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.
3
6
,
No
.
2
,
N
o
v
em
b
e
r
20
24
,
p
p
.
1
0
7
0
~
1
082
I
SS
N:
2
502
-
4
7
52
,
DOI
: 1
0
.
1
1
5
9
1
/ijee
cs
.v
3
6
.
i
2
.
pp
1
0
7
0
-
1
0
8
2
1070
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs
.
ia
esco
r
e.
co
m
Foreca
sting
resea
rch inf
luence
:
a
r
ecurren
t
neu
ra
l n
etwork
a
ppro
a
ch t
o
cit
a
ti
o
n predict
io
n
Na
s
er
J
a
m
a
l
1
,
M
o
ha
m
ma
d
Ala
uthm
a
n
2,
,
M
uh
a
nn
a
d M
a
lhi
s
3
,
Ab
delra
o
uf
M
.
I
s
hta
iwi
1
1
D
e
p
a
r
t
me
n
t
o
f
D
a
t
a
S
c
i
e
n
c
e
a
n
d
A
r
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
F
a
c
u
l
t
y
o
f
I
n
f
o
r
ma
t
i
o
n
Te
c
h
n
o
l
o
g
y
,
U
n
i
v
e
r
si
t
y
o
f
P
e
t
r
a
,
A
mm
a
n
,
J
o
r
d
a
n
2
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
mat
i
o
n
S
e
c
u
r
i
t
y
,
F
a
c
u
l
t
y
o
f
I
n
f
o
r
ma
t
i
o
n
T
e
c
h
n
o
l
o
g
y
,
U
n
i
v
e
r
si
t
y
o
f
P
e
t
r
a
,
A
mm
a
n
,
J
o
r
d
a
n
3
D
i
r
e
c
t
o
r
o
f
I
C
T,
U
n
i
v
e
r
si
t
y
o
f
P
e
t
r
a
,
A
mm
a
n
,
J
o
r
d
a
n
Art
icle
I
n
fo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Dec
16
,
2
0
2
3
R
ev
is
ed
J
u
n
10
,
2
0
2
4
Acc
ep
ted
J
u
n
25
,
2
0
2
4
As
th
e
v
o
lu
m
e
o
f
sc
ien
ti
fic
p
u
b
l
ica
ti
o
n
s
c
o
n
ti
n
u
e
s
to
p
r
o
li
fe
ra
te,
e
ffe
c
ti
v
e
e
v
a
lu
a
ti
o
n
to
o
ls
to
d
e
term
in
e
th
e
imp
a
c
t
a
n
d
q
u
a
li
t
y
o
f
re
se
a
rc
h
a
rti
c
les
a
re
in
c
re
a
sin
g
ly
n
e
c
e
ss
a
ry
.
Cit
a
ti
o
n
s
se
rv
e
a
s
a
wid
e
ly
u
ti
li
z
e
d
m
e
tri
c
fo
r
g
a
u
g
i
n
g
sc
ien
ti
fic
imp
a
c
t.
Ho
w
e
v
e
r,
a
c
c
u
ra
tely
p
ro
g
n
o
stica
ti
n
g
th
e
lo
n
g
-
term
c
it
a
ti
o
n
imp
a
c
t
o
f
n
a
sc
e
n
t
p
u
b
li
sh
e
d
re
se
a
rc
h
p
re
se
n
ts
a
f
o
rm
id
a
b
le
c
h
a
ll
e
n
g
e
d
u
e
t
o
t
h
e
in
tri
c
a
c
y
a
n
d
u
n
p
re
d
icta
b
i
li
ty
in
n
a
te
to
th
e
sc
ien
ti
fic
e
c
o
sy
ste
m
.
S
o
p
h
isti
c
a
ted
m
a
c
h
in
e
lea
rn
in
g
m
e
th
o
d
o
l
o
g
ies
,
p
a
rti
c
u
larly
re
c
u
rre
n
t
n
e
u
ra
l
n
e
two
r
k
s
(RNN
s),
h
a
v
e
re
c
e
n
tl
y
d
e
m
o
n
stra
ted
p
ro
m
isin
g
p
o
ten
ti
a
l
i
n
a
d
d
re
ss
in
g
th
is
tas
k
.
Th
is
re
se
a
rc
h
p
ro
p
o
se
s
a
n
RNN
a
rc
h
it
e
c
tu
re
lev
e
ra
g
in
g
e
n
c
o
d
e
r
-
d
e
c
o
d
e
r
se
q
u
e
n
c
e
m
o
d
e
li
n
g
c
a
p
a
b
il
it
ies
to
i
n
g
e
s
t
h
isto
rica
l
c
h
r
o
n
icle
s
a
n
d
p
re
d
ic
t
su
c
c
e
e
d
in
g
e
v
o
lu
ti
o
n
v
ia
late
n
t
tem
p
o
ra
l
d
y
n
a
m
ics
lea
rn
i
n
g
.
C
o
m
p
a
ra
ti
v
e
a
n
a
ly
sis
b
e
twe
e
n
t
h
e
RNN
a
p
p
ro
a
c
h
a
n
d
b
a
se
li
n
e
s,
in
c
lu
d
in
g
ra
n
d
o
m
fo
r
e
st,
su
p
p
o
rt
v
e
c
to
r
re
g
re
ss
io
n
,
a
n
d
m
u
lt
i
-
lay
e
r
p
e
rc
e
p
tro
n
,
d
e
m
o
n
stra
te
su
p
e
rio
r
p
e
rfo
rm
a
n
c
e
o
n
u
n
se
e
n
tes
t
d
a
ta
a
n
d
rig
o
r
o
u
s
k
-
fo
l
d
c
ro
ss
-
v
a
li
d
a
ti
o
n
.
On
a
c
o
rp
u
s
fr
o
m
P
e
tra
Un
iv
e
rsity
,
th
e
RNN
m
e
th
o
d
o
lo
g
y
a
tt
a
in
e
d
th
e
lo
we
st
e
rro
rs
(
r
o
o
t
m
e
a
n
sq
u
a
re
d
e
rro
r
(
RM
S
E
)
1
.
8
4
)
a
n
d
h
i
g
h
e
st
a
c
c
u
ra
c
y
(0
.
9
1
),
a
re
a
u
n
d
e
r
th
e
c
u
r
v
e
(
AUC
)
(0
.
9
6
),
a
n
d
F
1
-
sc
o
re
(
0
.
9
2
).
S
tatisti
c
a
l
t
e
sts
fu
rth
e
r
v
e
rify
s
ig
n
ifi
c
a
n
t
imp
ro
v
e
m
e
n
ts.
