I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
pu
t
er
Science
Vo
l.
23
,
No
.
2
,
A
u
g
u
s
t
2
0
2
1
,
p
p
.
6
86
~
6
9
3
I
SS
N:
2
5
0
2
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/ijeecs.v
23
.i
2
.
pp
6
86
-
6
9
3
686
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
O
ptimi
za
tion o
f
l
ea
rning
alg
o
rithms in mul
tilay
er
percept
ro
n
for
sheet
resista
n
ce of r
e
duced
gra
phene o
x
ide
thin
-
film
No
o
r
Aim
a
n bin
Am
inu
dd
in
1
,
Nurla
ila
I
s
m
a
il
2
,
M
a
ria
na
h M
a
s
rie
3
,
Siti
Ai
s
ha
h M
o
ha
m
a
d B
a
da
rudd
in
4
1
-
3
C
o
lleg
e
o
f
E
n
g
in
ee
r
in
g
,
Un
i
v
er
s
iti T
ek
n
o
lo
g
i M
AR
A
(
UiT
M)
,
Ma
lay
s
ia
4
M
IM
OS
Be
rh
a
d
,
Tec
h
n
o
l
o
g
y
P
a
rk
M
a
lay
sia
,
K
u
a
la L
u
m
p
u
r,
M
a
l
a
y
sia
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ma
y
2
0
,
2
0
2
1
R
ev
is
ed
J
u
l
5
,
2
0
2
1
Acc
ep
ted
J
u
l
14
,
2
0
2
1
M
u
lt
il
a
y
e
r
p
e
rc
e
p
tro
n
(M
L
P
)
o
p
ti
m
iza
ti
o
n
is
c
a
rried
o
u
t
to
in
v
e
stig
a
te
th
e
c
las
sifier'
s
p
e
rfo
rm
a
n
c
e
in
d
isc
rimin
a
ti
n
g
th
e
u
n
if
o
rm
it
y
o
f
re
d
u
c
e
d
G
ra
p
h
e
n
e
Ox
id
e
(
r
G
O)
th
i
n
-
fil
m
sh
e
e
t
re
sista
n
c
e
.
Th
is
stu
d
y
u
se
d
t
h
re
e
lea
rn
in
g
a
l
g
o
ri
th
m
s:
re
sili
e
n
t
b
a
c
k
p
ro
p
a
g
a
ti
o
n
(RP
)
,
sc
a
led
c
o
n
ju
g
a
te
g
ra
d
ien
t
(S
CG
)
a
n
d
lev
e
n
b
e
rg
-
m
a
rq
u
a
rd
t
(LM
).
Th
e
d
a
tas
e
t
u
se
d
in
th
is
stu
d
y
is
t
h
e
sh
e
e
t
re
sista
n
c
e
o
f
r
G
O
th
in
fil
m
s
o
b
tain
e
d
fro
m
M
I
M
OS
Bh
d
.
Th
is
wo
rk
i
n
v
o
lv
e
d
sa
m
p
les
se
le
c
ti
o
n
fro
m
a
u
n
if
o
rm
a
n
d
n
o
n
-
u
n
ifo
rm
r
G
O
th
in
-
f
il
m
sh
e
e
t
re
sist
a
n
c
e
.
Th
e
i
n
p
u
t
a
n
d
o
u
tp
u
t
d
a
ta
we
re
u
n
d
e
r
g
o
i
n
g
d
a
ta
p
re
-
p
ro
c
e
ss
in
g
:
d
a
ta
n
o
rm
a
li
z
a
ti
o
n
,
d
a
ta
ra
n
d
o
m
iza
ti
o
n
,
a
n
d
d
a
ta
sp
li
tt
i
n
g
.
Th
e
d
a
ta
we
re
d
i
v
i
d
e
d
in
t
o
th
re
e
g
ro
u
p
s;
train
i
n
g
,
v
a
li
d
a
ti
o
n
a
n
d
t
e
stin
g
with
a
ra
ti
o
o
f
7
0
%
:
1
5
%
:
1
5
%
,
re
sp
e
c
ti
v
e
ly
.
A
v
a
ry
i
n
g
n
u
m
b
e
r
o
f
h
i
d
d
e
n
n
e
u
ro
n
s
o
p
ti
m
ize
d
th
e
lea
rn
in
g
a
l
g
o
ri
th
m
s
in
M
LP
fro
m
1
to
1
0
.
Th
e
ir
b
e
h
a
v
io
r
h
e
lp
e
d
e
sta
b
li
sh
th
e
b
e
st
lea
rn
in
g
a
lg
o
rit
h
m
s
in
d
isc
rimin
a
ti
n
g
M
L
P
fo
r
r
G
O
sh
e
e
t
re
sista
n
c
e
u
n
ifo
rm
it
y
.
Th
e
p
e
rfo
rm
a
n
c
e
s
m
e
a
su
re
d
we
re
th
e
a
c
c
u
ra
c
y
o
f
train
i
n
g
,
v
a
li
d
a
ti
o
n
a
n
d
tes
ti
n
g
d
a
tas
e
t,
m
e
a
n
sq
u
a
re
d
e
rro
rs
(
M
S
E)
a
n
d
e
p
o
c
h
s.
Al
l
t
h
e
a
n
a
ly
t
ica
l
wo
rk
i
n
t
h
is
stu
d
y
wa
s
a
c
h
iev
e
d
a
u
to
m
a
ti
c
a
ll
y
v
ia
M
ATLAB
so
ftwa
re
v
e
rsio
n
R2
0
1
8
a
.
It
wa
s
fo
u
n
d
th
a
t
t
h
e
LM
is
d
o
m
in
a
n
t
i
n
th
e
o
p
ti
m
iza
ti
o
n
o
f
a
lea
rn
i
n
g
a
lg
o
rit
h
m
i
n
M
L
P
f
o
r
r
G
O
sh
e
e
t
re
sista
n
c
e
.
Th
e
M
S
E
fo
r
LM
is
th
e
m
o
st r
e
d
u
c
e
d
a
m
id
S
CG
a
n
d
RP
.
K
ey
w
o
r
d
s
:
I
m
ag
e
class
if
icatio
n
LM
MLP
R
ed
u
ce
d
g
r
a
p
h
en
e
o
x
id
e
R
P
SC
G
Sh
ee
t r
esis
tan
ce
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
:
Ma
r
ian
ah
Ma
s
r
ie
S
ch
o
o
l o
f
E
lectr
ical
E
n
g
in
ee
r
in
g
C
o
lleg
e
o
f
E
n
g
i
n
ee
r
in
g
Un
iv
er
s
iti T
ek
n
o
lo
g
i M
AR
A,
4
0
4
5
0
,
Sh
ah
Alam
,
Selan
g
o
r
,
Ma
lay
s
ia
E
m
ail: m
ar
ian
ah
@
u
itm
.
ed
u
.
m
y
1.
I
NT
RO
D
UCT
I
O
N
Gr
ap
h
en
e
co
n
s
is
ts
o
f
o
n
e
lay
er
o
f
ca
r
b
o
n
ato
m
s
o
r
g
an
ize
d
in
a
v
er
y
h
o
n
e
y
co
m
b
p
atter
n
an
d
m
a
y
ev
en
b
e
d
elin
ea
te
d
as
a
o
n
e
-
ato
m
-
th
ick
lay
er
o
f
g
r
ap
h
ite
[1
]
-
[
4]
.
Gr
a
p
h
en
e
is
r
eliab
le
to
b
e
an
elec
tr
ical
co
n
d
u
ct
o
r
f
o
r
j
u
s
t
o
n
e
at
o
m
th
ick
f
o
r
r
em
ai
n
s
lig
h
t,
f
le
x
i
b
le
an
d
tr
an
s
p
ar
e
n
t
[5
]
-
[
7]
.
T
h
e
m
o
s
t
r
ec
o
g
n
ize
tech
n
iq
u
e
h
as
b
ee
n
d
e
v
elo
p
e
d
to
cr
ea
te
lar
g
e
-
s
ca
le
co
n
ti
n
u
o
u
s
g
r
ap
h
e
n
e
f
ilm
s
s
u
ch
a
s
ch
em
ical
v
ap
o
u
r
d
ep
o
s
itio
n
(
C
VD)
,
g
r
a
p
h
en
e
ep
itax
ial
g
r
o
wth
o
n
s
ilico
n
ca
r
b
id
e
(
SiC
)
an
d
g
r
ap
h
en
e
o
x
id
e
r
e
d
u
ctio
n
.
Am
o
n
g
s
t
th
ese
ap
p
r
o
ac
h
es,
th
e
o
x
id
e
r
ed
u
ctio
n
p
r
o
v
e
d
to
b
e
a
p
r
ac
tical
ap
p
r
o
ac
h
to
p
r
o
d
u
c
e
g
r
a
p
h
en
e
at
a
r
elativ
ely
lo
w
co
s
t w
ith
o
p
tim
al
q
u
ality
.
Gr
ap
h
e
n
e
O
x
id
e
(
G
O)
is
id
en
tifie
d
as a
n
elec
tr
ical
in
s
u
lato
r
with
lo
w
th
er
m
al
co
n
d
u
ctiv
ity
d
u
e
to
th
e
d
is
r
u
p
tio
n
o
f
its
s
p
2
b
o
n
d
i
n
g
n
etwo
r
k
s
.
T
o
r
ec
o
v
er
th
e
h
ex
ag
o
n
al
h
o
n
ey
c
o
m
b
lattice
an
d
elec
tr
ical
c
o
n
d
u
ctiv
ity
,
th
e
r
GO
m
u
s
t
b
e
p
r
o
d
u
c
ed
in
a
h
ig
h
-
tem
p
e
r
atu
r
e
v
ac
u
u
m
ch
a
m
b
er
with
a
ce
r
tain
d
eg
r
ee
o
f
tem
p
er
atu
r
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
Op
timiz
a
tio
n
o
f le
a
r
n
in
g
a
l
g
o
r
ith
ms in
mu
ltil
a
ye
r
p
ercep
tr
on
fo
r
…
(
N
o
o
r
A
ima
n
b
i
n
A
mi
n
u
d
d
in
)
687
ML
P
is
o
n
e
o
f
th
e
p
r
ef
er
r
ed
m
eth
o
d
s
u
s
ed
f
o
r
th
e
class
if
icatio
n
an
d
p
r
ed
ictio
n
o
f
n
an
o
m
ater
ials
p
r
o
p
er
ties
s
u
ch
as
th
in
f
ilm
s
,
n
an
o
f
lu
id
s
,
n
a
n
o
f
i
b
er
,
a
n
d
n
an
o
co
m
p
o
s
ites
r
ep
o
r
ted
in
t
h
e
p
r
ev
io
u
s
r
esear
ch
.
Kh
o
s
r
o
jer
d
i
et
a
l.
[
8
]
p
r
ed
icte
d
a
th
er
m
al
co
n
d
u
ctiv
ity
o
f
g
r
ap
h
en
e
n
an
o
f
lu
id
u
s
in
g
th
e
m
u
ltil
ay
er
p
e
r
ce
p
tr
o
n
(
ML
P)
o
f
a
n
ar
tific
ial
n
e
u
r
al
n
etwo
r
k
.
