
c
ts
on
c
r
o
ps
(
L
ingar
aju
N
av
e
e
n
)
645
F
igur
e
6
.
E
f
f
e
c
ts
of
f
or
e
c
a
s
ted
we
a
ther
pa
r
a
mete
r
s
on
‘
M
a
ize
’
c
r
op
a
long
with
mes
s
a
ge
to
f
a
r
mer
s
a
n
d
a
lt
e
r
na
te
mea
s
ur
e
s
f
or
‘
Ha
s
s
a
n’
c
it
y
f
or
2026
F
r
om
r
e
s
ult
s
s
hown
in
F
ig
ur
e
s
5
-
8
,
the
e
f
f
e
c
ts
of
f
o
r
e
c
a
s
ted
we
a
ther
pa
r
a
mete
r
s
on
c
r
ops
a
r
e
we
ll
obvious
.
T
he
e
f
f
e
c
ts
ha
ve
be
e
n
f
ur
ther
e
va
luate
d
a
ga
ins
t
idea
l
c
r
op
c
ondit
ions
a
long
with
the
ge
og
r
a
phica
l
da
ta
a
va
il
a
ble
in
or
de
r
to
c
onve
y
the
r
e
leva
nt
me
s
s
a
ge
to
the
f
a
r
mer
s
a
nd
s
ugge
s
ti
ng
a
lt
e
r
na
te
mea
s
ur
e
to
them
if
ne
e
de
d.
T
a
bles
1
-
4
s
how
s
the
c
ompar
is
on
a
na
lys
is
of
the
pr
opos
e
d
methodology
with
va
r
ious
r
e
leva
nt
e
xis
ti
ng
wor
ks
[
23]
-
[
25
]
in
ter
ms
of
f
e
w
pe
r
f
or
manc
e
pa
r
a
mete
r
s
s
uc
h
a
s
mea
n
s
qua
r
e
e
r
r
o
r
(
M
S
E
)
a
nd
P
e
a
r
s
on
c
or
r
e
lation
c
oe
f
f
icie
nt
(
P
C
C
)
ba
s
e
d
a
c
c
ur
a
c
y
dur
ing
f
o
r
e
c
a
s
t
of
a
ve
r
a
ge
r
a
in
f
a
ll
a
nd
tempe
r
a
tur
e
r
e
s
pe
c
ti
ve
ly
f
or
the
given
f
ive
dis
tr
icts
(
B
a
nga
lor
e
Ur
ba
n,
B
e
lgaum,
Ha
s
s
a
n,
M
ys
or
e
a
nd
T
u
mkur
)
of
Ka
r
n
a
taka
s
tate
a
s
input
f
or
ye
a
r
2020
.
F
r
om
c
o
mpar
is
on
a
na
lys
is
a
nd
gr
a
phica
l
r
e
s
ult
s
,
it
is
quit
e
obvious
that
the
pr
opos
e
d
wor
k
outper
f
o
r
ms
in
e
ve
r
y
a
s
pe
c
t
of
pe
r
f
or
manc
e
a
s
c
ompar
e
d
to
the
e
xis
ti
ng
wor
k
s
.
F
igur
e
7
.
F
or
e
c
a
s
ted
r
a
inf
a
l
l
a
nd
tempe
r
a
tur
e
pr
o
f
i
le
f
or
‘
M
ys
or
e
’
c
it
y
f
or
2021
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
12
,
No.
1
,
F
e
br
ua
r
y
20
22
:
639
-
648
646
F
igur
e
8
.
E
f
f
e
c
ts
of
f
or
e
c
a
s
ted
we
a
ther
pa
r
a
mete
r
s
on
‘
R
a
gi’
c
r
op
a
long
with
mes
s
a
ge
to
f
a
r
mer
s
a
nd
a
lt
e
r
na
te
mea
s
ur
e
s
f
or
‘
M
ys
or
e
’
c
it
y
f
o
r
2021
T
a
ble
1.
C
ompar
is
on
of
M
S
E
f
or
r
a
inf
a
ll
f
or
e
c
a
s
t
M
is
hr
a
e
t
al
.
[
23]
J
e
ongwoo
e
t
al
.
[
24]
K
a
it
li
n
e
t
al
.
[
25]
P
r
opos
e
d W
or
k
B
a
nga
lo
r
e
U
r
ba
n
35.14
31.67
33.58
30.85
B
e
lg
a
um
37.85
29.22
36.28
27.25
H
a
s
s
a
n
30.49
34.67
34.94
28.64
M
ys
or
e
39.5
32.5
37.77
25.71
T
umkur
33.29
36.45
32.78
29.38
T
a
ble
2.