Th
e
f
in
d
in
g
s
v
a
li
d
a
te
o
u
r
d
e
e
p
lea
rn
i
n
g
so
l
u
ti
o
n
'
s
e
ffica
c
y
,
ro
b
u
stn
e
ss
,
a
n
d
re
a
l
-
wo
rl
d
v
i
a
b
il
it
y
f
o
r
l
o
n
g
-
term
sc
ien
ti
fi
c
imp
a
c
t
q
u
a
n
t
ifi
c
a
ti
o
n
to
a
id
sta
k
e
h
o
ld
e
rs
in
re
se
a
rc
h
e
v
a
lu
a
ti
o
n
.
T
h
e
fin
d
i
n
g
s
in
ti
m
a
te
th
a
t
RNN
-
b
a
se
d
p
re
d
ictiv
e
m
o
d
e
li
n
g
c
o
n
st
it
u
tes
a
p
o
te
n
t
tec
h
n
o
l
o
g
y
fo
r
c
it
a
ti
o
n
-
d
r
iv
e
n
sc
i
e
n
ti
fic i
m
p
a
c
t
q
u
a
n
ti
f
ica
ti
o
n
.
K
ey
w
o
r
d
s
:
B
ib
lio
m
etr
ics
C
itatio
n
p
r
ed
ictio
n
Dee
p
lear
n
in
g
R
ec
u
r
r
en
t
n
eu
r
al
n
etwo
r
k
s
Sch
o
lar
ly
im
p
ac
t
Scien
tific
r
esear
ch
im
p
ac
t
Scien
to
m
etr
ics
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
:
Mo
h
am
m
ad
Alau
t
h
m
an
Dep
ar
tm
en
t o
f
I
n
f
o
r
m
atio
n
Secu
r
ity
,
Facu
lty
o
f
I
n
f
o
r
m
atio
n
T
ec
h
n
o
lo
g
y
,
Un
i
v
er
s
ity
o
f
Pet
r
a
Am
m
an
,
J
o
r
d
a
n
E
m
ail:
m
o
h
am
m
a
d
.
alau
th
m
aa
n
@
u
o
p
.
ed
u
.
jo
1.
I
NT
RO
D
UCT
I
O
N
As
th
e
v
o
lu
m
e
o
f
s
cien
tific
p
u
b
licatio
n
s
co
n
ti
n
u
es
to
g
r
o
w,
th
er
e
is
a
g
r
o
win
g
n
ee
d
f
o
r
to
o
ls
to
ev
alu
ate
th
e
q
u
ality
an
d
im
p
a
ct
o
f
r
esear
ch
ar
ticles.
C
ita
tio
n
s
ar
e
wid
ely
u
s
ed
to
m
ea
s
u
r
e
s
cien
tific
im
p
ac
t
an
d
ar
e
th
e
b
asis
f
o
r
m
an
y
o
th
er
m
etr
ics,
s
u
ch
as
th
e
h
-
in
d
ex
[
1
]
.
Ho
wev
e
r
,
p
r
e
d
ictin
g
th
e
lo
n
g
-
ter
m
citatio
n
im
p
ac
t
o
f
r
ec
en
tly
p
u
b
lis
h
ed
r
esear
ch
is
ch
allen
g
i
n
g
d
u
e
to
th
e
c
o
m
p
lex
ity
an
d
u
n
p
r
ed
ictab
ilit
y
o
f
th
e
s
cien
tific
lan
d
s
ca
p
e.
Ma
ch
in
e
lear
n
in
g
tech
n
iq
u
es,
s
p
ec
if
ical
ly
r
ec
u
r
r
e
n
t
n
eu
r
al
n
etwo
r
k
s
,
h
a
v
e
r
ec
en
tly
b
ee
n
ap
p
lied
to
th
is
task
with
p
r
o
m
i
s
in
g
r
esu
lts
.
T
h
e
g
lo
b
al
s
cien
tific
co
m
m
u
n
i
ty
'
s
en
g
ag
em
en
t in
d
ev
el
o
p
in
g
an
d
p
u
b
lis
h
in
g
r
esear
c
h
ar
ticl
es h
as led
to
a
p
r
o
lific
o
u
tp
u
t
o
f
s
cien
ti
f
ic
p
ap
e
r
s
,
ea
ch
v
ar
y
in
g
in
q
u
ality
an
d
im
p
ac
t.
T
h
is
p
r
o
f
u
s
io
n
n
ec
ess
itates
a
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:
2
5
0
2
-
4
7
52
F
o
r
ec
a
s
tin
g
r
esea
r
ch
in
flu
en
ce
:
a
r
ec
u
r
r
en
t n
eu
r
a
l n
etw
o
r
k
…
(
N
a
s
er J
a
ma
l
)
1071
r
o
b
u
s
t
s
y
s
tem
to
e
v
alu
ate
an
d
d
is
ce
r
n
th
e
q
u
ality
o
f
t
h
ese
p
u
b
lis
h
ed
wo
r
k
s
.
W
h
ile
th
er
e
ar
e
n
u
m
er
o
u
s
m
eth
o
d
s
to
ju
d
g
e
th
e
q
u
ality
o
f
a
s
cien
tific
p
ap
er
,
t
h
e
n
u
m
b
e
r
o
f
citatio
n
s
it r
ec
eiv
es is
wid
ely
r
ec
o
g
n
ized
as a
cr
u
cial
m
ea
s
u
r
e.
C
itatio
n
co
u
n
t
s
ig
n
if
ican
tly
in
d
icate
s
a
p
ap
er
'
s
in
f
lu
en
ce
a
n
d
r
ea
ch
with
in
th
e
ac
ad
em
ic
co
m
m
u
n
ity
.
B
ey
o
n
d
citatio
n
co
u
n
t,
o
th
er
m
etr
ics ar
e
also
em
p
lo
y
e
d
to
a
s
s
es
s
s
ch
o
lar
ly
im
p
ac
t,
in
clu
d
i
n
g
[
2
]
-
[
5
]
:
−
H
-
I
n
d
ex
:
t
h
is
m
etr
ic
m
ea
s
u
r
es
a
r
esear
ch
er
'
s
p
u
b
licatio
n
s
'
p
r
o
d
u
ctiv
ity
an
d
citatio
n
i
m
p
ac
t.
A
h
ig
h
h
-
in
d
ex
in
d
icate
s
a
co
n
s
is
ten
t
p
r
o
d
u
ctio
n
o
f
in
f
lu
e
n
tial p
ap
er
s
.
−
I
m
p
ac
t
f
ac
to
r
:
o
f
ten
u
s
ed
f
o
r
jo
u
r
n
als,
t
h
e
im
p
ac
t
f
ac
to
r
m
ea
s
u
r
es
th
e
av
er
a
g
e
n
u
m
b
e
r
o
f
citatio
n
s
to
ar
ticles
p
u
b
lis
h
ed
in
a
p
ar
ticu
lar
jo
u
r
n
al.