M
o
d
el
ac
cu
r
ac
y
was
ev
alu
ated
u
s
in
g
s
q
u
a
r
e
m
ea
n
q
u
ad
r
atu
r
e
(
R
MS)
in
d
ex
es
.
T
h
e
ANN
al
g
o
r
ith
m
was
u
s
ed
to
m
o
d
el
C
d
(
I
I
)
eli
m
in
atio
n
e
f
f
icien
cy
an
d
o
p
tim
ize
p
r
o
ce
s
s
v
ar
iab
les
o
f
C
d
(
I
I
)
c
o
n
ce
n
t
r
atio
n
,
in
itia
l
p
H
v
alu
es,
co
n
tact
tim
es
an
d
o
p
er
atin
g
tem
p
er
at
u
r
es
[
9
]
.
Am
an
i
et
a
l.
[
1
0
]
p
er
f
o
r
m
ed
m
u
lti
-
cr
iter
ia
m
o
d
elin
g
an
d
o
p
tim
izatio
n
o
f
th
e
r
h
eo
lo
g
ical
an
d
t
h
er
m
o
p
h
y
s
i
ca
l
p
r
o
p
e
r
ties
o
f
an
en
v
ir
o
n
m
en
tally
-
f
r
ien
d
ly
co
v
alen
tly
f
u
n
ctio
n
alize
d
n
an
o
f
l
u
id
co
n
tain
in
g
g
r
ap
h
en
e
n
a
n
o
p
latelets
(
C
GN
Ps
)
.
T
h
e
Nar
x
-
ANN
m
ath
em
atica
l
m
o
d
el
was
d
ev
elo
p
ed
t
o
s
h
if
t
th
e
q
u
ar
tz
r
eso
n
at
o
r
'
s
f
r
eq
u
en
cy
s
h
if
t
o
n
GO
lan
g
m
u
ir
b
lad
g
ett
t
h
in
-
f
ilm
s
[
1
1
]
.
T
h
e
a
p
p
licatio
n
o
f
ANN
to
th
e
class
if
icatio
n
an
d
p
r
ed
ictio
n
o
f
g
r
ap
h
en
e
n
an
o
m
ater
ial
is
v
er
y
m
in
im
al
,
b
u
t
it
is
ex
ten
s
iv
e
f
o
r
o
th
er
s
[
1
2
]
.
G
u
o
et
a
l.
[
1
3
]
r
e
p
o
r
ted
ten
m
u
ltil
ay
er
p
er
ce
p
tr
o
n
s
wer
e
in
te
g
r
ated
in
to
a
r
an
d
o
m
f
o
r
est
an
d
m
u
ltil
ay
er
p
er
ce
p
tr
o
n
(
R
F
-
ML
P)
m
o
d
el
u
s
in
g
th
e
r
an
d
o
m
f
o
r
est
(
R
F)
m
eth
o
d
f
o
r
p
r
ed
ictin
g
th
e
d
ielec
tr
ic
lo
s
s
o
f
p
o
ly
im
id
e
n
an
o
co
m
p
o
s
ite
f
ilm
s
.
T
h
ey
also
ap
p
lied
th
e
ML
P.
A
m
u
ltil
ay
e
r
p
er
ce
p
tr
o
n
an
d
a
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
i
n
e
(
SVM)
b
ased
o
n
a
PUK
k
er
n
el
wer
e
u
s
ed
to
class
if
y
b
o
th
th
e
s
in
g
l
e
-
lay
er
an
d
th
r
ee
-
lay
er
p
o
l
y
im
id
e
n
an
o
co
m
p
o
s
ite
f
ilm
s
[
1
4
]
.
Ko
n
o
m
i
et
a
l.
[
1
5
]
h
av
e
d
e
v
elo
p
e
d
a
n
o
v
el
m
eth
o
d
to
ch
ar
ac
ter
ize
th
i
n
-
f
ilm
c
o
n
d
u
ctiv
ity
in
E
FM
b
ased
o
n
f
ee
d
-
f
o
r
war
d
n
eu
r
a
l
n
etwo
r
k
s
an
d
ev
o
lu
tio
n
ar
y
al
g
o
r
ith
m
s
.
ML
P
h
as
also
b
ee
n
co
n
d
u
cte
d
to
p
r
ed
ict
th
e
o
p
tical
p
r
o
p
er
ties
o
f
Plas
m
o
n
ic
th
in
-
f
ilm
s
o
lar
ce
l
ls
an
d
o
p
tim
ize
th
eir
s
tr
u
ct
u
r
es
[
1
6
]
.
T
h
e
ML
P
is
also
ap
p
lied
to
p
r
ed
ict
t
h
e
ef
f
icien
cy
o
f
a
d
o
u
b
le
-
walled
r
ea
cto
r
u
s
in
g
n
an
o
f
l
u
id
s
a
s
h
ea
t
tr
an
s
f
er
an
d
in
p
r
ed
ic
tin
g
th
e
n
a
n
o
f
lu
i
d
s
r
elativ
e
v
is
co
s
ity
[
1
7
]
,
[
18]
.
H
ass
an
et
a
l.
[
1
9
]
h
av
e
d
ev
elo
p
ed
a
m
o
d
el
b
ased
o
n
th
e
p
r
ed
i
ctio
n
o
f
R
-
s
q
u
ar
ed
v
alu
e
th
at
ca
n
b
e
im
p
lem
en
te
d
to
esti
m
ate
th
e
v
alu
es
o
f
s
p
ec
if
ic
h
ea
t
ca
p
ac
ity
f
o
r
n
an
o
f
lu
id
s
s
am
p
les.
Fo
r
n
an
o
f
ib
e
r
m
ater
ials
,
th
e
ML
P
-
b
ased
ANN
m
o
d
el
was
u
s
ed
to
p
r
ed
ict
th
e
m
ea
n
d
iam
eter
o
f
t
h
e
elec
tr
o
s
p
u
n
f
ib
er
[
2
0
]
,
[
21]
.
A
p
a
r
t
f
r
o
m
t
h
i
s
,
tw
o
A
NN
m
o
d
e
ls
h
a
v
e
b
e
en
d
e
v
e
l
o
p
e
d
t
o
m
o
d
e
l
t
h
e
e
li
m
i
n
a
t
i
o
n
e
f
f
i
ci
e
n
c
y
o
f
n
a
n
o
m
a
t
e
r
i
a
ls
h
e
a
v
y
m
e
t
a
l
s
an
d
t
h
e
e
s
t
i
m
a
t
i
o
n
o
f
c
h
e
m
i
c
a
l
m
a
t
e
r
i
a
l
a
d
s
o
r
p
t
i
o
n
o
n
n
a
n
o
c
o
m
p
o
s
i
t
e
[
2
2
]
,
[
23]
.
A
c
c
o
r
d
i
n
g
t
o
p
r
e
v
i
o
u
s
r
es
e
a
r
ch
,
t
h
e
r
e
is
n
o
t
y
e
t
r
e
p
o
r
t
o
n
t
h
e
u
s
e
o
f
i
n
t
el
l
i
g
e
n
t
c
o
m
p
u
t
i
n
g
a
n
d
n
e
u
r
a
l
n
e
tw
o
r
k
s
t
e
c
h
n
o
l
o
g
y
t
o
c
l
as
s
i
f
y
n
a
n
o
m
a
t
e
r
i
a
l
t
h
i
n
-
f
i
l
m
s
h
e
et
r
e
s
i
s
t
a
n
c
e
u
n
i
f
o
r
m
i
t
y
.
T
h
is
s
t
u
d
y
h
a
s
g
i
v
e
n
r
i
s
e
t
o
n
e
w
c
h
a
l
l
e
n
g
e
s
i
n
t
h
e
f
i
e
l
d
o
f
n
a
n
o
m
a
t
e
r
i
a
l
s
h
ee
t
r
e
s
is
ta
n
c
e
.
ML
P
c
a
n
p
r
o
v
i
d
e
t
h
e
b
e
s
t
m
o
d
e
l
b
e
h
a
v
i
o
u
r
f
o
r
a
u
n
i
f
o
r
m
s
h
e
e
t
r
e
s
i
s
t
a
n
c
e
.
T
h
is
s
tu
d
y
p
r
o
p
o
s
es
th
r
ee
lea
r
n
in
g
alg
o
r
ith
m
s
o
f
m
u
ltil
ay
e
r
p
er
ce
p
tr
o
n
(
ML
P)
class
if
ier
:
r
esil
ien
t
b
ac
k
p
r
o
p
ag
ati
o
n
(
R
P),
s
ca
le
co
n
ju
g
ate
g
r
ad
ien
t
(
SC
G)
an
d
lev
en
b
er
g
-
m
ar
q
u
ar
d
t
(
L
M)
f
o
r
p
r
o
ce
s
s
m
o
d
elin
g
an
d
ac
c
o
m
p
lis
h
in
g
o
p
tim
al
c
o
atin
g
p
ar
am
eter
s
f
o
r
in
v
esti
g
atin
g
t
h
e
n
a
n
o
m
ater
ial
th
in
f
ilm
p
r
o
p
er
ty
.
T
h
e
alg
o
r
ith
m
s
h
av
e
b
ee
n
d
e
v
elo
p
ed
to
o
p
tim
i
ze
th
e
u
n
if
o
r
m
ity
o
f
r
GO
th
in
-
f
ilm
s
h
ee
t
r
esis
tan
ce
.
T
h
e
r
GO
s
h
ee
t
r
esis
tan
ce
d
atasets
wer
e
ac
q
u
ir
ed
f
r
o
m
th
e
p
r
e
v
io
u
s
r
esear
c
h
er
in
MI
MO
S
B
er
h
ad
.
T
h
e
d
ata
wer
e
p
r
o
ce
s
s
ed
b
ef
o
r
e
h
an
d
an
d
th
e
d
atasets
w
er
e
u
s
ed
in
th
r
ee
p
h
ases
:
tr
ain
in
g
,
v
alid
atio
n
an
d
test
in
g
.
T
h
e
p
r
o
ce
s
s
co
n
tin
u
es
with
th
e
d
ev
elo
p
m
en
t o
f
th
e
ML
P m
o
d
el
th
r
o
u
g
h
R
P,
L
M
an
d
SC
G.
T
h
en
,
all
th
r
ee
m
o
d
els d
ev
elo
p
ed
m
o
d
els
wer
e
test
ed
an
d
ac
ce
p
ted
o
n
c
e
ea
ch
m
o
d
el
m
et
p
er
f
o
r
m
a
n
c
e
cr
iter
ia.
Fin
ally
,
th
e
r
esu
lts
o
b
tain
ed
h
av
e
also
b
ee
n
v
alid
ated
ex
p
er
i
m
en
tally
.
2.
RE
S
E
ARCH
M
E
T
H
O
D
T
h
e
ex
p
er
im
en
tal
s
etu
p
f
o
r
t
h
e
o
p
tim
izatio
n
o
f
lear
n
in
g
alg
o
r
ith
m
s
is
d
ep
icted
in
Fig
u
r
e
1
.
T
h
e
p
r
o
ce
s
s
b
eg
in
s
with
a
d
ata
co
llectio
n
f
r
o
m
MI
MO
S
B
er
h
ad
.