C
ompar
is
on
of
P
C
C
f
or
r
a
inf
a
ll
f
o
r
e
c
a
s
t
M
is
hr
a
e
t
al
.
[
23]
J
e
ongwoo
e
t
al
.
[
24]
K
a
it
li
n
e
t
al
.
[
25]
P
r
opos
e
d W
or
k
B
a
nga
lo
r
e
U
r
ba
n
0.6925
0.7315
0.7098
0.7445
B
e
lg
a
um
0.6695
0.7629
0.6845
0.7871
H
a
s
s
a
n
0.7487
0.7007
0.6992
0.7746
M
ys
or
e
0.6253
0.7188
0.6711
0.7994
T
umkur
0.7134
0.6817
0.7216
0.7605
T
a
ble
3.
C
ompar
is
on
of
M
S
E
f
or
tempe
r
a
tur
e
f
or
e
c
a
s
t
M
is
hr
a
e
t
al
.
[
23]
J
e
ongwoo
e
t
al
.
[
24]
K
a
it
li
n
e
t
al
.
[
25]
P
r
opos
e
d W
or
k
B
a
nga
lo
r
e
U
r
ba
n
34.87
32.65
34.57
28.47
B
e
lg
a
um
36.65
34.92
31.84
30.58
H
a
s
s
a
n
38.18
33.05
33.69
31.48
M
ys
or
e
32.45
34.78
29.95
27.95
T
umkur
35.23
32.38
36.14
30.92
T
a
ble
4.
C
ompar
is
on
of
P
C
C
f
or
tempe
r
a
tur
e
f
or
e
c
a
s
t
M
is
hr
a
e
t
al
.
[
23]
J
e
ongwoo
e
t
al
.
[
24]
K
a
it
li
n
e
t
al
.
[
25]
P
r
opos
e
d W
or
k
B
a
nga
lo
r
e
U
r
ba
n
0.7002
0.7205
0.7167
0.7781
B
e
lg
a
um
0.6813
0.7174
0.7291
0.7479
H
a
s
s
a
n
0.6637
0.7111
0.7052
0.7304
M
ys
or
e
0.7229
0.7183
0.7582
0.7815
T
umkur
0.6946
0.7229
0.6839
0.7443
5.
CONC
L
USI
ON
I
n
thi
s
pa
pe
r
,
a
uthor
ha
s
pr
e
s
e
nted
a
nove
l
s
c
he
m
e
f
or
f
or
e
c
a
s
t
of
the
ke
y
we
a
ther
pa
r
a
mete
r
s
s
uc
h
a
s
tempe
r
a
tur
e
a
nd
r
a
in
f
a
ll
by
making
e
xtens
iv
e
us
e
of
f
it
ne
s
s
f
unc
ti
on
ba
s
e
d
a
da
pti
ve
a
r
ti
f
ici
a
l
ne
ur
a
l
ne
twor
k
f
or
upc
omi
ng
de
c
a
de
(
2021
-
2030)
us
ing
his
tor
ica
l
we
a
ther
da
ta
of
20
dif
f
e
r
e
nt
dis
tr
icts
of
K
a
r
na
taka
s
tate
.
M
or
e
ove
r
,
thes
e
f
or
e
c
a
s
ted
we
a
ther
pa
r
a
met
e
r
s
ha
ve
be
e
n
invol
ve
d
in
o
r
de
r
to
e
va
luate
e
f
f
ici
e
nt
c
r
op
mana
ge
ment
in
ter
ms
of
the
e
f
f
e
c
t
of
f
or
e
c
a
s
ted
we
a
ther
pa
r
a
mete
r
s
on
the
c
r
op
yield,
r
e
lev
a
nt
me
s
s
a
ge
s
to
the
f
a
r
mer
s
a
nd
a
lt
e
r
na
te
mea
s
ur
e
s
s
uc
h
a
s
s
ugge
s
ti
ng
other
ge
ogr
a
phica
l
loca
ti
ons
to
gr
ow
the
s
a
m
e
c
r
op
o
r
gr
owing
other
s
uit
a
ble
c
r
ops
a
t
s
a
me
ge
ogr
a
phica
l
loca
ti
on.
T
he
r
e
s
e
a
r
c
h
is
c
a
r
r
ied
out
ove
r
f
ive
maj
or
c
r
ops
of
Ka
r
na
taka
na
mely
ri
c
e
,
whe
a
t,
jowa
r
,
maiz
e
,
a
nd
r
a
gi.