I
t
is
u
s
ed
to
c
o
m
p
ar
e
th
e
r
elativ
e
im
p
o
r
tan
ce
o
f
jo
u
r
n
als
with
in
s
p
ec
if
ic
f
ield
s
.
−
I
-
1
0
i
n
d
e
x
:
t
h
is
in
d
ex
tr
ac
k
s
t
h
e
n
u
m
b
er
o
f
p
u
b
licatio
n
s
b
y
an
au
th
o
r
with
at
least
ten
cita
tio
n
s
.
I
t
h
elp
s
in
u
n
d
er
s
tan
d
i
n
g
th
e
b
r
ea
d
th
o
f
s
ig
n
if
ican
t c
o
n
t
r
ib
u
tio
n
s
b
y
a
r
esear
ch
er
.
T
h
ese
m
etr
ics,
co
llectiv
ely
w
ith
citatio
n
co
u
n
ts
,
o
f
f
er
a
m
u
ltid
im
en
s
io
n
al
v
iew
o
f
a
p
a
p
er
'
s
o
r
r
esear
ch
er
'
s
im
p
ac
t,
aid
in
g
i
n
id
en
tif
y
in
g
h
ig
h
-
q
u
ality
,
in
f
lu
e
n
tial scien
tific
wo
r
k
.
W
e
lo
o
k
at
th
e
p
r
o
b
lem
o
f
d
et
er
m
in
in
g
h
o
w
o
f
ten
a
s
cien
tific
p
ap
er
will
b
e
cited
.
T
h
is
p
r
o
b
lem
ca
n
b
e
u
s
ed
in
m
an
y
d
if
f
er
e
n
t
ar
ea
s
.
W
i
th
th
e
n
u
m
b
er
o
f
p
u
b
lis
h
ed
d
o
cu
m
en
ts
g
o
in
g
u
p
,
r
esear
ch
er
s
n
ee
d
to
k
n
o
w
wh
ich
p
a
p
er
s
will
b
e
th
e
m
o
s
t
im
p
o
r
tan
t
s
o
th
e
y
ca
n
p
lan
th
e
d
ir
ec
tio
n
o
f
th
eir
r
esear
ch
[
6
]
.
B
y
g
u
ess
in
g
h
o
w
m
an
y
tim
es
a
p
ap
er
will
b
e
cited
in
th
e
f
u
tu
r
e,
we
ca
n
also
f
ig
u
r
e
o
u
t
h
o
w
im
p
o
r
tan
t
th
e
p
ap
er
'
s
au
th
o
r
s
will
b
e
.
T
h
is
c
o
u
ld
h
elp
u
s
h
ir
e
r
esear
ch
er
s
an
d
p
r
o
f
ess
o
r
s
an
d
g
iv
e
awa
r
d
s
an
d
f
u
n
d
s
.
Ma
n
y
attem
p
ts
h
av
e
b
ee
n
m
ad
e
to
d
eter
m
in
e
h
o
w
r
esear
ch
er
s
'
wo
r
k
w
ill af
f
ec
t th
e
f
u
t
u
r
e
[
7
]
.
T
h
e
m
o
tiv
atio
n
f
o
r
th
e
r
esear
ch
o
n
p
r
ed
ictin
g
citatio
n
n
u
m
b
er
s
u
s
in
g
r
ec
u
r
r
en
t
n
eu
r
al
n
etwo
r
k
(
R
NN)
lear
n
in
g
tech
n
iq
u
es is
th
e
n
ee
d
f
o
r
r
ev
iew
an
d
q
u
alit
y
ass
ess
m
en
t to
o
ls
f
o
r
r
esear
ch
ar
ticles in
th
e
f
ac
e
o
f
a
g
r
o
win
g
n
u
m
b
e
r
o
f
s
cien
tific
p
u
b
licatio
n
s
wo
r
ld
wid
e.
T
h
e
s
h
ee
r
v
o
l
u
m
e
o
f
s
cien
tific
liter
atu
r
e
m
ak
es
it
d
if
f
icu
lt
f
o
r
r
esear
c
h
er
s
an
d
s
ch
o
lar
s
to
k
ee
p
u
p
with
th
e
latest
f
ield
d
ev
elo
p
m
en
ts
.
Q
u
an
titativ
e
an
aly
tic
m
eth
o
d
s
an
d
m
et
r
ics
h
av
e
b
ee
n
d
ev
elo
p
ed
f
o
r
ev
alu
atin
g
s
cien
ti
f
ic
wo
r
k
s
b
y
s
cien
tific
f
ield
s
,
in
clu
d
in
g
b
ib
lio
m
etr
ics,
in
f
o
r
m
et
r
ic,
an
d
s
cien
tific
m
etr
ics.
On
e
o
f
th
e
m
o
s
t
cr
itical
m
ea
s
u
r
es
in
th
is
co
n
tex
t
is
th
e
n
u
m
b
er
o
f
citatio
n
s
to
wo
r
k
.
T
h
e
ab
ilit
y
to
p
r
ed
ict
th
e
lo
n
g
-
ter
m
im
p
ac
t
o
f
r
ec
e
n
tly
p
u
b
li
s
h
ed
r
esear
ch
i
s
o
f
g
r
ea
t
s
ig
n
if
ican
ce
,
p
r
im
ar
ily
b
ec
au
s
e
citatio
n
c
o
u
n
ts
ar
e
a
co
r
n
er
s
to
n
e
in
ass
ess
in
g
s
cien
tific
ar
ticles
an
d
f
o
r
m
th
e
f
o
u
n
d
atio
n
f
o
r
v
a
r
io
u
s
o
th
er
m
etr
ics,
in
clu
d
in
g
th
e
h
-
in
d
ex
.
Ho
wev
er
,
ac
cu
r
ately
f
o
r
ec
asti
n
g
th
e
en
d
u
r
in
g
in
f
lu
en
ce
o
f
n
ew
s
ch
o
lar
ly
wo
r
k
s
p
o
s
es
a
s
ig
n
if
ican
t
ch
allen
g
e.
An
ea
r
ly
d
is
tin
ctio
n
o
f
p
u
b
licatio
n
s
in
to
ca
teg
o
r
ies
o
f
im
p
o
r
tan
ce
o
r
tr
iv
iality
co
u
ld
h
av
e
co
n
s
id
er
ab
le
ap
p
licatio
n
s
.
T
h
er
ef
o
r
e,
d
ev
is
in
g
ac
cu
r
ate
m
eth
o
d
s
to
esti
m
ate
th
e
f
u
t
u
r
e
citatio
n
n
u
m
b
er
s
o
f
r
esear
c
h
p
ap
e
r
s
is
c
r
u
cial.