T
h
e
m
eth
o
d
o
f
p
r
o
d
u
cin
g
r
GO
s
h
ee
t
r
esis
tan
ce
d
atasets
was
s
tar
ted
with
th
e
s
p
r
ay
o
f
g
r
ap
h
e
n
e
o
x
id
e
(
GO)
with
3
x
an
d
4
x
s
p
r
a
y
p
ass
es
o
n
s
ilico
n
d
io
x
id
e
(
SiO2
)
waf
er
b
y
u
s
in
g
an
ato
m
izer
s
y
s
tem
d
ev
elo
p
ed
b
y
M
I
MO
S
B
er
h
ad
[
2
4
]
.
T
h
e
p
r
o
ce
s
s
was
r
ep
ea
ted
f
o
r
f
iv
e
r
u
n
s
o
f
th
e
ex
p
er
im
en
t.
T
h
e
GO
s
am
p
les
wer
e
th
en
r
ed
u
ce
d
th
r
o
u
g
h
a
h
ig
h
tem
p
e
r
atu
r
e
o
f
th
e
th
er
m
al
r
ed
u
ctio
n
p
r
o
ce
s
s
to
p
r
o
d
u
ce
r
GO
s
am
p
les
[
2
5
]
.
T
h
e
f
o
u
r
-
p
o
in
t
m
ac
h
in
e
m
ea
s
u
r
ed
elec
tr
ical
co
n
d
u
ctiv
ity
,
wh
ich
is
th
e
s
h
ee
t r
e
s
is
tan
ce
a
t 4
9
d
if
f
er
en
t c
o
o
r
d
in
ate
p
o
i
n
ts
d
is
tr
ib
u
ted
r
ad
ially
f
r
o
m
th
e
ce
n
ter
o
f
th
e
wh
o
le
8
-
in
ch
SiO2
waf
er
s
h
o
wn
in
Fig
u
r
e
2
r
ig
h
t
af
ter
th
e
r
e
d
u
ctio
n
p
r
o
ce
s
s
.
T
h
e
f
ig
u
r
e
also
illu
s
tr
ate
s
th
e
d
is
tr
ib
u
tio
n
o
f
th
e
s
h
ee
t r
esis
tan
ce
v
alu
es f
o
r
4
x
s
p
r
ay
p
ass
es
.
T
h
e
d
atasets
wer
e
u
n
d
er
g
o
in
g
d
ata
p
r
e
-
p
r
o
ce
s
s
in
g
wh
er
e
7
0
%
was
u
s
ed
f
o
r
tr
ain
in
g
,
1
5
%
f
o
r
v
alid
atio
n
a
n
d
th
e
r
em
ai
n
in
g
d
ata,
1
5
%
was
u
s
ed
f
o
r
test
in
g
.
T
h
e
p
r
o
ce
s
s
co
n
tin
u
ed
with
th
e
m
u
ltil
ay
er
p
er
ce
p
tr
o
n
(
ML
P)
tr
ai
n
in
g
u
s
in
g
two
d
i
f
f
er
en
t d
atasets
tr
ain
ed
s
ep
ar
ately
.
I
n
th
is
p
r
o
ce
s
s
,
lear
n
in
g
alg
o
r
ith
m
s
wer
e
v
ar
ied
,
wh
ich
in
clu
d
es
t
r
ain
in
g
u
s
in
g
R
P,
L
M
an
d
S
C
G.
T
h
e
n
eu
r
o
n
s
in
th
e
h
id
d
e
n
lay
er
wer
e
v
a
r
ied
u
s
in
g
p
atter
n
r
ec
o
g
n
itio
n
n
etw
o
r
k
(
p
a
tter
n
et
)
f
u
n
ctio
n
in
M
AT
L
AB
R
2
0
1
8
a,
s
et
with
1
to
1
0
h
id
d
e
n
n
e
u
r
o
n
s
.
T
h
en
it
was
f
o
llo
we
d
b
y
th
e
v
alid
atio
n
an
d
test
in
g
o
f
th
e
tr
ain
ed
n
etwo
r
k
f
o
r
ea
c
h
lear
n
in
g
alg
o
r
ith
m
.
T
o
ac
ce
p
t
th
e
d
ev
el
o
p
ed
ML
P
m
o
d
el,
th
e
f
o
llo
win
g
p
er
f
o
r
m
an
ce
was
m
ea
s
u
r
ed
b
y
t
h
e
cr
iter
ia
o
f
th
e
co
n
f
u
s
io
n
m
atr
ix
,
ac
cu
r
ac
y
,
s
en
s
itiv
ity
,
s
p
ec
if
icity
an
d
p
r
ec
is
io
n
th
at
a
p
p
ea
r
ed
in
n
eu
r
al
n
etwo
r
k
tr
ai
n
in
g
(
n
n
tr
a
in
to
o
l)
.
T
h
e
m
o
d
el
was
ac
ce
p
ted
if
th
e
m
o
d
el
p
ass
ed
.
B
u
t
if
n
o
t,
it
en
d
u
r
e
d
th
e
d
ata
p
r
o
ce
s
s
in
g
p
r
ec
is
e
to
eith
er
th
r
ee
p
r
o
ce
s
s
es
as
s
h
o
wn
in
Fig
u
r
e
1
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
23
,
No
.
2
,
Au
g
u
s
t 2
0
2
1
:
6
86
-
6
9
3
688
As
s
h
o
wn
in
Fig
u
r
e
3
,
o
n
e
ca
n
s
ee
th
e
ar
ch
itectu
r
e
o
f
m
u
ltil
ay
er
p
er
ce
p
tr
o
n
(
ML
P)
with
in
p
u
t,
h
id
d
en
an
d
o
u
tp
u
t
lay
e
r
s
.
T
h
e
p
r
o
ce
s
s
s
tar
ts
f
r
o
m
th
e
f
ir
s
t
lay
er
tak
in
g
in
in
p
u
ts
an
d
t
h
e
l
ast
lay
er
p
r
o
d
u
cin
g
o
u
tp
u
t.
I
n
th
e
m
id
d
le
o
f
th
e
la
y
er
is
a
h
id
d
en
lay
er
.
T
h
e
in
p
u
t
lay
er
h
as
4
9
n
eu
r
o
n
s
,
r
ep
r
e
s
en
tin
g
4
9
p
o
i
n
ts
o
n
r
GO
s
h
ee
t
r
esis
tan
ce
b
y
r
ea
d
i
n
g
th
e
4
-
p
o
i
n
t
p
r
o
b
e
m
ac
h
in
e.
Af
ter
th
at
h
id
d
en
lay
e
r
is
o
p
ti
m
ized
f
r
o
m
1
to
1
0
n
eu
r
o
n
s
an
d
th
e
o
u
tp
u
t
lay
er
h
as
o
n
e
n
eu
r
o
n
is
r
ep
r
esen
tin
g
th
e
u
n
if
o
r
m
ity
o
f
th
e
s
h
ee
t
r
esis
tan
ce
.
E
ac
h
p
er
ce
p
tr
o
n
is
co
n
n
ec
te
d
to
ev
er
y
p
er
ce
p
tr
o
n
o
n
th
e
n
ex
t
la
y
er
.
So
th
at
th
e
i
n
f
o
r
m
atio
n
i
s
co
n
s
tan
tly
"f
ee
d
-
f
o
r
war
d
"
f
r
o
m
o
n
e
la
y
er
to
th
e
n
ex
t; h
en
ce
th
is
n
etwo
r
k
is
al
s
o
ca
lled
a
f
o
r
war
d
f
ee
d
n
etw
o
r
k
.
Fig
u
r
e
1
.
Flo
wch
ar
t
o
f
th
e
ex
p
er
im
en
tal
s
et
-
u
p
in
m
o
d
elin
g
f
o
r
r
GO
s
h
ee
t r
esis
tan
ce
class
if
ier
Fig
u
r
e
2
.
T
h
e
co
n
to
u
r
p
lo
ts
f
o
r
s
h
ee
t r
esis
tan
ce
d
is
tr
ib
u
tio
n
o
n
4
9
p
o
in
ts
f
o
r
4
x
s
p
r
a
y
p
ass
es
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
Op
timiz
a
tio
n
o
f le
a
r
n
in
g
a
l
g
o
r
ith
ms in
mu
ltil
a
ye
r
p
ercep
tr
on
fo
r
…
(
N
o
o
r
A
ima
n
b
i
n
A
mi
n
u
d
d
in
)
689
Fig
u
r
e
3
.
T
h
e
ML
P
Ar
ch
itectu
r
e
m
o
d
elin
g
f
o
r
r
GO
s
h
ee
t r
esis
tan
ce
class
if
ier
3.
RE
SU
L
T
S
A
ND
D
IS
CU
SS
I
O
N
T
h
is
s
ec
tio
n
p
r
esen
ts
th
e
cl
ass
if
ier
'
s
p
er
f
o
r
m
an
ce
in
d
is
cr
im
in
atin
g
t
h
e
u
n
if
o
r
m
ity
o
f
r
ed
u
ce
d
g
r
ap
h
e
n
e
o
x
id
e
(
r
GO)
th
in
-
f
ilm
s
h
ee
t
r
esis
tan
ce
r
esu
lts
o
b
ta
in
ed
b
y
o
p
tim
izin
g
R
P,
SC
G
an
d
L
M
alg
o
r
ith
m
s
in
test
in
g
,
v
alid
atio
n
an
d
tr
ain
in
g
.
B
esid
es,
th
ese
o
p
tim
izatio
n
lear
n
i
n
g
alg
o
r
ith
m
s
,
th
e
n
u
m
b
er
o
f
ep
o
ch
s
,
th
e
MSE
v
alu
e
an
d
t
h
e
n
u
m
b
er
o
f
iter
atio
n
s
ar
e
also
h
ig
h
li
g
h
ted
.
T
ab
le
1
an
d
T
ab
le
2
tab
u
late
th
e
ac
cu
r
ac
ies
f
o
r
tr
ain
in
g
,
v
al
id
atio
n
an
d
test
in
g
f
o
r
b
o
th
d
atasets
:
3
x
s
p
r
ay
p
ass
es
an
d
4
x
s
p
r
ay
p
ass
es.
T
h
e
m
ac
h
in
e
lear
n
in
g
f
r
o
m
th
e
MA
T
L
AB
s
o
f
twar
e
im
p
lem
en
tatio
n
o
p
tim
ized
th
e
m
o
d
el
d
e
v
elo
p
m
en
t
th
at
f
o
r
all
th
e
t
r
ain
in
g
,
v
alid
atio
n
a
n
d
test
in
g
g
iv
e
th
e
1
0
0
%
ac
cu
r
ac
y
f
o
r
b
o
th
d
atasets
.
F
r
o
m
t
h
e
s
e
d
a
t
a,
i
t
c
a
n
b
e
s
e
e
n
t
h
a
t
t
h
e
p
r
o
b
l
e
m
o
f
u
n
i
f
o
r
m
i
t
y
o
f
r
G
O
s
h
e
e
t
r
e
s
is
t
a
n
c
e
is
n
o
t
t
o
o
d
i
f
f
i
c
u
l
t
f
o
r
M
A
T
L
A
B
v
e
r
s
i
o
n
R
2
0
1
8
a
t
o
o
p
t
i
m
i
z
e
a
n
d
t
h
e
m
o
d
e
l
i
s
g
e
n
u
i
n
e
l
y
p
e
r
f
o
r
m
i
n
g
ex
c
e
p
t
i
o
n
a
l
l
y
w
el
l
.