F
u
r
ther
mor
e
,
th
is
a
ppli
c
a
ti
on
c
a
n
be
de
ve
loped
in
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
E
lec
&
C
omp
E
ng
I
S
S
N:
2088
-
8708
A
N
ov
e
l
w
e
ather
par
ame
ter
s
pr
e
diction
s
c
he
me
an
d
their
e
ff
e
c
ts
on
c
r
o
ps
(
L
ingar
aju
N
av
e
e
n
)
647
ha
r
dwa
r
e
a
nd
s
e
t
up
a
s
a
kios
k
to
f
a
r
mer
s
whe
r
e
they
c
a
n
r
e
ly
by
r
e
duc
ing
the
ne
e
d
o
f
a
g
r
icult
ur
e
a
nd
we
a
ther
e
xpe
r
ts
.
T
his
wor
k
c
a
n
be
e
xtende
d
with
other
we
a
ther
pa
r
a
met
e
r
s
s
uc
h
a
s
wi
nd
s
pe
e
d,
a
nd
r
e
lative
humi
dit
y
with
mul
ti
ple
c
r
ops
a
t
na
ti
ona
l
leve
l
in
or
de
r
to
e
nha
nc
e
c
r
op
yield
a
nd
he
nc
e
boos
t
up
the
I
ndian
e
c
onomy
in
upc
omi
ng
de
c
a
de
.
AC
KNOWL
E
DGE
M
E
NT
S
W
e
would
li
ke
to
e
xpr
e
s
s
s
ince
r
e
thanks
to
M
s
.
S
ha
li
ni
De
e
pa
k,
A
s
s
is
tant
Dir
e
c
tor
o
f
Agr
icultur
e
,
B
a
nga
lor
e
,
I
ndia,
Ka
r
na
taka
S
tate
Na
tur
a
l
Dis
a
s
ter
M
onit
or
ing
C
e
ntr
e
,
B
a
nga
lor
e
,
Dr
.
K.
C
.
Gouda
,
P
r
incipa
l
S
c
ientis
t,
C
S
I
R
F
our
th
P
a
r
a
digm
I
ns
ti
tut
e
(
C
-
M
M
AC
S
)
,
B
a
nga
lor
e
,
I
ndia.
W
e
would
a
ls
o
thank
De
pa
r
tm
e
nt
of
I
S
E
,
S
J
B
I
T
a
nd
Vis
ve
s
va
r
a
ya
T
e
c
hnologi
c
a
l
Unive
r
s
it
y,
B
e
lgaum,
Ka
r
na
taka
f
or
pr
ovidi
ng
e
xtende
d
s
uppor
t
dur
ing
r
e
s
e
a
r
c
h
wor
k.
RE
F
E
RE
NC
E
S
[
1]
S
.
S
.
B
a
boo
a
nd
I
.
K
.
S
he
r
e
e
f
,
“
A
n
e
f
f
ic
ie
nt
w
e
a
th
e
r
f
or
e
c
a
s
ti
ng
s
ys
te
m
us
in
g
a
r
ti
f
ic
ia
l
ne
ur
a
l
ne
twor
k,”
I
nt
e
r
nat
io
nal
J
our
na
l
of
E
nv
ir
onm
e
nt
al
Sc
ie
nc
e
and De
v
e
lo
pm
e
nt
, vol
. 1, no. 4, no. 321
-
326, 2010, doi:
10.7763/I
J
E
S
D
.2010.V1.63
.
[
2]
A
.
S
ma
r
gi
a
s
s
i,
M
.
F
our
ni
e
r
,
C
.
G
r
io
t,
Y
.
B
a
udouin,
a
nd T
.
K
o
s
a
ts
ky,
“
P
r
e
di
c
ti
on
of
th
e
in
door
te
mpe
r
a
tu
r
e
s
of
a
n
ur
ba
n
a
r
e
a
w
it
h
a
n
in
-
ti
me
r
e
gr
e
s
s
io
n
ma
ppi
ng
a
ppr
oa
c
h,
”
J
ou
r
nal
of
E
x
po
s
ur
e
Sc
ie
nc
e
and
E
nv
ir
onm
e
nt
al
E
pi
de
m
io
lo
gy
,
vol
.
18,
no
.
3,
pp. 282
-
288, 2008, doi:
10.1038/s
j.
je
s
.7500588.
[
3]
G
.