T
h
is
ca
p
a
b
ilit
y
wo
u
ld
e
n
ab
le
id
en
tify
in
g
th
e
m
o
s
t
im
p
ac
tf
u
l
an
d
p
er
tin
en
t
r
esear
ch
,
th
er
eb
y
f
ac
ilit
atin
g
r
esear
ch
er
s
an
d
s
ch
o
lar
s
in
k
ee
p
in
g
ab
r
ea
s
t o
f
th
e
latest ad
v
a
n
ce
m
en
ts
with
in
th
eir
r
esp
ec
tiv
e
f
ie
ld
s
.
T
h
e
p
r
im
a
r
y
k
n
o
wled
g
e
g
ap
ad
d
r
ess
ed
b
y
r
esear
ch
o
n
f
o
r
ec
asti
n
g
citatio
n
n
u
m
b
e
r
s
v
ia
R
NN
tech
n
iq
u
es
p
er
tai
n
s
to
th
e
in
a
d
eq
u
ac
y
o
f
p
r
ec
is
e
an
d
d
ep
e
n
d
ab
le
m
eth
o
d
o
lo
g
ies
f
o
r
f
o
r
ec
asti
n
g
th
e
en
d
u
r
in
g
in
f
lu
en
ce
o
f
n
ewly
p
u
b
lis
h
ed
s
ch
o
lar
ly
wo
r
k
s
.
Alth
o
u
g
h
cit
atio
n
tallies
ar
e
a
p
r
ev
alen
t
m
etr
ic
f
o
r
ev
al
u
atin
g
s
cien
tific
ar
ticles
an
d
u
n
d
er
p
in
n
u
m
er
o
u
s
o
th
er
in
d
icato
r
s
,
p
r
ec
is
ely
an
ticip
atin
g
th
e
lo
n
g
-
ter
m
citatio
n
co
u
n
t
o
f
a
r
esear
ch
p
a
p
er
r
em
ain
s
a
f
o
r
m
id
a
b
le
ch
allen
g
e
.
T
h
is
la
cu
n
a
u
n
d
er
s
co
r
es
t
h
e
n
ec
ess
ity
f
o
r
m
o
r
e
ac
cu
r
ate
an
d
r
eliab
le
ap
p
r
o
ac
h
es to
p
r
e
d
ict
th
e
s
u
s
tain
ed
im
p
ac
t o
f
s
cien
tific
p
u
b
licatio
n
s
.
T
h
e
ex
p
ec
ted
c
o
n
tr
ib
u
tio
n
o
f
th
e
r
esear
ch
o
n
p
r
ed
ictin
g
citatio
n
n
u
m
b
er
s
u
s
in
g
R
N
N
lear
n
in
g
tech
n
iq
u
es
is
d
e
v
elo
p
in
g
a
m
eth
o
d
f
o
r
ca
lcu
latin
g
a
m
an
u
s
cr
ip
t'
s
lo
n
g
-
ter
m
citatio
n
s
.
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
u
tili
ze
s
an
ar
tific
ial
n
eu
r
al
n
e
two
r
k
(
ANN)
,
s
p
ec
if
ically
a
R
NN
,
to
p
r
ed
ict
th
e
n
u
m
b
er
o
f
citatio
n
s
a
p
ap
e
r
will
o
b
tain
in
th
e
f
u
tu
r
e
b
a
s
ed
o
n
its
in
itial
citatio
n
co
u
n
ts
.
T
h
e
m
eth
o
d
o
u
tp
e
r
f
o
r
m
s
s
tate
-
of
-
th
e
-
ar
t
tech
n
iq
u
es
r
eg
a
r
d
in
g
f
o
r
ec
ast
ac
cu
r
ac
y
f
o
r
y
e
a
r
ly
an
d
o
v
er
all
esti
m
ates
o
f
th
e
n
u
m
b
e
r
o
f
citatio
n
s
.
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
ca
n
ass
is
t
i
n
id
en
tify
i
n
g
th
e
m
o
s
t
im
p
ac
t
f
u
l
an
d
r
elev
an
t
r
esear
ch
p
ap
e
r
s
,
m
ak
in
g
it
ea
s
ier
f
o
r
r
esear
ch
e
r
s
an
d
s
ch
o
lar
s
t
o
s
tay
u
p
-
to
-
d
ate
with
th
e
late
s
t
d
ev
elo
p
m
en
ts
in
th
eir
f
ie
ld
s
.
Fu
r
th
er
m
o
r
e,
th
e
p
r
o
p
o
s
ed
m
eth
o
d
m
ay
b
e
h
elp
f
u
l
f
o
r
s
cien
tific
in
s
titu
tio
n
s
,
f
u
n
d
in
g
ag
e
n
cies,
an
d
p
o
licy
m
ak
er
s
in
e
v
alu
atin
g
th
e
im
p
ac
t o
f
s
cien
tific
r
esear
c
h
an
d
allo
ca
tin
g
r
eso
u
r
ce
s
ac
co
r
d
in
g
l
y
.
I
n
th
is
p
ap
er
,
we
s
u
g
g
est
a
w
ay
to
f
ig
u
r
e
o
u
t
h
o
w
m
a
n
y
ti
m
es
a
s
cien
tific
p
ap
er
will
b
e
cited
b
ased
o
n
h
o
w
m
a
n
y
tim
es
it
is
cited
in
its
f
ir
s
t
f
ew
y
ea
r
s
.
I
n
o
th
e
r
wo
r
d
s
,
t
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
lo
o
k
s
at
h
o
w
m
a
n
y
tim
es
a
p
ap
er
is
cited
th
r
ee
y
ea
r
s
af
ter
it
co
m
es
o
u
t
an
d
p
r
ed
icts
h
o
w
m
an
y
tim
es
it
will
b
e
cited
.
W
e
o
n
ly
in
p
u
t
th
e
ea
r
l
y
p
u
b
licatio
n
y
e
ar
citatio
n
p
atter
n
i
n
th
is
p
r
o
b
lem
.
I
n
o
u
r
f
r
am
ewo
r
k
,
we
m
ad
e
a
cu
s
to
m
ize
d
R
NN
to
d
eter
m
in
e
th
e
citatio
n
co
u
n
t.
On
e
o
f
th
e
p
a
r
am
o
u
n
t
ch
allen
g
es
with
in
b
ib
lio
m
etr
ics
is
th
e
f
o
r
ec
ast
o
f
th
e
im
p
ac
t
an
d
s
i
g
n
if
ican
ce
o
f
n
a
s
ce
n
t
s
cien
tific
p
u
b
licatio
n
s
.