T
ab
le
1.
Acc
u
r
ac
y
f
o
r
tr
ain
in
g
,
v
alid
atio
n
a
n
d
test
in
g
p
er
f
o
r
m
an
ce
f
o
r
R
P,
SC
G
an
d
L
M
f
o
r
3
x
s
p
r
ay
p
ass
es
N
o
.
o
f
h
i
d
d
e
n
n
e
u
r
o
n
A
c
c
u
r
a
c
y
(
%)
Tr
a
i
n
i
n
g
V
a
l
i
d
a
t
i
o
n
Te
st
i
n
g
1
1
0
0
1
0
0
1
0
0
2
1
0
0
1
0
0
1
0
0
3
1
0
0
1
0
0
1
0
0
4
1
0
0
1
0
0
1
0
0
5
1
0
0
1
0
0
1
0
0
6
1
0
0
1
0
0
1
0
0
7
1
0
0
1
0
0
1
0
0
8
1
0
0
1
0
0
1
0
0
9
1
0
0
1
0
0
1
0
0
10
1
0
0
1
0
0
1
0
0
T
ab
le
2
.
Acc
u
r
ac
y
f
o
r
tr
ain
in
g
,
v
alid
atio
n
a
n
d
test
in
g
p
er
f
o
r
m
an
ce
f
o
r
R
P,
SC
G
an
d
L
M
f
o
r
4
x
s
p
r
ay
p
ass
es
N
o
.
o
f
h
i
d
d
e
n
n
e
u
r
o
n
A
c
c
u
r
a
c
y
(
%)
Tr
a
i
n
i
n
g
V
a
l
i
d
a
t
i
o
n
Te
st
i
n
g
1
1
0
0
1
0
0
1
0
0
2
1
0
0
1
0
0
1
0
0
3
1
0
0
1
0
0
1
0
0
4
1
0
0
1
0
0
1
0
0
5
1
0
0
1
0
0
1
0
0
6
1
0
0
1
0
0
1
0
0
7
1
0
0
1
0
0
1
0
0
8
1
0
0
1
0
0
1
0
0
9
1
0
0
1
0
0
1
0
0
10
1
0
0
1
0
0
1
0
0
R
esu
lts
illu
s
tr
ated
in
T
ab
le
3
an
d
T
ab
le
4
s
h
o
w
th
e
av
e
r
a
g
e
MSE
v
er
s
u
s
th
e
n
u
m
b
er
o
f
h
id
d
e
n
n
eu
r
o
n
s
u
s
in
g
R
P,
SC
G
an
d
L
M
alg
o
r
ith
m
s
f
o
r
3
x
s
p
r
ay
p
ass
es
an
d
4
x
s
p
r
ay
p
ass
es
d
atasets
.
Fo
r
3
x
s
p
r
ay
p
ass
es
,
th
e
av
er
ag
e
MSE
am
o
n
g
th
e
alg
o
r
ith
m
s
s
h
o
ws
th
at
L
M
was
th
e
b
est
p
er
f
o
r
m
an
ce
am
o
n
g
th
e
o
t
h
er
s
.
T
h
e
s
m
allest
v
alu
e
f
o
r
L
M
is
o
n
ly
7
.
2
6
×
10
−
10
at
h
id
d
e
n
n
e
u
r
o
n
s
2
f
o
llo
we
d
b
y
SC
G
(
1
.
2
5
×
10
−
7
)
al
g
o
r
ith
m
at
h
id
d
e
n
n
eu
r
o
n
s
1
0
an
d
last
ly
,
R
P
(
1
.
9
8
×
10
−
7
)
alg
o
r
ith
m
f
o
r
n
in
e
h
id
d
en
n
eu
r
o
n
s
.
I
n
4
x
s
p
r
ay
p
ass
es
,
th
e
a
v
er
ag
e
MSE
f
o
r
L
M
is
g
o
o
d
co
m
p
a
r
ed
to
SC
G
an
d
R
P
al
g
o
r
ith
m
s
.
T
h
e
L
M
alg
o
r
ith
m
g
iv
es
th
e
m
in
i
m
u
m
er
r
o
r
,
wh
ich
is
o
n
ly
1
.
0
8
×
10
−
9
at
h
id
d
en
n
eu
r
o
n
7
wh
ile
SC
G
(
1
.
2
0
×
10
−
7
)
at
8
h
id
d
en
n
eu
r
o
n
s
an
d
R
P
(
2
.
2
6
×
10
−
7
)
f
o
r
9
h
id
d
en
n
eu
r
o
n
s
.
Fu
r
th
er
m
o
r
e
,
SC
G
an
d
R
P
alg
o
r
ith
m
s
s
till
g
iv
e
th
e
s
lig
h
test
er
r
o
r
f
o
r
b
o
th
3
x
s
p
r
ay
p
ass
es
an
d
4
x
s
p
r
ay
p
ass
es
,
th
e
r
an
g
e
v
alu
e
clo
s
e
to
0
.
T
ab
le
5
an
d
T
ab
le
6
s
u
m
m
ar
i
ze
th
e
p
ar
am
eter
f
o
r
t
h
e
b
est
-
h
id
d
en
n
eu
r
o
n
s
o
b
tain
i
n
SC
G,
L
M
an
d
R
P
alg
o
r
ith
m
s
(
f
o
r
b
o
th
d
atasets
:
3
x
s
p
r
ay
p
ass
es
an
d
4
x
s
p
r
ay
p
ass
es
.
A
s
im
ilar
f
in
d
in
g
was
ac
h
iev
ed
b
y
3
x
s
p
r
ay
p
ass
es
an
d
4
x
s
p
r
ay
p
ass
es
in
wh
ich
,
f
o
r
b
o
t
h
s
p
r
a
y
s
p
ass
,
th
e
L
M
tr
ain
in
g
alg
o
r
ith
m
o
u
tp
er
f
o
r
m
s
o
th
er
s
.
I
t
i
s
s
h
o
w
n
b
y
h
a
v
i
n
g
th
e
l
o
w
es
t
M
S
E
a
t
7
.
6
8
×
10
−
10
(
3
x
s
p
r
a
y
p
a
s
s
e
s
)
a
n
d
1
.
0
8
×
10
−
9
(
4
x
s
p
r
a
y
p
a
s
s
e
s
)
.
B
es
i
d
e
s
t
h
a
t
,
t
h
e
n
u
m
b
e
r
o
f
n
o
d
e
s
i
n
h
i
d
d
e
n
n
e
u
r
o
n
s
f
o
r
L
M
i
s
2
(
3
x
s
p
r
a
y
p
ass
e
s
)
a
n
d
7
(
4
x
s
p
r
a
y
p
a
s
s
es
)
a
c
c
o
m
p
a
n
i
e
d
w
i
t
h
t
h
e
e
p
o
c
h
o
f
1
4
(
3
x
s
p
r
a
y
p
a
s
s
e
s
)
a
n
d
8
(
4
x
s
p
r
a
y
p
a
s
s
es
)
.
I
n
t
e
r
es
t
i
n
g
l
y
,
S
C
G
a
n
d
R
P
a
ls
o
g
i
v
e
t
h
e
s
m
al
l
es
t
v
a
l
u
e
o
f
MS
E
,
w
h
i
c
h
i
s
n
ea
r
e
r
t
o
0
a
n
d
u
s
e
d
t
h
e
m
i
n
i
m
u
m
o
f
h
i
d
d
e
n
n
e
u
r
o
n
s
a
n
d
e
p
o
c
h
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
23
,
No
.
2
,
Au
g
u
s
t 2
0
2
1
:
6
86
-
6
9
3
690
T
ab
le
3
.
Av
e
r
ag
e
MSE
v
e
r
s
u
s
th
e
n
u
m
b
er
o
f
h
id
d
en
n
e
u
r
o
n
s
u
s
in
g
R
P,
SC
G
an
d
L
M
alg
o
r
i
th
m
s
f
o
r
3
x
s
p
r
ay
p
ass
es d
atasets
N
o
.
o
f
h
i
d
d
e
n
n
e
u
r
o
n
M
e
a
n
sq
u
a
r
e
e
r
r
o
r
(
M
S
E)
RP
M
e
a
n
sq
u
a
r
e
e
r
r
o
r
(
M
S
E)
S
C
G
M
e
a
n
sq
u
a
r
e
e
r
r
o
r
(
M
S
E)
LM
*
1
3
.
3
5
×
10
−
6
3
.
2
1
×
10
−
7
1
.
2
8
×
10
−
8
2
2
.
7
8
×
10
−
6
3
.
1
0
×
10
−
7
7
.
6
8
×
10
−
10
3
2
.
2
7
×
10
−
6
1
.
4
7
×
10
−
7
2
.
2
6
×
10
−
8
4
2
.
2
8
×
10
−
6
1
.
2
8
×
10
−
7
4
.
7
2
×
10
−
9
5
2
.
0
1
×
10
−
6
1
.
8
4
×
10
−
7
1
.
3
1
×
10
−
8
6
9
.
9
0
×
10
−
7
2
.
1
9
×
10
−
7
3
.
0
4
×
10
−
9
7
1
.
8
7
×
10
−
6
1
.
3
3
×
10
−
7
9
.
6
6
×
10
−
9
8
3
.
9
6
×
10
−
7
1
.
7
0
×
10
−
7
3
.
6
9
×
10
−
9
9
1
.
9
8
×
10
−
7
1
.
5
6
×
10
−
7
2
.
1
8
×
10
−
9
10
1
.
6
1
×
10
−
6
1
.
2
5
×
10
−
7
1
.
8
2
×
10
−
9
*
Th
e
b
e
s
t
p
e
r
f
o
r
ma
n
c
e
T
ab
le
4
.
Av
e
r
ag
e
MSE
v
er
s
u
s
th
e
n
u
m
b
er
o
f
h
id
d
en
n
e
u
r
o
n
s
u
s
in
g
R
P,
SC
G
an
d
L
M
alg
o
r
i
th
m
s
f
o
r
4
x
s
p
r
ay
p
ass
es d
atasets
N
o
.
o
f
h
i
d
d
e
n
n
e
u
r
o
n
M
e
a
n
sq
u
a
r
e
e
r
r
o
r
(
M
S
E)
RP
M
e
a
n
sq
u
a
r
e
e
r
r
o
r
(
M
S
E)
S
C
G
M
e
a
n
sq
u
a
r
e
e
r
r
o
r
(
M
S
E)
LM
*
1
4
.
1
7
×
10
−
6
4
.
6
7
×
10
−
7
2
.
3
7
×
10
−
8
2
2
.
3
9
×
10
−
6
1
.
8
0
×
10
−
7
2
.
4
1
×
10
−
8
3
9
.
3
0
×
10
−
7
3
.
7
4
×
10
−
7
8
.
7
2
×
10
−
9
4
1
.
9
6
×
10
−
6
1
.
4
7
×
10
−
7
5
.
8
7
×
10
−
9
5
1
.
2
8
×
10
−
6
3
.