G
.
T
ir
une
ha
,
A
.
R
.
F
a
ye
kb,
a
nd
V
.
S
uma
ti
,
“
N
e
ur
o
-
f
uz
z
y
s
ys
te
ms
in
c
ons
tr
uc
ti
on
e
ngi
ne
e
r
in
g
ma
na
ge
me
nt
a
nd
r
e
s
e
a
r
c
h,”
A
ut
om
at
io
n i
n c
ons
tr
uc
ti
on
, vol
. 119, 2020, Ar
t.
no. 103348, doi:
10.1016/j
.a
ut
c
on.2020.103348.
[
4]
S
.
P
a
pa
nt
oni
ou,
D
.
K
ol
okot
s
a
,
a
nd
K
.
K
a
la
it
z
a
ki
s
,
“
B
ui
ld
in
g
o
pt
im
iz
a
ti
on
a
nd
c
ont
r
ol
a
lg
o
r
it
hms
im
pl
e
me
nt
e
d
in
e
xi
s
ti
ng
B
E
M
S
us
in
g
a
w
e
b
ba
s
e
d
e
ne
r
gy
ma
na
ge
me
nt
a
nd
c
ont
r
ol
s
ys
te
m,”
E
ne
r
gy
and
B
ui
ld
in
gs
,
vol
.
98,
pp.
45
-
55,
2014,
doi
:
10.1016/j
.e
nbui
ld
.2014.10.083.
[
5]
A
.
K
a
ur
a
nd
H
.
S
in
gh,
“
A
r
ti
f
ic
ia
l
ne
ur
a
l
n
e
twor
ks
in
f
or
e
c
a
s
ti
ng
ma
xi
mum
a
nd
mi
ni
mum
r
e
la
ti
ve
humi
di
ty
,”
I
nt
e
r
nat
io
nal
J
our
nal
of
C
om
put
e
r
S
c
ie
nc
e
and
N
e
tw
o
r
k
Se
c
u
r
it
y
, vol
. 11, no
. 5, pp. 197
-
199, 2011.
[
6]
Y
.
R
a
dhi
ka
a
nd
M
.
S
ha
s
hi
,
“
A
tm
os
ph
e
r
ic
te
mpe
r
a
tu
r
e
pr
e
di
c
ti
on
us
in
g
s
uppor
t
ve
c
to
r
ma
c
hi
ne
s
,”
I
nt
e
r
nat
io
nal
J
our
na
l
of
C
om
put
e
r
T
he
or
y
and E
ngi
ne
e
r
in
g,
vol
. 1, no. 1, pp. 1793
-
820
1, 2009, doi:
10.7763/I
J
C
T
E
.2009.V1.9.
[
7]
K
.
K
uw
a
ta
a
nd
R
.
S
hi
ba
s
a
ki
,
“
E
s
ti
ma
ti
ng
c
r
op
yi
e
ld
s
w
it
h
d
e
e
p
le
a
r
ni
ng
a
nd
r
e
mot
e
ly
s
e
ns
e
d
d
a
ta
,”
2015
I
E
E
E
I
nt
e
r
nat
io
nal
G
e
os
c
ie
nc
e
and R
e
m
ot
e
Se
ns
in
g S
y
m
pos
iu
m
(
I
G
A
R
SS)
,
2015, p
p. 858
-
861, doi:
10.1109/I
G
A
R
S
S
.2015.7325900.
[
8]
L
.
Z
ja
vka
,
“
N
ume
r
ic
a
l
w
e
a
th
e
r
pr
e
di
c
ti
on
r
e
vi
s
io
n
s
us
in
g
th
e
lo
c
a
ll
y
tr
a
in
e
d
di
f
f
e
r
e
nt
ia
l
pol
ynomi
a
l
ne
twor
k,”
E
x
pe
r
t
Sy
s
te
m
s
w
it
h A
ppl
ic
at
io
ns
,
vol
. 44, pp. 265
-
2
74, 2016, doi:
10.1016/j
.e
s
w
a
.2015.08.057.
[
9]
J
.
W
u,
J
.
L
ong,
a
nd
M
.
L
iu
,
“
E
vol
vi
ng
R
B
F
ne
ur
a
l
ne
twor
k
s
f
or
r
a
in
f
a
ll
pr
e
di
c
ti
on
us
in
g
hybr
id
pa
r
ti
c
le
s
w
a
r
m
opt
im
iz
a
ti
on
a
n
d
ge
ne
ti
c
a
lg
or
it
hm,”
N
e
ur
oc
o
m
put
in
g,
vol
. 148, pp. 136
-
142, 2015, doi:
10.1016/j
.ne
uc
om
.2012.10.043.