C
itatio
n
f
r
eq
u
e
n
cy
,
as
a
g
au
g
e
o
f
s
cien
tific
in
f
lu
e
n
ce
,
is
f
u
n
d
am
en
tal,
a
n
d
th
e
lo
n
g
-
ter
m
citatio
n
p
r
ed
ict
io
n
f
o
r
a
p
ap
e
r
h
o
l
d
s
s
u
b
s
tan
tial
im
p
o
r
tan
ce
.
Pre
cise
p
r
e
d
ictio
n
o
f
a
p
ap
er
'
s
citatio
n
im
p
ac
t
is
in
s
tr
u
m
en
tal
f
o
r
r
esea
r
c
h
er
s
an
d
p
o
licy
m
a
k
er
s
in
id
en
tify
i
n
g
p
i
v
o
tal
an
d
r
elev
an
t
r
esear
ch
,
g
u
id
in
g
r
eso
u
r
ce
allo
ca
tio
n
,
a
n
d
s
tr
ateg
izin
g
f
u
tu
r
e
r
esear
ch
tr
ajec
to
r
ies.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
In
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
3
6
,
No
.
2
,
No
v
em
b
er
20
24
:
1
0
7
0
-
1
0
8
2
1072
T
h
e
en
s
u
in
g
s
ec
tio
n
s
o
f
t
h
is
p
ap
er
ar
e
s
tr
u
ct
u
r
ed
as
f
o
llo
ws:
s
ec
tio
n
2
o
f
f
er
s
an
in
-
d
e
p
th
r
ev
iew
o
f
ex
is
tin
g
liter
atu
r
e
in
citatio
n
p
r
ed
ictio
n
,
en
c
o
m
p
ass
in
g
b
o
t
h
s
tatis
tical
an
d
m
ac
h
in
e
-
lear
n
in
g
m
eth
o
d
o
lo
g
ies.
Sectio
n
3
d
elin
ea
tes
o
u
r
p
r
o
p
o
s
ed
m
eth
o
d
s
f
o
r
citatio
n
p
r
ed
ictio
n
,
h
i
g
h
lig
h
tin
g
th
e
d
i
s
tin
ct
f
ea
tu
r
es
an
d
alg
o
r
ith
m
s
in
co
r
p
o
r
ated
in
o
u
r
m
o
d
el.
I
n
s
ec
tio
n
4
,
we
elu
cid
ate
th
e
o
u
tco
m
es
o
f
o
u
r
ex
p
er
im
en
tal
s
tu
d
ies,
in
clu
d
in
g
a
co
m
p
ar
ativ
e
an
al
y
s
is
o
f
o
u
r
m
eth
o
d
a
g
ain
s
t
o
th
er
lead
in
g
-
ed
g
e
tech
n
i
q
u
es.
Fin
ally
,
s
ec
tio
n
5
d
elv
es
in
to
th
e
b
r
o
ad
er
im
p
licatio
n
s
o
f
o
u
r
f
in
d
in
g
s
an
d
p
r
o
p
o
s
es
av
en
u
es
f
o
r
p
r
o
s
p
ec
ti
v
e
r
es
ea
r
ch
in
th
is
d
o
m
ain
.
2.
RE
L
AT
E
D
WO
RK
Nu
m
er
o
u
s
en
d
ea
v
o
r
s
h
av
e
b
ee
n
u
n
d
er
tak
en
t
o
p
r
e
d
ict
th
e
s
u
cc
ess
o
f
s
cien
tific
wo
r
k
s
,
v
ar
y
in
g
s
ig
n
if
ican
tly
in
th
eir
m
eth
o
d
o
l
o
g
ies an
d
o
u
tc
o
m
es.
E
x
is
tin
g
r
esear
ch
in
th
is
d
o
m
ain
h
as f
o
cu
s
ed
o
n
p
r
e
d
ictin
g
d
iv
er
s
e
m
etr
ics.
T
h
ese
in
clu
d
e
esti
m
atin
g
th
e
to
tal
citatio
n
co
u
n
t
a
s
p
ec
if
ic
s
cien
tific
p
a
p
er
will
r
ec
eiv
e,
as
ex
p
lo
r
ed
in
r
e
f
er
en
ce
s
[
8
]
,
[
9
]
;
f
o
r
ec
asti
n
g
th
e
citatio
n
n
u
m
b
er
s
f
o
r
a
s
elec
ted
g
r
o
u
p
o
f
h
ig
h
ly
cited
p
ap
er
s
,
d
is
cu
s
s
ed
in
[
1
0
]
; p
r
e
d
ictin
g
a
n
in
d
iv
id
u
al
r
esear
ch
er
'
s
h
-
in
d
ex
,
as p
er
[
1
1
]
; a
n
d
ass
ess
in
g
t
h
e
im
p
ac
t f
ac
to
r
o
f
a
s
et
o
f
s
cien
tific
jo
u
r
n
als,
wh
i
ch
is
th
e
s
u
b
ject
o
f
[
1
2
]
.
T
o
p
r
o
ject
th
e
n
u
m
b
er
o
f
ci
tatio
n
s
an
au
th
o
r
m
ig
h
t
ac
c
r
u
e
o
v
e
r
a
f
o
r
th
c
o
m
in
g
n
-
y
e
ar
p
er
io
d
,
Ma
zlo
u
m
ian
'
s
s
tu
d
y
[
1
3
]
in
c
o
r
p
o
r
ates
a
r
an
g
e
o
f
au
th
o
r
-
s
p
ec
if
ic
ch
ar
ac
ter
is
tics
.
T
h
ese
en
co
m
p
ass
th
e
to
tal
n
u
m
b
er
o
f
p
ap
er
s
a
u
th
o
r
e
d
,
t
h
e
av
er
a
g
e
an
n
u
al
citatio
n
r
a
te,
an
d
th
e
au
th
o
r
'
s
h
-
in
d
ex
.
Su
ch
an
a
p
p
r
o
a
ch
u
n
d
er
s
co
r
es
th
e
m
u
ltifa
ce
ted
n
atu
r
e
o
f
b
ib
lio
m
etr
ic
an
aly
s
es,
wh
er
e
b
o
th
q
u
an
titativ
e
o
u
t
p
u
t
an
d
q
u
alitativ
e
im
p
ac
t a
r
e
co
n
s
id
er
e
d
to
e
v
alu
ate
s
cien
tific
in
f
lu
en
ce
an
d
s
u
cc
ess
.
C
asti
l
lo
et
a
l.
[
1
4
]
u
s
e
th
e
au
th
o
r
s
'
p
r
io
r
p
u
b
licatio
n
s
an
d
th
e
co
au
th
o
r
-
s
h
ip
n
etwo
r
k
t
o
f
o
r
etell
a
p
ap
er
'
s
citatio
n
co
u
n
t
in
th
e
f
ir
s
t
f
ew
y
ea
r
s
o
f
p
u
b
licatio
n
.