7
4
×
10
−
7
6
.
7
4
×
10
−
9
6
2
.
1
3
×
10
−
6
1
.
5
0
×
10
−
7
2
.
4
3
×
10
−
9
7
2
.
2
6
×
10
−
7
1
.
2
1
×
10
−
7
1
.
0
8
×
10
−
9
8
1
.
3
7
×
10
−
6
1
.
2
0
×
10
−
7
5
.
2
0
×
10
−
9
9
3
.
8
5
×
10
−
7
1
.
6
4
×
10
−
7
3
.
4
0
×
10
−
9
10
1
.
4
8
×
10
−
6
1
.
3
0
×
10
−
7
3
.
8
7
×
10
−
9
T
ab
le
5
.
Par
am
eter
s
f
o
r
th
e
b
e
s
t RP
,
S
C
G
an
d
L
M
alg
o
r
ith
m
s
f
o
r
3
x
s
p
r
ay
p
ass
es d
atasets
P
a
r
a
me
t
e
r
s
RP
LM
S
C
G
M
S
E
1
.
9
8
×
10
−
7
7
.
6
8
×
10
−
10
1
.
2
5
×
10
−
7
N
u
mb
e
r
o
f
n
o
d
e
s
i
n
h
i
d
d
e
n
n
e
u
r
o
n
9
2
10
Ep
o
c
h
s
5
14
22
T
ab
le
6
.
Par
am
eter
s
f
o
r
th
e
b
e
s
t RP
,
S
C
G
an
d
L
M
alg
o
r
ith
m
s
f
o
r
4
x
s
p
r
ay
p
ass
es d
atasets
P
a
r
a
me
t
e
r
s
RP
LM
S
C
G
M
S
E
2
.
2
6
×
10
−
7
1
.
0
8
×
10
−
9
1
.
2
0
×
10
−
7
N
u
mb
e
r
o
f
n
o
d
e
s
i
n
h
i
d
d
e
n
n
e
u
r
o
n
7
7
8
Ep
o
c
h
s
9
8
27
T
h
e
o
v
er
all
ac
cu
r
ac
y
r
esu
lts
f
o
r
b
o
th
3
x
an
d
4
x
s
p
r
a
y
p
as
s
es
wer
e
q
u
ite
s
im
ilar
.
T
h
er
e
f
o
r
e,
o
n
ly
ML
P
f
in
al
d
esig
n
p
ar
a
m
eter
f
o
r
3
x
s
p
r
a
y
p
ass
es
is
d
is
cu
s
s
e
d
in
th
is
s
tu
d
y
.
T
ab
le
7
s
u
m
m
ar
izes
th
e
last
d
esig
n
p
ar
am
eter
f
o
r
ML
P
a
r
ch
itectu
r
e
an
d
th
e
tr
ain
i
n
g
p
ar
am
eter
.
T
h
e
h
i
g
h
est
MSE
f
o
u
n
d
in
L
M
b
y
u
s
in
g
d
ataset
3
x
s
p
r
ay
p
ass
es
wh
ich
is
7
.
6
8
×
10
−
10
with
th
e
ep
o
ch
s
o
f
1
4
iter
at
io
n
s
.
T
h
e
n
u
m
b
er
o
f
th
e
i
n
p
u
t
lay
er
is
4
9
p
o
in
ts
b
y
r
ea
d
in
g
f
r
o
m
r
GO
s
h
ee
t r
esis
tan
ce
an
d
th
e
h
id
d
en
n
eu
r
o
n
s
o
f
2
n
eu
r
o
n
s
in
th
e
h
i
d
d
en
lay
er
o
b
tain
ed
th
e
lo
west
MSE
.
T
h
e
o
u
tp
u
t,
wh
ich
is
th
e
u
n
if
o
r
m
ity
o
f
r
GO
s
h
ee
t
r
esis
tan
ce
,
is
r
ep
r
e
s
en
ted
b
y
1
o
u
tp
u
t
lay
er
.
T
h
e
co
n
f
u
s
io
n
m
atr
ix
,
a
s
d
ep
icted
in
Fig
u
r
e
4
,
s
h
o
ws
th
at
th
e
ac
cu
r
ac
y
o
f
th
e
tr
ain
i
n
g
d
ataset
is
1
0
0
%,
an
d
it
is
p
r
o
v
en
th
at
L
M
o
u
tp
er
f
o
r
m
s
o
th
er
s
.
T
h
is
f
in
d
in
g
is
s
u
p
p
o
r
ted
b
y
th
e
r
esu
lts
as
t
ab
u
lated
in
T
ab
le
8
in
wh
ich
th
e
ac
c
u
r
ac
y
,
s
en
s
itiv
ity
,
s
p
ec
if
icity
an
d
p
r
ec
is
io
n
ar
e
all
1
0
0
%.
T
ab
le
7
.
MLP
d
esig
n
p
ar
am
et
er
s
P
a
r
a
me
t
e
r
V
a
l
u
e
Le
a
r
n
i
n
g
a
l
g
o
r
i
t
h
m
s
LM
D
a
t
a
s
e
t
3
x
sp
r
a
y
p
a
sses
Ep
o
c
h
s
14
M
e
a
n
S
q
u
a
r
e
Er
r
o
r
(
M
S
E)
7
.
6
8
×
10
−
10
N
u
mb
e
r
o
f
I
n
p
u
t
s
l
a
y
e
r
49
N
u
mb
e
r
o
f
n
o
d
e
s
i
n
t
h
e
h
i
d
d
e
n
l
a
y
e
r
2
N
u
mb
e
r
o
f
t
h
e
o
u
t
p
u
t
l
a
y
e
r
1
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
Op
timiz
a
tio
n
o
f le
a
r
n
in
g
a
l
g
o
r
ith
ms in
mu
ltil
a
ye
r
p
ercep
tr
on
fo
r
…
(
N
o
o
r
A
ima
n
b
i
n
A
mi
n
u
d
d
in
)
691
Fig
u
r
e
4
.
T
h
e
ML
P c
lass
if
icati
o
n
p
er
f
o
r
m
a
n
ce
co
n
f
u
s
io
n
m
atr
ix
T
ab
le
8
.
Acc
u
r
ac
y
,
s
en
s
itiv
ity
,
s
p
ec
if
icity
an
d
p
r
ec
is
io
n
P
a
r
a
me
t
e
r
V
a
l
u
e
(
%)
A
c
c
u
r
a
c
y
1
0
0
S
e
n
s
i
t
i
v
i
t
y
1
0
0
S
p
e
c
i
f
i
c
i
t
y
1
0
0
P
r
e
c
i
s
i
o
n
1
0
0
I
n
g
en
er
al,
it
was
f
o
u
n
d
th
at
lev
en
b
er
g
m
ar
q
u
ar
d
t
(
L
M)
tr
ai
n
in
g
alg
o
r
it
h
m
s
o
u
tp
er
f
o
r
m
o
t
h
er
s
f
o
r
all
tr
ain
in
g
,
test
in
g
an
d
v
alid
atio
n
d
atasets
.
T
h
is
s
tu
d
y
p
r
o
v
ed
th
at
L
M
c
o
u
ld
o
b
tain
th
e
lo
west
m
ea
n
s
q
u
ar
ed
er
r
o
r
(
MSE
)
a
n
d
p
er
f
o
r
m
s
1
0
0
%
ac
cu
r
ac
y
f
o
r
o
p
tim
a
l
b
e
h
a
v
io
u
r
(
s
en
s
itiv
ity
,
s
p
ec
if
icity
an
d
p
r
ec
is
io
n
)
.
Du
e
to
th
is
,
L
M
ca
p
ab
le
o
f
ap
p
r
o
ac
h
in
g
th
e
d
ir
ec
tio
n
o
f
th
e
s
teep
est
d
escen
t.
T
h
is
f
in
d
in
g
ag
r
ee
s
th
at
SC
G
is
f
aster
in
co
m
p
u
tatio
n
al
tim
e
b
u
t
m
o
r
e
s
ig
n
if
ican
t
er
r
o
r
co
m
p
ar
es
to
L
M.
T
h
is
i
s
b
ec
au
s
e
i
t
a
v
o
id
s
lin
e
s
ea
r
ch
p
er
lear
n
in
g
iter
atio
n
b
y
th
e
L
M
ap
p
r
o
ac
h
.
T
h
e
r
ef
o
r
e,
it is
p
r
o
v
id
in
g
th
e
lo
west su
m
o
f
s
q
u
ar
ed
er
r
o
r
s
.
4.
CO
NCLU
SI
O
N
T
h
e
o
p
tim
izatio
n
alg
o
r
ith
m
s
in
th
e
ML
P
n
etwo
r
k
f
o
r
r
GO
t
h
in
-
f
ilm
s
h
ee
t
r
esis
tan
ce
u
n
if
o
r
m
ity
f
o
r
3
x
a
n
d
4
x
s
p
r
ay
p
ass
es
wer
e
s
u
cc
ess
f
u
lly
p
er
f
o
r
m
e
d
in
th
is
s
tu
d
y
.
I
t
is
clea
r
t
h
at,
am
i
d
th
r
ee
alg
o
r
ith
m
s
u
s
ed
in
n
etwo
r
k
tr
ai
n
in
g
d
ev
elo
p
m
en
t
to
o
p
tim
ize
r
GO
s
h
ee
t
r
esis
tan
ce
u
n
if
o
r
m
ity
,
L
M
g
ain
s
t
h
e
b
est.
T
h
is
s
tu
d
y
h
as
b
ee
n
p
r
o
v
en
th
r
o
u
g
h
u
p
p
er
m
o
s
t
ac
cu
r
ac
y
(
1
0
0
%)
y
iel
d
ed
b
y
L
M
f
o
r
test
in
g
,
v
alid
atio
n
an
d
tr
ain
in
g
.
Ho
wev
er
,
th
e
r
esu
lt
is
r
ed
u
ce
d
to
th
e
MSE
v
alu
e
o
n
ly
an
d
t
h
e
n
u
m
b
er
o
f
ep
o
ch
s
an
d
a
n
u
m
b
er
o
f
th
e
h
id
d
en
n
eu
r
o
n
r
esear
ch
ar
ea
.
T
h
is
s
tu
d
y
is
ess
en
tial
an
d
b
en
ef
icia
l,
esp
ec
ially
in
th
e
u
n
if
o
r
m
ity
o
f
g
r
ap
h
en
e
s
h
ee
t
r
esis
tan
ce
f
o
r
class
if
ier
an
aly
s
is
.
ACK
NO
WL
E
DG
E
M
E
NT
S
T
h
e
au
th
o
r
s
ac
k
n
o
wled
g
e
f
u
n
d
in
g
f
r
o
m
th
e
Min
is
ter
o
f
Hig
h
er
E
d
u
ca
tio
n
(
MO
HE
)
o
f
Ma
lay
s
ia
u
n
d
er
th
e
FR
GS
Gr
an
t
No
:
6
0
0
-
I
R
MI
/FR
GS
5
/3
/
(
0
3
1
/2
0
1
9
)
a
n
d
th
e
Sch
o
o
l
o
f
E
lec
tr
ical
E
n
g
in
ee
r
in
g
,
C
o
lleg
e
o
f
E
n
g
i
n
ee
r
in
g
,
Un
iv
e
r
s
iti T
ek
n
o
lo
g
i M
AR
A
(
UiT
M)
f
o
r
s
u
p
p
o
r
tin
g
th
is
r
esear
ch
.