[
10]
R
.
A
.
V
e
r
z
ij
lb
e
r
gh,
P
.
H
e
ij
ne
n,
S
.
R
.
d
e
R
oode
,
A
.
L
os
,
a
nd
H
.
J
.
J
.
J
onke
r
,
“
I
mpr
ove
d
mode
l
out
put
s
t
a
ti
s
ti
c
s
of
nume
r
ic
a
l
w
e
a
th
e
r
pr
e
di
c
ti
on
ba
s
e
d
ir
r
a
di
a
nc
e
f
or
e
c
a
s
ts
f
or
s
ol
a
r
po
w
e
r
a
ppl
ic
a
ti
ons
,”
Sol
ar
E
ne
r
gy
,
vol
.
118,
pp.
634
-
645,
2
015
,
doi
:
10.1016/j
.s
ol
e
ne
r
.2015.06.005.
[
11]
K
.
C
.
P
e
r
e
r
a
,
A
.
W
.
W
e
s
te
r
n,
B
.
N
a
w
a
r
a
th
na
,
a
nd
B
.
G
e
or
ge
,
“
F
or
e
c
a
s
ti
ng
da
il
y
r
e
f
e
r
e
nc
e
e
va
pot
r
a
ns
pi
r
a
ti
on
f
or
A
us
tr
a
li
a
u
s
in
g
nume
r
ic
a
l
w
e
a
th
e
r
pr
e
di
c
ti
on
out
put
s
,”
A
gr
ic
ul
tu
r
al
and
F
or
e
s
t
M
e
te
or
ol
ogy
,
vol
.
194,
pp.
50
-
63,
20
14,
doi
:
10.1016/j
.a
gr
f
or
me
t.
2014.03.014.
[
12]
B
.
K
r
ie
s
c
he
,
R
.
H
e
s
s
,
B
.
K
.
R
e
i
c
he
r
t,
a
nd
V
.
S
c
hmi
dt
,
“
A
pr
ob
a
bi
li
s
ti
c
a
ppr
oa
c
h
to
th
e
pr
e
di
c
ti
on
of
a
r
e
a
w
e
a
th
e
r
e
ve
nt
s
,
a
pp
li
e
d
to
pr
e
c
ip
it
a
ti
on,”
Spat
ia
l
St
at
is
ti
c
s
,
vol
. 12, pp. 15
-
30, 2015, doi:
10.1
016/
j.
s
pa
s
ta
.2015.01.002.
[
13]
E
.
P
e
na
ba
d
e
t
al
.,
“
C
ompa
r
a
ti
ve
a
na
ly
s
is
b
e
twe
e
n
op
e
r
a
ti
ona
l
w
e
a
th
e
r
pr
e
di
c
ti
on
mode
l
s
a
nd
Q
ui
kS
C
A
T
w
in
d
d
a
ta
ne
a
r
th
e
G
a
li
c
ia
n c
oa
s
t,
”
J
our
nal
of
M
ar
in
e
S
y
s
te
m
s
,
vol
. 72, no. 1
-
4, pp. 256
-
270, 2008, doi:
10.1016/j
.j
ma
r
s
ys
.2007.07.008
.
[
14]
R
.
B
us
ta
mi
,
N
.
B
e
s
s
a
ih
,
C
.
H
.
J
oo
B
ong
,
“
A
r
ti
f
ic
ia
l
ne
ur
a
l
ne
t
w
or
k
f
or
pr
e
c
ip
it
a
ti
on
a
nd
w
a
te
r
le
ve
l
pr
e
di
c
ti
ons
of
be
d
up
r
iv
e
r
,”
I
A
E
N
G
I
nt
e
r
nat
io
nal
J
our
nal
of
C
om
put
e
r
Sc
ie
nc
e
, vol
. 34, no.
2, pp. 228
-
233, 2007.
[
15]
C
.
V
oya
nt
,
M
.
M
us
e
ll
i,
C
.
P
a
ol
i,
a
nd
M
.
L
.
N
iv
e
t,
“
O
pt
im
iz
a
ti
on
of
a
n
a
r
ti
f
ic
ia
l
ne
ur
a
l
ne
twor
k
de
di
c
a
te
d
to
th
e
mul
t
iv
a
r
ia
te
f
or
e
c
a
s
ti
ng of
da
il
y gl
oba
l
r
a
di
a
ti
on,”
E
ne
r
gy
, vol
. 36, no. 1, pp. 348
-
359, 2011, doi:
10.1016/j
.e
ne
r
gy.2010.10.032.