I
n
th
ei
r
wo
r
k
,
B
o
r
n
m
a
n
n
et
a
l.
[
1
5
]
r
ely
o
n
n
u
m
er
o
u
s
au
th
o
r
s
,
citatio
n
s
,
a
n
d
citatio
n
s
f
r
o
m
o
t
h
er
w
o
r
k
s
.
Sp
ec
if
ically
,
we
f
o
c
u
s
o
n
h
o
w
o
f
ten
a
s
cien
tific
wo
r
k
h
as b
ee
n
m
en
tio
n
ed
in
th
e
p
ast s
ev
er
al
y
ea
r
s
.
T
h
er
e
is
n
o
o
th
e
r
co
n
s
id
er
atio
n
.
M
an
s
o
u
r
et
a
l.
[
1
6
]
cr
ea
te
d
a
m
ac
h
in
e
-
lear
n
in
g
a
p
p
r
o
ac
h
i
n
th
eir
2
0
1
9
s
tu
d
y
f
o
r
p
r
o
ject
in
g
f
u
t
u
r
e
r
esear
ch
p
a
p
er
m
en
tio
n
s
.
T
h
e
y
m
ad
e
u
s
e
o
f
d
ata
c
o
v
er
in
g
a
d
ec
ad
e
f
r
o
m
t
h
e
I
n
ter
n
atio
n
al
Ar
ab
J
o
u
r
n
al
o
f
I
n
f
o
r
m
atio
n
T
e
ch
n
o
l
o
g
y
an
d
t
ested
s
ix
teen
m
ac
h
in
e
-
lear
n
in
g
alg
o
r
ith
m
s
.
T
h
e
f
in
d
in
g
s
o
f
th
eir
s
tu
d
y
s
h
o
wed
th
at
th
e
s
ig
n
if
ica
n
ce
o
f
f
o
r
ec
a
s
tin
g
f
u
tu
r
e
m
en
ti
o
n
s
lies
m
o
r
e
in
th
e
n
u
m
b
er
o
f
r
ef
e
r
en
ce
s
th
an
th
e
n
u
m
b
er
o
f
wr
iter
s
.
Ou
t
o
f
all
test
ed
alg
o
r
ith
m
s
,
n
eu
r
al
n
etwo
r
k
an
d
v
o
tin
g
class
if
ier
1
ca
m
e
o
u
t
a
h
ea
d
f
o
r
f
o
r
ec
asti
n
g
f
u
tu
r
e
m
en
tio
n
s
.
Seco
n
d
ly
was
Naïv
e
B
ay
es,
with
o
th
er
s
p
er
f
o
r
m
in
g
o
n
a
co
m
p
ar
ab
le
lev
el.
T
h
is
r
esear
c
h
m
ar
k
s
a
n
o
ta
b
le
ad
v
a
n
ce
in
th
e
u
s
e
o
f
m
ac
h
in
e
lear
n
in
g
f
o
r
b
ib
lio
m
etr
ics.
I
n
th
eir
s
tu
d
y
,
Ab
r
is
h
am
i
an
d
Aliak
b
ar
y
[
1
7
]
d
ev
elo
p
ed
a
way
to
f
o
r
ec
ast
th
e
n
u
m
b
er
o
f
tim
es
th
at
a
r
esear
ch
p
a
p
er
will
b
e
cited
o
v
e
r
th
e
lo
n
g
te
r
m
.
I
n
s
tead
o
f
r
ely
in
g
o
n
th
e
ac
tu
al
c
o
u
n
t
o
f
citatio
n
s
,
an
im
p
r
ac
tical
ap
p
r
o
ac
h
with
a
lo
n
g
lead
tim
e,
to
m
ak
e
th
is
p
r
ed
ictio
n
,
th
e
au
th
o
r
s
tr
ain
a
m
o
d
el
u
s
in
g
ANNs,
a
p
o
wer
f
u
l
m
ac
h
i
n
e
lear
n
in
g
m
eth
o
d
o
lo
g
y
th
at
h
as
b
ee
n
ap
p
lied
s
u
cc
ess
f
u
lly
to
an
in
cr
ea
s
i
n
g
ly
wid
e
r
an
g
e
o
f
task
s
-
m
o
s
t
f
am
o
u
s
ly
in
im
ag
e
an
d
tex
t
p
r
o
ce
s
s
in
g
.
E
m
p
ir
ic
al
ex
p
er
im
en
ts
s
h
o
wed
t
h
at
th
e
p
r
ed
ictio
n
s
m
ad
e
u
s
in
g
ANNs a
r
e,
to
d
ate,
th
e
m
o
s
t a
cc
u
r
ate.
Ma
ts
u
i
e
t
a
l.
[
1
8
]
in
tr
o
d
u
ce
d
a
r
eg
r
ess
io
n
an
aly
s
is
-
b
ased
m
ac
h
in
e
lear
n
in
g
a
p
p
r
o
ac
h
to
p
r
ed
ict
th
e
f
u
tu
r
e
citatio
n
co
u
n
t
o
f
a
r
ese
ar
ch
ar
ticle.
Ah
u
ja
[
1
9
]
p
r
ese
n
ted
two
ty
p
es
o
f
an
aly
s
is
aim
ed
at
p
r
ed
ictin
g
th
e
g
r
o
wth
o
f
u
n
iv
e
r
s
ities
ab
o
v
e
an
d
b
elo
w
th
e
av
er
a
g
e
c
o
n
ce
r
n
in
g
t
h
e
to
tal
n
u
m
b
er
o
f
u
n
i
v
er
s
ities
.
T
h
e
au
th
o
r
em
p
lo
y
ed
a
tr
ain
in
g
d
ataset
f
r
o
m
2
0
1
1
to
2
0
1
6
,
in
cl
u
d
in
g
all
u
n
iv
er
s
ities
'
p
u
b
licatio
n
s
a
n
d
citatio
n
d
etails.
T
h
e
p
r
ed
ictio
n
s
f
o
r
2
0
1
7
an
d
o
n
war
d
s
esti
m
ated
ab
o
v
e
-
av
er
ag
e
g
r
o
wth
in
u
n
iv
er
s
ity
p
u
b
licatio
n
s
an
d
citatio
n
s
b
y
7
.
8
5
% a
n
d
6
.
6
2
%,
r
esp
ec
tiv
ely
.
Su
[
2
0
]
c
o
n
d
u
cted
a
s
tu
d
y
b
ased
o
n
2
,
6
0
0
p
ap
e
r
s
o
n
p
h
y
s
io
lo
g
y
ex
tr
ac
te
d
f
r
o
m
th
e
W
eb
o
f
Scien
ce
.