RE
F
E
R
E
NC
E
S
[1
]
A.
Da
h
a
l,
"
S
u
rfa
c
e
S
c
ien
c
e
S
tu
d
i
e
s
o
f
G
ra
p
h
e
n
e
I
n
terfa
c
e
s,
"
G
ra
d
u
a
te
Th
e
se
s,
De
p
a
rtme
n
t
o
f
P
h
y
s
ics
,
Un
iv
e
rsit
y
o
f
S
o
u
t
h
F
l
o
ri
d
a
,
USA,
2
0
1
5
.
[2
]
Y.
Ha
n
,
"
El
e
c
tro
n
ic
P
ro
p
e
rti
e
s
o
f
G
r
a
p
h
e
n
e
Tu
n
e
d
b
y
Tw
o
-
d
ime
n
sio
n
a
l
Cry
sta
ls
,
"
T
h
e
sis
,
De
p
a
rtme
n
t
o
f
P
h
y
sic
s
,
Ho
n
g
Ko
n
g
Un
iv
e
rsity
o
f
S
c
ien
c
e
a
n
d
Tec
h
n
o
l
o
g
y
,
H
o
n
g
Ko
n
g
,
2
0
1
5
.
[3
]
V.
K.
Na
g
a
re
d
d
y
,
"
F
a
b
r
ica
ti
o
n
,
fu
n
c
ti
o
n
a
li
sa
ti
o
n
a
n
d
c
h
a
ra
c
terisa
ti
o
n
o
f
e
p
it
a
x
ial
g
ra
p
h
e
n
e
d
e
v
ice
s,"
Th
e
sis
,
De
p
a
rtme
n
t
o
f
El
e
c
tri
c
a
l
a
n
d
El
e
c
tro
n
ic E
n
g
in
e
e
rin
g
,
Ne
wc
a
stle Un
iv
e
rsity
,
Un
it
e
d
Kin
g
d
o
m
,
2
0
1
5
.
[4
]
M
.
G
o
o
se
y
,
"
A
sh
o
rt
in
tr
o
d
u
c
ti
o
n
t
o
g
ra
p
h
e
n
e
a
n
d
it
s
p
o
te
n
ti
a
l
in
terc
o
n
n
e
c
t
a
p
p
li
c
a
ti
o
n
s,"
Circ
u
i
t
W
o
rld
,
v
o
l.
3
8
,
n
o
.
2
,
p
p
.
8
3
-
8
6
,
2
0
1
2
,
d
o
i:
1
0
.
1
1
0
8
/
0
3
0
5
6
1
2
1
2
1
1
2
2
2
3
0
9
.
[5
]
W.
Wa
n
g
,
"
De
v
e
lo
p
m
e
n
t
o
f
N
a
n
o
c
o
m
p
o
sites
Us
in
g
G
ra
p
h
e
n
e
S
y
n
t
h
e
siz
e
d
b
y
S
o
lv
e
n
t
Ex
f
o
li
a
ti
o
n
M
e
th
o
d
,
"
Un
iv
e
rsity
o
f
To
led
o
,
2
0
1
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
23
,
No
.
2
,
Au
g
u
s
t 2
0
2
1
:
6
86
-
6
9
3
692
[6
]
L.
B.
M
o
d
e
sto
-
Ló
p
e
z
,
M
.
M
iett
in
e
n
,
T.
T
o
rv
e
la,
A.
Läh
d
e
,
a
n
d
J.
Jo
k
i
n
iem
i,
"
Dire
c
t
d
e
p
o
si
ti
o
n
o
f
g
ra
p
h
e
n
e
n
a
n
o
m
a
teria
l
fil
m
s
o
n
p
o
l
y
m
e
r
-
c
o
a
ted
g
las
s
b
y
u
lt
ra
so
n
ic
sp
ra
y
in
g
,
"
T
h
in
S
o
li
d
Fi
lms,
v
o
l.
5
7
8
,
p
p
.
4
5
-
5
2
,
2
0
1
5
,
d
o
i:
1
0
.
1
0
1
6
/j
.
tsf.
2
0
1
5
.
0
1
.
0
7
3
.
[7
]
D.
S
tein
b
e
rg
,
R.
G
e
ro
sa
,
F
.
P
e
l
li
c
e
r,
S
.
Do
m
in
g
u
e
s,
E.
T.
d
e
S
o
u
z
a
,
a
n
d
L.
S
a
it
o
,
"
S
u
b
-
3
0
0
fs
m
o
d
e
-
lo
c
k
e
d
e
rb
i
u
m
d
o
p
e
d
f
ib
e
r
las
e
r
u
si
n
g
g
ra
p
h
e
n
e
o
x
i
d
e
a
n
d
re
d
u
c
e
d
g
ra
p
h
e
n
e
o
x
id
e
o
n
t
o
D
-
sh
a
p
e
d
o
p
ti
c
a
l
f
ib
e
rs,"
in
2
0
1
7
S
BM
O/IE
EE
M
T
T
-
S
In
ter
n
a
ti
o
n
a
l
M
icr
o
wa
v
e
a
n
d
Op
t
o
e
lec
tro
n
ics
Co
n
fer
e
n
c
e
(I
M
OC)
,
IEE
E,
2
0
1
7
,
p
p
.
1
-
3
,
d
o
i:
1
0
.
1
1
0
9
/IM
OC.
2
0
1
7
.
8
1
2
1
0
8
6
.
[8
]
S.
K
h
o
sro
jerd
i
,
M
.
Va
k
il
i,
M
.
Ya
h
y
a
e
i,
a
n
d
K.
Ka
lh
o
r,
"
Th
e
rm
a
l
c
o
n
d
u
c
ti
v
i
ty
m
o
d
e
li
n
g
o
f
g
ra
p
h
e
n
e
n
a
n
o
p
late
lets/d
e
io
n
ize
d
wa
ter
n
a
n
o
fl
u
id
b
y
M
L
P
n
e
u
ra
l
n
e
tw
o
r
k
a
n
d
th
e
o
re
ti
c
a
l
m
o
d
e
li
n
g
u
si
n
g
e
x
p
e
rime
n
tal
re
su
l
ts,"
I
n
ter
n
a
ti
o
n
a
l
Co
mm
u
n
ica
ti
o
n
s
i
n
He
a
t
a
n
d
M
a
ss
T
ra
n
sfe
r,
v
o
l.
7
4
,
p
p
.
1
1
-
1
7
,
2
0
1
6
,
d
o
i:
1
0
.
1
0
1
6
/j
.
ich
e
a
tma
ss
tran
sfe
r.
2
0
1
6
.
0
3
.
0
1
0
.
[9
]
M
.
F
a
n
,
e
t
a
l
.
,
"
Art
ifi
c
ial
n
e
u
ra
l
n
e
two
rk
m
o
d
e
li
n
g
a
n
d
g
e
n
e
ti
c
a
l
g
o
rit
h
m
o
p
ti
m
iza
ti
o
n
fo
r
c
a
d
m
i
u
m
re
m
o
v
a
l
fro
m
a
q
u
e
o
u
s
so
l
u
ti
o
n
s
b
y
re
d
u
c
e
d
g
ra
p
h
e
n
e
o
x
i
d
e
-
su
p
p
o
rte
d
n
a
n
o
sc
a
le
z
e
ro
-
v
a
len
t
ir
o
n
(n
ZVI/r
G
O)
c
o
m
p
o
sites
,
"
M
a
ter
ia
ls,
v
o
l.
1
0
,
n
o
.
5
,
p
.
5
4
4
,
2
0
1
7
,
d
o
i:
1
0
.
3
3
9
0
/ma
1
0
0
5
0
5
4
4
.
[1
0
]
P
.
Am
a
n
i
a
n
d
K.
Va
jrav
e
lu
,
"
I
n
te
ll
ig
e
n
t
m
o
d
e
li
n
g
o
f
r
h
e
o
l
o
g
ica
l
a
n
d
t
h
e
rm
o
p
h
y
sic
a
l
p
ro
p
e
rti
e
s
o
f
g
re
e
n
c
o
v
a
len
tl
y
fu
n
c
ti
o
n
a
li
z
e
d
g
ra
p
h
e
n
e
n
a
n
o
fl
u
i
d
s
c
o
n
tain
i
n
g
n
a
n
o
p
late
lets,"
In
t
e
rn
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
He
a
t
a
n
d
M
a
ss
T
ra
n
sfe
r,
v
o
l.
1
2
0
,
p
p
.
9
5
-
1
0
5
,
2
0
1
8
,
d
o
i:
1
0
.
1
0
1
6
/j
.
i
jh
e
a
tma
ss
tran
sfe
r.
2
0
1
7
.
1
2
.
0
2
5
.
[1
1
]
K.
Bü
y
ü
k
k
a
b
a
sa
k
a
l,
S
.
C
.
Ac
i
k
b
a
s,
A.
De
n
iz,
Y.
Ac
i
k
b
a
s,
R.
Ca
p
a
n
,
a
n
d
M
.
Er
d
o
g
a
n
,
"
Ch
e
m
ica
l
se
n
so
r
p
r
o
p
e
rti
e
s
a
n
d
m
a
th
e
m
a
ti
c
a
l
m
o
d
e
li
n
g
o
f
g
ra
p
h
e
n
e
o
x
i
d
e
lan
g
m
u
ir
-
b
lo
d
g
e
tt
th
i
n
fil
m
s,"
IE
EE
S
e
n
so
rs
J
o
u
rn
a
l
,
v
o
l.
1
9
,
n
o
.
2
0
,
p
p
.
9
0
9
7
-
9
1
0
4
,
2
0
1
9
,
d
o
i:
1
0
.
1
1
0
9
/JS
EN.
2
0
1
9
.
2
9
2
6
3
6
7
.
[1
2
]
M
.
Ka
m
ra
n
,
S
.
Ha
id
e
r,
T.
Ak
ra
m
,
S
.
Na
q
v
i
,
a
n
d
S
.
He
,
"
P
re
d
ict
io
n
o
f
IV
c
u
r
v
e
s
fo
r
a
su
p
e
rc
o
n
d
u
c
ti
n
g
th
i
n
fil
m
u
sin
g
a
rti
ficia
l
n
e
u
ra
l
n
e
tw
o
rk
s,"
S
u
p
e
rl
a
tt
ice
s
a
n
d
M
i
c
ro
stru
c
tu
re
s,
v
o
l.
9
5
,
p
p
.
8
8
-
9
4
,
2
0
1
6
,
d
o
i:
1
0
.
1
0
1
6
/j
.
s
p
m
i.
2
0
1
6
.
0
4
.
0
1
8
.
[1
3
]
H.
G
u
o
,
J.
Zh
a
o
,
a
n
d
J.
Yin
,
"
Ra
n
d
o
m
f
o
re
st
a
n
d
m
u
l
ti
lay
e
r
p
e
rc
e
p
tro
n
fo
r
p
re
d
ictin
g
t
h
e
d
i
e
lec
tri
c
lo
ss
o
f
p
o
l
y
imid
e
n
a
n
o
c
o
m
p
o
site
fil
m
s,"
RS
C
a
d
v
a
n
c
e
s,
v
o
l
.