[
16]
A
.
H
a
s
ni
,
A
.
S
e
hl
i,
B
.
D
r
a
oui
,
A
.
B
a
s
s
ou,
a
nd
B
.
A
mi
e
ur
,
“
E
s
ti
ma
ti
ng
gl
oba
l
s
ol
a
r
r
a
di
a
ti
on
us
in
g
a
r
ti
f
ic
ia
l
ne
ur
a
l
ne
twor
k
a
nd
c
li
ma
te
da
ta
in
th
e
S
out
h
-
W
e
s
te
r
n
r
e
gi
on
of
A
lg
e
r
i
a
,”
E
ne
r
gy
P
r
oc
e
di
a
,
vol
.
18,
pp.
531
-
537,
2012,
doi
:
10.1016/j
.e
gypr
o.2012.05.064
.
[
17]
J
.
L
u,
H
.
H
u,
a
nd
Y
.
B
a
i,
“
R
a
di
a
l
ba
s
is
f
unc
ti
on
ne
ur
a
l
ne
tw
or
k
ba
s
e
d
on
a
n
im
pr
ove
d
e
xpone
nt
ia
l
de
c
r
e
a
s
in
g
in
e
r
ti
a
w
e
i
ght
-
pa
r
ti
c
le
s
w
a
r
m
opt
im
iz
a
ti
on
a
lg
or
it
hm
f
or
A
Q
I
pr
e
di
c
ti
on,”
A
bs
tr
ac
t
and
A
ppl
ie
d
A
nal
y
s
is
,
vol
.
2014,
2014,
A
r
t.
no.
178
313,
doi
:
10.1155/2014/
178313.
[
18]
Z
. Q
. Z
ha
o,
X
. D
. W
u, C
. Y
i
L
u, H
.
G
lo
ti
n, a
nd J
.
G
a
o, “
O
pt
im
iz
in
g w
id
th
s
w
it
h P
S
O
f
or
c
e
nt
e
r
s
e
le
c
ti
on of
G
a
us
s
ia
n r
a
di
a
l
ba
s
is
f
unc
ti
on ne
twor
ks
,”
Sc
ie
nc
e
C
hi
na I
nf
or
m
at
io
n
s
c
ie
nc
e
s
, vol
. 5
7, pp. 1
-
17, 2014, doi:
10.1007/s
11432
-
013
-
4850
-
5.
[
19]
K
.
S
ubr
a
ma
ni
a
n.,
R
.
S
a
vi
th
a
,
a
nd
S
.
S
ur
e
s
h,
“
A
c
ompl
e
x
-
va
lu
e
d
ne
ur
o
-
f
uz
z
y
in
f
e
r
e
nc
e
s
y
s
te
m
a
nd
it
s
le
a
r
ni
ng
me
c
ha
ni
s
m,”
N
e
ur
oc
om
put
in
g
, vol
. 123, pp. 110
-
120, 2014, doi:
10.1016/j
.ne
uc
om.2013.06.009.
[
20]
L
.
N
a
ve
e
n
a
nd
H
.
S
.
M
oh
a
n,
“
H
ig
h
-
r
e
s
ol
ut
io
n
w
e
a
th
e
r
pr
e
di
c
ti
on
us
in
g
modi
f
ie
d
ne
ur
a
l
ne
twor
k
a
ppr
oa
c
h
ove
r
th
e
di
s
tr
ic
ts
of
ka
r
na
ta
ka
s
ta
te
,”
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
C
om
put
e
r
N
e
tw
o
r
k
s
and
C
o
m
m
uni
c
at
io
n
T
e
c
hnol
ogi
e
s
,
vol
.
15,
pp.
125
-
143,
20
18,
doi
:
10.1007/978
-
981
-
10
-
8681
-
6_14.
[
21]
L
.
N
a
ve
e
n
a
nd
H
.
S
.
M
oh
a
n,
“
A
tm
os
phe
r
ic
w
e
a
th
e
r
pr
e
di
c
ti
o
n
us
in
g
a
dva
nc
e
d
te
c
hni
que
s
:
A
s
ur
ve
y,”
2019
3
r
d
I
n
te
r
nat
io
nal
C
onf
.
on C
om
put
in
g M
e
th
odol
ogi
e
s
and C
om
m
uni
c
at
io
n (
I
C
C
M
C
)
,
2019, pp. 440
-
446, doi:
10.1109/I
C
C
M
C
.2019.8819643.