T
h
e
au
th
o
r
s
elec
ted
eig
h
t
b
ib
l
io
m
etr
ic
f
ea
tu
r
es
o
f
citin
g
p
a
p
er
s
in
th
e
f
ir
s
t
th
r
ee
y
ea
r
s
af
ter
p
u
b
licatio
n
.
T
h
e
au
th
o
r
b
u
ilt
th
r
ee
m
ac
h
in
e
le
ar
n
in
g
m
o
d
els
an
d
a
n
eu
r
al
n
etwo
r
k
to
test
wh
eth
er
th
ese
f
ea
tu
r
es
ef
f
ec
tiv
ely
p
r
ed
icted
f
u
t
u
r
e
citatio
n
co
u
n
t
s
.
T
h
e
ex
p
er
im
en
tal
r
esu
lts
in
d
icate
d
th
at
th
e
s
elec
ted
f
ea
tu
r
es
wer
e
v
alu
ab
le
in
p
r
ed
ictin
g
lo
n
g
-
ter
m
citatio
n
c
o
u
n
ts
,
a
n
d
th
e
m
ac
h
in
e
lear
n
in
g
a
n
d
n
eu
r
al
n
etwo
r
k
m
o
d
els
h
elp
ed
p
r
e
d
ict
f
u
tu
r
e
citatio
n
c
o
u
n
ts
.
Du
[
2
1
]
a
p
p
lied
s
ev
er
al
m
ac
h
in
e
-
lear
n
in
g
tech
n
iq
u
es
t
o
r
an
k
r
esear
ch
in
s
titu
tio
n
s
b
ased
o
n
p
r
ed
ictin
g
t
h
e
n
u
m
b
er
o
f
ac
c
ep
ted
p
a
p
er
s
at
u
p
c
o
m
in
g
to
p
co
n
f
er
en
ce
s
.
T
h
e
au
th
o
r
p
r
o
p
o
s
ed
a
th
r
ee
-
p
h
ase
ex
p
er
im
en
t,
b
e
g
in
n
in
g
with
a
s
im
p
le
av
er
ag
e
m
eth
o
d
an
d
ex
ten
d
in
g
th
e
tr
ain
i
n
g
d
at
aset
b
y
f
in
d
in
g
th
e
s
im
ilar
ity
o
f
co
n
f
er
en
ce
s
en
g
in
ee
r
in
g
t
r
en
d
f
ea
tu
r
es
an
d
u
tili
zin
g
lin
ea
r
r
e
g
r
ess
io
n
,
r
an
k
s
u
p
p
o
r
t
v
ec
t
o
r
m
ac
h
in
e
(
SVM
)
,
an
d
en
s
em
b
l
e
m
o
d
els to
im
p
r
o
v
e
p
r
ed
ictio
n
s
.
W
en
et
a
l.
[
2
2
]
p
r
o
p
o
s
ed
a
citatio
n
n
u
m
b
er
p
r
ed
ictio
n
m
o
d
el,
g
ate
d
r
ec
u
r
r
en
t
u
n
it
-
co
n
tin
u
o
u
s
p
ar
am
eter
m
o
d
e
(
GR
U
-
C
PM
)
,
b
ased
o
n
th
e
R
NN
m
eth
o
d
with
a
g
ated
r
ec
u
r
r
en
t
u
n
it.
T
h
e
au
th
o
r
s
ex
tr
a
cted
f
ea
tu
r
es
f
r
o
m
r
e
al
d
atasets
th
at
ar
e
u
s
ef
u
l
in
p
r
ed
ictin
g
th
e
n
u
m
b
er
o
f
citatio
n
s
in
p
ap
e
r
s
an
d
in
p
u
t
th
em
in
to
th
e
GR
U
-
C
P
M
f
o
r
p
r
ed
ictio
n
.
T
h
ey
co
m
p
ar
ed
th
e
p
r
ed
ictio
n
r
esu
lts
with
o
th
er
r
eg
r
ess
io
n
m
o
d
els
an
d
f
o
u
n
d
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:
2
5
0
2
-
4
7
52
F
o
r
ec
a
s
tin
g
r
esea
r
ch
in
flu
en
ce
:
a
r
ec
u
r
r
en
t n
eu
r
a
l n
etw
o
r
k
…
(
N
a
s
er J
a
ma
l
)
1073
th
at
th
e
GR
U
-
C
PM
h
as
h
ig
h
er
ac
c
u
r
ac
y
an
d
f
aster
c
o
n
v
er
g
e
n
ce
s
p
ee
d
.
Mo
r
eo
v
e
r
,
th
e
GR
U
-
C
PM
o
u
tp
er
f
o
r
m
ed
ex
is
tin
g
m
eth
o
d
s
in
th
e
tim
e
s
er
ies p
r
ed
ictio
n
o
f
citatio
n
co
u
n
t.
C
r
o
f
t
an
d
Sack
[
2
3
]
co
n
d
u
cte
d
a
s
tu
d
y
o
n
two
r
eg
r
ess
io
n
ta
s
k
s
:
p
r
ed
ictin
g
th
e
n
u
m
b
e
r
o
f
citatio
n
s
a
jo
u
r
n
al
will
r
ec
eiv
e
d
u
r
i
n
g
t
h
e
n
ex
t
ca
len
d
a
r
y
ea
r
an
d
p
r
ed
ictin
g
th
e
E
ls
ev
ier
cite
-
co
r
e
a
jo
u
r
n
al
will
b
e
ass
ig
n
ed
f
o
r
th
e
f
o
llo
win
g
ca
len
d
ar
y
ea
r
.
T
h
e
a
u
th
o
r
s
cr
ea
ted
a
d
ataset
o
f
h
is
to
r
ical
b
ib
lio
m
etr
ic
d
ata
f
o
r
jo
u
r
n
als
in
d
ex
ed
in
Sco
p
u
s
an
d
p
r
o
p
o
s
ed
u
s
in
g
n
e
u
r
al
n
et
wo
r
k
m
o
d
els
to
p
r
ed
ict
th
e
f
u
tu
r
e
p
er
f
o
r
m
an
ce
o
f
jo
u
r
n
als.
T
h
ey
p
er
f
o
r
m
e
d
f
ea
tu
r
e
s
elec
tio
n
an
d
m
o
d
el
c
o
n
f
i
g
u
r
atio
n
f
o
r
a
m
u
lti
-
lay
er
p
er
ce
p
tr
o
n
an
d
a
lo
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
.
T
h
ey
co
m
p
ar
ed
th
e
e
x
p
er
im
e
n
tal
r
esu
lts
with
h
eu
r
is
tic
p
r
e
d
ictio
n
b
aselin
es
an
d
class
ic
m
ac
h
in
e
lear
n
in
g
m
o
d
els
.