7
,
n
o
.
4
9
,
p
p
.
3
0
9
9
9
-
3
1
0
0
8
,
Ju
n
e
2
0
1
7
,
d
o
i:
1
0
.
1
0
3
9
/C7
RA
0
4
1
4
7
K.
[1
4
]
H.
G
u
o
,
J.
Zh
a
o
,
J.
Yi
n
,
a
n
d
L.
Ya
o
,
"
S
t
ru
c
tu
ra
l
tes
ti
n
g
o
f
p
o
l
y
i
m
id
e
n
a
n
o
c
o
m
p
o
si
te
fil
m
s
wit
h
S
AX
S
a
n
d
S
VM
-
P
UK
,
"
P
o
lym
e
r T
e
stin
g
,
v
o
l.
7
0
,
p
p
.
3
0
-
3
8
,
2
0
1
8
,
d
o
i:
1
0
.
1
0
1
6
/
j.
p
o
ly
m
e
rtes
ti
n
g
.
2
0
1
8
.
0
6
.
0
2
5
.
[1
5
]
M
.
Ko
n
o
m
i
a
n
d
G
.
S
a
c
h
a
,
"
F
e
e
d
fo
rwa
rd
n
e
u
ra
l
n
e
tw
o
rk
m
e
th
o
d
o
lo
g
y
t
o
c
h
a
ra
c
teriz
e
th
in
f
il
m
s
b
y
El
e
c
tr
o
sta
ti
c
F
o
rc
e
M
icro
sc
o
p
y
,
"
Ultr
a
mic
ro
sc
o
p
y
,
v
o
l
.
1
8
2
,
p
p
.
2
4
3
-
2
4
8
,
2
0
1
7
,
d
o
i:
1
0
.
1
0
1
6
/j
.
u
lt
ra
m
ic.2
0
1
7
.
0
7
.
0
1
5
.
[1
6
]
M
.
Ka
y
a
a
n
d
S
.
Ha
ji
m
irza
,
"
S
u
rr
o
g
a
te
b
a
se
d
m
o
d
e
li
n
g
a
n
d
o
p
ti
m
i
z
a
ti
o
n
o
f
p
las
m
o
n
ic
t
h
in
fi
lm
o
rg
a
n
ic
so
lar
c
e
ll
s,"
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
He
a
t
a
n
d
M
a
ss
T
ra
n
sfe
r,
v
o
l.
1
1
8
,
p
p
.
1
1
2
8
-
1
1
4
2
,
M
a
rc
h
2
0
1
8
,
doi
:
1
0
.
1
0
1
6
/j
.
ij
h
e
a
tma
ss
tran
sfe
r.
2
0
1
7
.
1
1
.
0
4
4
.
[1
7
]
A.
M
o
h
a
m
m
a
d
i
G
h
a
h
d
a
rij
a
n
i,
F
.
H
o
rm
o
z
i,
a
n
d
A.
Ha
g
h
i
g
h
i
As
l,
"
Co
n
v
e
c
ti
v
e
h
e
a
t
tran
sfe
r
a
n
d
p
re
s
su
re
d
r
o
p
stu
d
y
o
n
n
a
n
o
fl
u
i
d
s
in
d
o
u
b
le
-
wa
ll
e
d
r
e
a
c
to
r
b
y
d
e
v
e
lo
p
in
g
a
n
o
p
ti
m
a
l
m
u
lt
il
a
y
e
r
p
e
rc
e
p
tr
o
n
a
rti
ficia
l
n
e
u
ra
l
n
e
two
rk
,
"
In
ter
n
a
t
io
n
a
l
Co
mm
u
n
ica
ti
o
n
s
in
He
a
t
a
n
d
M
a
ss
T
ra
n
sfe
r,
v
o
l.
8
4
,
p
p
.
1
1
-
1
9
,
M
a
y
2
0
1
7
,
d
o
i:
1
0
.
1
0
1
6
/j
.
ich
e
a
tma
ss
tran
sfe
r.
2
0
1
7
.
0
3
.
0
1
4
.
[1
8
]
H.
R
eza
An
sa
ri,
M
.
J
a
v
a
d
Zare
i,
S
.
S
a
b
b
a
g
h
i,
a
n
d
P
.
Ke
sh
a
v
a
rz
,
"
A
n
e
w
c
o
m
p
re
h
e
n
si
v
e
m
o
d
e
l
fo
r
re
lati
v
e
v
isc
o
sity
o
f
v
a
rio
u
s
n
a
n
o
fl
u
id
s
u
sin
g
fe
e
d
-
fo
rwa
r
d
b
a
c
k
-
p
ro
p
a
g
a
ti
o
n
M
LP
n
e
u
ra
l
n
e
two
r
k
s,
"
In
ter
n
a
ti
o
n
a
l
Co
mm
u
n
ica
ti
o
n
s
i
n
He
a
t
a
n
d
M
a
ss
T
r
a
n
sfe
r,
v
o
l.
9
1
,
p
p
.
1
5
8
-
1
6
4
,
F
e
b
ru
a
ry
2
0
1
8
,
d
o
i:
1
0
.
1
0
1
6
/j
.
ich
e
a
tma
ss
tran
sfe
r.
2
0
1
7
.
1
2
.
0
1
2
.
[1
9
]
M
.
A.
Ha
ss
a
n
a
n
d
D.
Ba
n
e
rjee
,
"
A
so
ft
c
o
m
p
u
t
in
g
a
p
p
ro
a
c
h
fo
r
e
stim
a
ti
n
g
th
e
s
p
e
c
ifi
c
h
e
a
t
c
a
p
a
c
it
y
o
f
m
o
l
ten
sa
lt
-
b
a
se
d
n
a
n
o
flu
id
s,"
J
o
u
r
n
a
l
o
f
M
o
lec
u
la
r L
i
q
u
i
d
s,
v
o
l
.
2
8
1
,
p
p
.
3
6
5
-
3
7
5
,
2
0
1
9
,
d
o
i:
1
0
.
1
0
1
6
/
j.
m
o
ll
iq
.
2
0
1
9
.
0
2
.
1
0
6
.
[2
0
]
C.
Yilma
z
,
D.
Us
tu
n
,
a
n
d
A.
Ak
d
a
g
li
,
"
Us
a
g
e
o
f
a
rti
ficia
l
n
e
u
ra
l
n
e
two
rk
f
o
r
e
stim
a
ti
n
g
o
f
th
e
e
lec
tro
sp
u
n
n
a
n
o
fi
b
e
r
d
iam
e
ter,"
i
n
2
0
1
7
In
t
e
rn
a
ti
o
n
a
l
Arti
fi
c
i
a
l
In
telli
g
e
n
c
e
a
n
d
Da
t
a
Pro
c
e
ss
in
g
S
y
mp
o
si
u
m
(IDAP
)
,
2
0
1
7
,
p
p
.
1
-
5
,
d
o
i:
1
0
.
1
1
0
9
/IDAP.
2
0
1
7
.
8
0
9
0
3
2
9
.
[2
1
]
C.
Ie
ra
c
it
a
n
o
,
A.
P
a
v
i
g
li
a
n
it
i,
M
.
Ca
m
p
o
lo
,
A.
Hu
ss
a
in
,
E.
P
a
se
ro
,
a
n
d
F
.
C.
M
o
ra
b
it
o
,
"
A
n
o
v
e
l
a
u
t
o
m
a
ti
c
c
las
sifica
ti
o
n
sy
ste
m
b
a
se
d
o
n
h
y
b
ri
d
u
n
s
u
p
e
rv
ise
d
a
n
d
su
p
e
rv
is
e
d
m
a
c
h
in
e
lea
rn
in
g
fo
r
e
lec
tro
s
p
u
n
n
a
n
o
f
ib
e
rs,"
IEE
E/
CAA
J
o
u
rn
a
l
o
f
A
u
to
m
a
ti
c
a
S
in
ica
,
v
o
l
.
8
,
n
o
.
1
,
p
p
.
6
4
-
7
6
,
2
0
2
1
,
d
o
i:
1
0
.
1
1
0
9
/JAS.
2
0
2
0
.
1
0
0
3
3
8
7
.
[2
2
]
A.
H.
Ha
m
id
ian
,
S
.
Esfa
n
d
e
h
,
Y.
Zh
a
n
g
,
a
n
d
M
.
Ya
n
g
,
"
S
imu
lati
o
n
a
n
d
o
p
t
imiz
a
ti
o
n
o
f
n
a
n
o
m
a
teria
ls
a
p
p
li
c
a
ti
o
n
fo
r
h
e
a
v
y
m
e
tal
re
m
o
v
a
l
fro
m
a
q
u
e
o
u
s
so
lu
ti
o
n
s,
"
I
n
o
r
g
a
n
ic
a
n
d
N
a
n
o
-
M
e
t
a
l
C
h
e
mistry
,
vo
l.
4
9
,
n
o
.
7
,
p
p
.
2
1
7
-
2
3
0
,
2
0
1
9
,
d
o
i:
1
0
.
1
0
8
0
/
2
4
7
0
1
5
5
6
.
2
0
1
9
.
1
6
5
3
3
2
1
.
[2
3
]
M.
S
a
d
e
g
h
M
a
z
lo
o
m
,
F
.
Re
z
a
e
i,
A.
He
m
m
a
ti
-
S
a
ra
p
a
rd
e
h
,
M
.
M
.
Hu
se
in
,
S
.
Zen
d
e
h
b
o
u
d
i,
a
n
d
A.
Be
m
a
n
i,
"
Artifi
c
ial
i
n
telli
g
e
n
c
e
b
a
se
d
m
e
t
h
o
d
s
fo
r
a
sp
h
a
lt
e
n
e
s
a
d
so
rp
ti
o
n
b
y
n
a
n
o
c
o
m
p
o
sites
:
A
p
p
li
c
a
ti
o
n
o
f
g
ro
u
p
m
e
th
o
d
o
f
d
a
ta
h
a
n
d
li
n
g
,
lea
st
sq
u
a
re
s
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
,
a
n
d
a
rti
f
icia
l
n
e
u
ra
l
n
e
two
rk
s
,
"
Na
n
o
m
a
t
e
ria
ls,
v
o
l
.
1
0
,
n
o
.
5
,
p
p
.
8
9
0
,
2
0
2
0
,
d
o
i:
1
0
.
3
3
9
0
/n
a
n
o
1
0
0
5
0
8
9
0
.
[2
4
]
M.
R
o
fe
i
M
a
t
Hu
ss
in
,
S
.
Aish
a
h
M
o
h
a
m
a
d
Ba
d
a
ru
d
d
in
,
N.
M
o
h
d
Ra
z
a
li
M
o
h
d
N
o
r,
a
n
d
M
.
Hilmy
Az
u
a
n
Ha
m
z
a
h
,
"
Ultras
o
n
ic
a
to
m
iza
ti
o
n
o
f
g
ra
p
h
e
n
e
d
e
riv
a
ti
v
e
s
fo
r
h
e
a
t
sp
re
a
d
e
r
th
in
fil
m
d
e
p
o
siti
o
n
o
n
sili
c
o
n
su
b
stra
te,
"
M
a
ter
ia
ls T
o
d
a
y
:
Pro
c
e
e
d
in
g
s,
v
o
l.