[
22]
L
.
N
a
ve
e
n
a
nd
H
.
S
.
M
oha
n,
“
A
na
ly
z
in
g
im
pa
c
t
of
w
e
a
th
e
r
f
or
e
c
a
s
ti
ng
th
r
ough
de
e
p
le
a
r
ni
ng
in
a
gr
ic
ul
tu
r
a
l
c
r
op
m
ode
l
pr
e
di
c
ti
ons
,”
I
nt
e
r
nat
io
nal
J
our
nal
of
A
ppl
ie
d E
ngi
ne
e
r
in
g R
e
s
e
ar
c
h
, vol
. 14, no. 23, pp. 4379
-
4386, 2019.
[
23]
N
.
M
is
hr
a
,
H
.
K
.
S
oni
,
S
.
S
ha
r
ma
,
a
nd
A
.
K
.
U
pa
dhya
y,
“
D
e
ve
lo
pme
nt
a
nd
a
na
ly
s
is
of
a
r
ti
f
ic
ia
l
ne
ur
a
l
ne
twor
k
mode
ls
f
or
r
a
in
f
a
ll
pr
e
di
c
ti
on
by
us
in
g
ti
me
-
s
e
r
ie
s
d
a
t
a
,”
I
nt
e
r
nat
io
nal
J
ou
r
nal
of
I
nt
e
ll
ig
e
nt
Sy
s
te
m
s
and
A
ppl
ic
at
io
ns
(
I
J
I
SA
)
,
vol
.
10,
no
.
1,
pp. 16
-
23, 2018, doi:
10.5815/i
ji
s
a
.2018.01.03.
[
24]
L
.
J
e
ongwoo,
K
.
C
.
G
yum,
L
.
J
e
ong,
K
.
N
a
m,
a
nd
K
.
H
ye
onj
u
n,
“
A
ppl
ic
a
ti
on
of
a
r
ti
f
ic
ia
l
ne
ur
a
l
ne
twor
ks
to
r
a
in
f
a
ll
f
or
e
c
a
s
t
in
g
in
t
he
ge
um r
iv
e
r
ba
s
in
, kor
e
a
,”
W
at
e
r
, vol
. 10, no. 10, 2018, A
r
t.
no. 1448, doi:
10.3390/w10101448.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
12
,
No.
1
,
F
e
br
ua
r
y
20
22
:
639
-
648
648
[
25]
T
.
K
a
it
li
n,
S
. V
.
A
r
c
hont
ou
l
i
s
,
D
.
R
a
n
a
e
,
P
.
L
a
i
l
a
,
V
.
L
.
A
n
d
y
,
“
H
o
w
d
o
e
s
i
n
c
l
u
s
i
o
n
o
f
w
e
a
t
h
e
r
f
o
r
e
c
a
s
t
i
n
g
i
m
p
a
c
t
i
n
-
s
e
a
s
o
n
c
r
o
p
m
o
d
e
l
p
r
e
d
i
c
t
i
o
n
s
?
,
”
F
i
e
l
d
C
r
o
p
s
R
e
s
e
a
r
c
h
,
v
o
l
.
2
1
4
,
p
p
.
2
6
1
-
272, 2017, doi
:
1016/j
.f
c
r
.2017.09.008.
B
I
OG
RA
P
HI
E
S
OF
AU
T
HO
RS
Li
ng
a
r
a
ju
N
a
v
een
o
b
t
ai
n
ed
h
i
s
Bac
h
el
o
r’s
d
e
g
ree
i
n
In
fo
rma
t
i
o
n
Sci
e
n
ce
an
d
E
n
g
i
n
eeri
n
g
fr
o
m
V
i
s
v
es
v
aray
a
T
ech
n
o
l
o
g
i
ca
l
U
n
i
v
e
rs
i
t
y
(
V
T
U
),
Bel
g
a
u
m,
K
ar
n
at
a
k
a,
In
d
i
a.
T
h
e
n
h
e
o
b
t
ai
n
ed
h
i
s
Mas
t
er’
s
d
e
g
ree
i
n
C
o
mp
u
t
er
Sci
e
n
ce
an
d
E
n
g
i
n
eeri
n
g
fro
m
V
T
U
.
Cu
rren
t
l
y
,
h
e
i
s
p
u
r
s
u
i
n
g
h
i
s
Ph
.
D
.
i
n
V
T
U
.
H
i
s
res
earc
h
i
n
t
ere
s
t
s
are
i
n
n
eu
ra
l
n
et
w
o
r
k
s
,
art
i
f
i
ci
a
l
i
n
t
e
l
l
i
g
e
n
ce,
d
at
a
mi
n
i
n
g
,
b
i
g
d
at
a,
w
i
re
l
es
s
s
en
s
o
r
an
d
cel
l
u
l
ar
n
e
t
w
o
rk
s
.
H
e
h
as
p
u
b
l
i
s
h
e
d
an
d
p
re
s
en
t
ed
p
a
p
ers
i
n
j
o
u
rn
a
l
s
,
i
n
t
er
n
at
i
o
n
al
an
d
n
at
i
o
n
al
l
ev
e
l
co
n
fere
n
ces
.
H
e
can
b
e
co
n
t
ac
t
ed
at
ema
i
l
:
n
av
e
en
l
i
n
g
ar
aj
u
@
g
ma
i
l
.
co
m.
Ho
s
a
a
g
ra
ha
ra
S
a
v
a
l
e
g
o
w
da
M
o
h
a
n
recei
v
ed
h
i
s
Bach
e
l
o
r’
s
d
e
g
ree
i
n
co
mp
u
t
er
Sci
en
ce
an
d
E
n
g
i
n
eeri
n
g
fro
m
Mal
n
a
d
co
l
l
eg
e
o
f
E
n
g
i
n
eeri
n
g
,
H
as
s
an
d
u
ri
n
g
t
h
e
y
ear
1
9
9
9
,
M
.
T
ech
i
n
co
m
p
u
t
er
Sci
en
ce
a
n
d
E
n
g
i
n
eer
i
n
g
fr
o
m
J
aw
a
h
arl
a
l
N
eh
r
u
N
at
i
o
n
al
Co
l
l
eg
e
o
f
E
n
g
i
n
eer
i
n
g
,
Sh
i
m
o
g
a
d
u
ri
n
g
t
h
e
y
ear
2
0
0
4
an
d
Ph
.
D
.
i
n
Co
mp
u
t
er
Sc
i
en
ce
an
d
E
n
g
i
n
eeri
n
g
fr
o
m
D
r.
MG
R
U
n
i
v
ers
i
t
y
C
h
en
n
ai
.
H
e
i
s
w
o
r
k
i
n
g
a
s
a
Pr
o
fes
s
o
r
a
n
d
H
ead
i
n
t
h
e
D
e
p
art
m
e
n
t
o
f
In
f
o
rmat
i
o
n
Sci
e
n
ce
an
d
E
n
g
i
n
eeri
n
g
at
S
J
B
In
s
t
i
t
u
t
e
o
f
T
ech
n
o
l
o
g
y
,
Ben
g
a
l
u
r
u
-
6
0
.
H
e
i
s
h
av
i
n
g
t
o
t
a
l
2
1
y
ears
o
f
t
each
i
n
g
ex
p
eri
en
ce.
H
i
s
area
o
f
i
n
t
eres
t
s
i
s
n
et
w
o
r
k
s
s
ec
u
ri
t
y
,
i
ma
g
e
p
ro
ces
s
i
n
g
,
d
at
a
s
t
r
u
ct
u
res
,
co
mp
u
t
er
g
rap
h
i
c
s
,
fin
i
t
e
au
t
o
mat
a
an
d
fo
rmal
l
a
n
g
u
ag
e
s
,
co
mp
i
l
er
d
e
s
i
g
n
.
H
e
h
a
s
o
b
t
a
i
n
e
d
a
b
es
t
t
eac
h
er
aw
ar
d
fo
r
h
i
s
t
eac
h
i
n
g
d
u
r
i
n
g
t
h
e
y
ear
2
0
0
8
a
t
SJ
BI
T
Be
n
g
a
l
u
r
u
-
6
0
.
H
e
h
as
p
u
b
l
i
s
h
ed
a
n
d
p
res
e
n
t
e
d
p
a
p
ers
i
n
j
o
u
rn
a
l
s
,
i
n
t
ern
at
i
o
n
al
an
d
n
at
i
o
n
al
l
ev
e
l
co
n
fere
n
ce
s
.
H
e
can
b
e
co
n
t
ac
t
ed
at
e
ma
i
l
:
mo
h
a
n
_
k
i
t
@
y
ah
o
o
.
co
m
.
Evaluation Warning : The document was created with Spire.PDF for Python.