7
,
2
0
1
9
,
p
p
.
7
6
3
-
7
6
9
,
d
o
i:
1
0
.
1
0
1
6
/j
.
m
a
tp
r.
2
0
1
8
.
1
2
.
0
7
2
.
[2
5
]
M
.
M
a
srie
,
S
.
Ba
d
a
ru
d
d
in
,
M
.
H
u
ss
in
,
N.
No
r
,
a
n
d
J.
J
o
e
,
"
Ra
p
i
d
Re
d
u
c
ti
o
n
o
f
G
ra
p
h
e
n
e
Ox
id
e
Th
in
F
il
m
s
o
n
Larg
e
-
Are
a
S
il
ico
n
S
u
b
stra
te,"
i
n
J
o
u
rn
a
l
o
f
P
h
y
sic
s:
Co
n
fer
e
n
c
e
S
e
rie
s
,
v
o
l.
1
5
3
5
,
n
o
.
1
,
2
0
2
0
,
p
.
0
1
2
0
2
7
,
d
o
i:
1
0
.
1
0
8
8
/1
7
4
2
-
6
5
9
6
/
1
5
3
5
/
1
/0
1
2
0
2
7
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
Op
timiz
a
tio
n
o
f le
a
r
n
in
g
a
l
g
o
r
ith
ms in
mu
ltil
a
ye
r
p
ercep
tr
on
fo
r
…
(
N
o
o
r
A
ima
n
b
i
n
A
mi
n
u
d
d
in
)
693
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
No
o
r
Aim
a
n
b
in
A
m
in
u
d
d
i
n
re
c
iev
e
d
th
e
B.
En
g
.
i
n
El
e
c
tro
n
ic
En
g
i
n
e
rin
g
fro
m
th
e
F
a
c
u
lt
y
o
f
El
e
c
tri
c
a
l
En
g
i
n
e
e
rin
g
,
Un
i
v
e
rsiti
Tek
n
o
lo
g
i
M
a
ra
(UiTM
)
i
n
2
0
1
9
.
He
is cu
rre
n
tl
y
w
o
rk
i
n
g
a
s
a
P
ro
c
e
ss
En
g
in
e
e
r
a
t
S
a
m
su
n
g
S
d
iem
.
He
is
re
sp
o
n
si
b
le
fo
r
d
e
sig
n
i
n
g
,
imp
lem
e
n
ti
n
g
,
a
n
d
tes
ti
n
g
n
e
w p
r
o
c
e
d
u
re
s to
o
p
ti
m
iz
e
li
th
i
u
m
io
n
b
a
tt
e
r
y
p
r
o
d
u
c
ti
o
n
.
Ir
Ts
Dr
Nurla
il
a
Is
m
a
il
is
a
se
n
io
r
lec
tu
re
r
a
t
S
c
h
o
o
l
o
f
El
e
c
tri
c
a
l
En
g
i
n
e
e
rin
g
,
C
o
ll
e
g
e
o
f
En
g
i
n
e
e
rin
g
,
Un
iv
e
rsit
i
Tek
n
o
lo
g
i
M
ARA
(UiTM
),
S
h
a
h
Ala
h
,
S
e
lan
g
o
r
.
S
h
e
o
b
tain
e
d
h
e
r
BS
c
,
M
S
c
a
n
d
P
h
D
i
n
El
e
c
tri
c
a
l
En
g
i
n
e
e
rin
g
fro
m
UiTM
S
h
a
h
Ala
m
.
S
h
e
jo
in
e
d
UiTM
d
u
ri
n
g
h
e
r
p
o
st
d
o
c
t
o
ra
l
se
rv
ice
i
n
2
0
1
6
.
S
h
e
is
a
p
r
o
fe
ss
io
n
a
l
e
n
g
in
e
e
r
i
n
t
h
e
d
isc
ip
li
n
e
o
f
tea
c
h
in
g
re
c
o
g
n
ize
d
b
y
th
e
Bo
a
r
d
o
f
En
g
in
e
e
r
M
a
lay
sia
(BEM
),
a
n
a
c
ti
v
e
m
e
m
b
e
r
in
se
v
e
ra
l
o
rg
a
n
iza
ti
o
n
s,
i
n
c
lu
d
in
g
IEE
E
M
a
lay
sia
,
e
sp
e
c
ially
Co
n
tr
o
l
S
y
ste
m
S
o
c
iety
,
M
a
lay
sia
n
S
o
c
iety
f
o
r
Co
m
p
u
te
d
To
m
o
g
r
a
p
h
y
a
n
d
Im
a
g
i
n
g
Tec
h
n
o
lo
g
y
(M
y
CT)
a
n
d
I
n
stit
u
te
o
f
En
g
i
n
e
e
r
M
a
lay
sia
(IE
M
)
a
n
d
M
a
lay
sia
Bo
a
rd
o
f
Tec
h
n
o
l
o
g
ists
(M
BOT)
a
s
we
ll
a
s
A
S
EAN
En
g
i
n
e
e
rin
g
Re
g
ister
(AER).
S
h
e
h
a
s
p
u
b
li
sh
e
d
m
o
re
th
a
n
5
0
tec
h
n
ica
l
p
a
p
e
rs,
l
o
c
a
ll
y
a
n
d
a
t
in
tern
a
ti
o
n
a
l
lev
e
l.
He
r
re
se
a
rc
h
in
tere
sts
a
re
in
a
d
v
a
n
c
e
d
si
g
n
a
l
p
ro
c
e
ss
in
g
a
n
d
a
rti
ficia
l
in
telli
g
e
n
c
e
.
Ir
Ts
Dr
Ma
r
ia
n
a
h
Ma
sr
ie
re
c
e
iv
e
d
th
e
B.
S
c
.
in
El
e
c
tri
c
,
El
e
c
tro
n
ic
&
S
y
ste
m
En
g
i
n
e
e
rin
g
,
Un
iv
e
rsiti
Ke
b
a
n
g
sa
a
n
M
a
lay
sia
(1
9
9
9
),
M
.
S
c
.
i
n
F
a
c
u
lt
y
o
f
E
lec
tri
c
a
l
En
g
in
e
e
rin
g
U
n
iv
e
rsiti
Tek
n
o
lo
g
i
M
ARA
(
2
0
0
9
)
a
n
d
P
h
.
D.
i
n
M
icro
e
n
g
in
e
e
ri
n
g
&
Na
n
o
e
lec
tro
n
ics
fro
m
In
sti
tu
te
o
f
M
icro
e
n
g
i
n
e
e
rin
g
&
Na
n
o
e
lec
tro
n
ics
(IM
EN)
Un
i
v
e
rsiti
Ke
b
a
n
g
s
a
a
n
M
a
lay
sia
(2
0
1
7
)
.
S
h
e
is
a
p
ro
fe
ss
io
n
a
l
e
n
g
in
e
e
r
i
n
t
h
e
d
isc
ip
li
n
e
o
f
tea
c
h
in
g
re
c
o
g
n
ize
d
b
y
th
e
B
o
a
rd
o
f
E
n
g
in
e
e
r
M
a
lay
sia
(BEM
).
S
h
e
is
c
u
rre
n
t
ly
,
th
e
Co
o
rd
i
n
a
to
r
f
o
r
Disc
ip
li
n
e
o
f
S
y
ste
m
En
g
in
e
e
rin
g
,
S
c
h
o
o
l
o
f
El
e
c
tri
c
a
l
E
n
g
i
n
e
e
rin
g
,
Co
ll
e
g
e
o
f
En
g
in
e
e
ri
n
g
Un
i
v
e
r
siti
Tek
n
o
lo
g
i
M
ARA
.
He
r
c
u
rre
n
t
re
se
a
rc
h
in
tere
sts in
c
l
u
d
e
M
EM
S
se
n
s
o
rs,
m
icro
fl
u
id
ic an
d
a
rti
ficia
l
in
telli
g
e
n
c
e
.
S
iti
Aisha
h
Mo
h
a
m
a
d
Ba
d
a
r
u
d
d
in
re
c
e
iv
e
d
t
h
e
B.
S
c
.
d
e
g
re
e
s in
El
e
c
tri
c
a
l
En
g
i
n
e
e
rin
g
fr
o
m
Un
iv
e
rsity
o
f
M
iss
o
u
ri
R
o
ll
a
,
US
A
in
2
0
0
0
.
S
h
e
is
c
u
rre
n
tl
y
a
S
e
n
io
r
E
n
g
in
e
e
r
in
M
IM
OS’s
Ad
v
a
n
c
e
De
v
ice
La
b
o
f
Re
se
a
rc
h
&
De
v
e
l
o
p
m
e
n
t
d
e
p
a
rtme
n
t
sin
c
e
2
0
1
7
.
He
r
m
a
jo
r
re
sp
o
n
si
b
il
it
y
is
h
e
a
d
in
g
th
e
2
D
n
a
n
o
m
a
teria
l
g
r
o
u
p
f
o
r
th
e
d
e
v
e
lo
p
m
e
n
t
o
f
2
DN
M
m
a
in
ly
g
ra
p
h
e
n
e
a
n
d
h
e
x
a
g
o
n
a
l
b
o
r
o
n
n
it
rid
e
(h
BN)
i
n
v
a
rio
u
s
a
p
p
l
ica
ti
o
n
su
c
h
a
s
VO
C,
t
h
e
rm
a
l
m
a
n
a
g
e
m
e
n
t,
a
n
d
G
a
N
p
ro
c
e
ss
tec
h
n
o
l
o
g
y
.
S
h
e
in
v
o
l
v
e
d
i
n
t
h
e
d
e
v
e
lo
p
m
e
n
t
o
f
M
IM
O
S
a
to
m
ize
r
sy
ste
m
a
n
d
it
s
larg
e
a
re
a
G
ra
p
h
e
n
e
sp
ra
y
d
e
p
o
si
ti
o
n
p
r
o
c
e
ss
c
a
p
a
b
il
it
y
.
Ot
h
e
r
m
a
in
re
sp
o
n
si
b
il
it
y
is
p
ro
d
u
c
i
n
g
I
P
s
(
p
a
ten
ts,
c
o
p
y
ri
g
h
ts
a
n
d
trad
e
se
c
re
ts)
a
n
d
p
u
b
li
c
a
ti
o
n
s.
P
a
st
e
x
p
e
rien
c
e
s
in
c
lu
d
e
1
1
y
e
a
rs
a
s
p
ro
c
e
ss
e
n
g
i
n
e
e
rs
o
v
e
rse
e
in
g
t
h
e
p
ro
c
e
ss
d
e
v
e
lo
p
m
e
n
t
,
a
n
d
o
p
ti
m
iza
ti
o
n
f
o
r
t
h
in
fil
m
a
n
d
w
e
t
p
ro
c
e
ss
m
o
d
u
les
a
n
d
3
y
e
a
rs
a
s
re
se
a
r
c
h
e
r
fo
r
CM
OS
&
M
EM
S
p
ro
c
e
ss
g
r
o
u
p
.
Evaluation Warning : The document was created with Spire.PDF for Python.