T
h
e
au
th
o
r
s
f
o
u
n
d
th
at
th
eir
p
r
o
p
o
s
ed
m
o
d
els
f
o
r
p
r
e
d
ictin
g
f
u
tu
r
e
citatio
n
s
an
d
citesco
r
e
v
alu
es o
u
tp
er
f
o
r
m
e
d
th
e
o
th
er
m
o
d
els.
R
u
an
et
a
l.
[
2
4
]
u
tili
ze
d
a
f
o
u
r
-
lay
er
b
ac
k
p
r
o
p
ag
atio
n
(
B
P)
n
eu
r
al
n
etw
o
r
k
m
o
d
el
to
p
r
ed
ict
th
e
f
iv
e
-
y
ea
r
citatio
n
s
o
f
4
9
,
8
3
4
p
ap
er
s
in
th
e
lib
r
ar
y
,
in
f
o
r
m
atio
n
,
an
d
d
o
cu
m
en
tatio
n
f
ield
in
d
ex
ed
b
y
th
e
C
SS
C
I
d
atab
ase
f
r
o
m
2
0
0
0
to
2
0
1
3
.
T
h
e
au
th
o
r
s
ex
tr
ac
te
d
s
ev
er
al
f
ea
tu
r
es
to
p
r
e
d
ict
th
e
citatio
n
s
,
in
clu
d
in
g
p
ap
er
,
jo
u
r
n
al,
au
th
o
r
,
r
ef
er
e
n
ce
,
an
d
ea
r
l
y
citatio
n
f
ea
tu
r
es.
T
h
e
ex
p
er
im
en
tal
r
esu
lts
d
e
m
o
n
s
tr
ated
t
h
at
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
B
P
n
eu
r
a
l
n
etwo
r
k
m
o
d
el
was
s
ig
n
if
i
ca
n
tly
b
etter
th
a
n
th
e
s
ix
b
a
s
elin
e
m
o
d
els.
T
h
e
m
o
d
el
d
em
o
n
s
tr
ated
s
u
p
er
io
r
p
r
o
f
icien
cy
in
f
o
r
ec
asti
n
g
th
e
citatio
n
f
r
eq
u
en
cy
o
f
less
-
r
ef
er
en
ce
d
ac
ad
em
ic
p
ap
er
s
co
m
p
ar
e
d
to
th
o
s
e
f
r
eq
u
en
tly
cited
.
T
h
e
r
esear
ch
d
elin
ea
ted
f
iv
e
p
iv
o
tal
attr
ib
u
tes
m
ar
k
ed
ly
in
f
lu
en
cin
g
th
e
m
o
d
el'
s
p
r
ed
ictiv
e
ef
f
icac
y
:
th
e
c
o
u
n
t
o
f
citatio
n
s
with
in
th
e
in
itial
two
y
ea
r
s
p
o
s
t
-
p
u
b
licatio
n
,
th
e
a
g
e
wh
en
f
ir
s
t
cited
,
th
e
o
v
er
all
len
g
th
o
f
th
e
p
ap
er
,
th
e
m
o
n
th
o
f
p
u
b
licatio
n
,
an
d
t
h
e
p
r
ev
alen
ce
o
f
s
elf
-
citatio
n
s
wi
th
in
th
e
s
am
e
jo
u
r
n
al.
T
h
ese
f
ac
to
r
s
wer
e
m
o
r
e
im
p
ac
tf
u
l
th
an
o
th
e
r
e
x
am
in
ed
f
ea
tu
r
es in
d
eter
m
i
n
in
g
th
e
m
o
d
el'
s
p
r
ed
ictio
n
ac
cu
r
ac
y
.
As
s
h
o
wn
in
T
ab
le
1
,
p
r
ev
io
u
s
r
esear
ch
h
as
s
h
o
wn
th
at
v
ar
io
u
s
d
ee
p
lea
r
n
in
g
m
e
th
o
d
s
ca
n
ef
f
ec
tiv
ely
p
r
e
d
ict
s
tu
d
en
t
ac
h
iev
em
en
t,
in
clu
d
in
g
co
m
b
i
n
atio
n
s
o
f
co
n
v
o
lu
tio
n
al
an
d
R
NNs
,
atten
tio
n
-
b
ased
r
ec
u
r
r
en
t
n
etwo
r
k
s
,
an
d
h
y
b
r
i
d
d
ee
p
m
o
d
els
with
s
u
p
p
o
r
t
v
ec
to
r
r
eg
r
ess
io
n
.
Ho
wev
er
,
ea
ch
ex
is
tin
g
a
p
p
r
o
ac
h
h
as
lim
itatio
n
s
r
eg
ar
d
in
g
co
m
p
u
tatio
n
al
ex
p
en
s
e,
s
en
s
itiv
ity
to
p
ar
am
ete
r
tu
n
i
n
g
,
a
n
d
a
p
p
l
icab
ilit
y
to
n
ar
r
o
w
f
ea
tu
r
e
s
ets.
T
h
e
cu
r
r
e
n
t
s
tu
d
y
p
r
o
p
o
s
es
a
d
ee
p
-
lear
n
in
g
tech
n
iq
u
e
th
at
lev
er
ag
es
co
n
v
o
lu
tio
n
al
an
d
R
NNs
to
f
u
lly
ca
p
tu
r
e
m
ea
n
i
n
g
f
u
l
p
att
er
n
s
in
s
tu
d
en
t
d
ata
ac
r
o
s
s
ac
h
iev
em
en
t
m
etr
ics
an
d
tim
escales.
Ou
r
ap
p
r
o
ac
h
also
s
tr
ateg
ically
s
p
o
tlig
h
ts
th
e
m
o
s
t
r
elev
an
t
in
p
u
t
f
ea
t
u
r
es
u
s
in
g
atten
tio
n
m
ec
h
a
n
is
m
s
wh
ile
o
f
f
s
ettin
g
th
e
v
an
is
h
in
g
g
r
a
d
ien
t
p
r
o
b
lem
d
ee
p
er
m
o
d
els
f
ac
e.
W
e
ex
p
ec
t
th
at
th
o
u
g
h
tf
u
lly
b
len
d
i
n
g
th
ese
co
m
p
lem
en
tar
y
m
eth
o
d
s
will su
r
p
ass
cu
r
r
en
t p
r
ed
ictiv
e
p
er
f
o
r
m
an
ce
.
M
o
r
eo
v
er
,
b
y
c
o
m
p
r
e
h
en
s
iv
ely
ass
ess
in
g
ac
r
o
s
s
d
iv
er
s
e
d
atasets
,
s
tu
d
en
t
g
r
o
u
p
s
,
a
n
d
s
u
cc
ess
in
d
icato
r
s
,
we
d
em
o
n
s
tr
ate
wid
er
ap
p
licab
ilit
y
co
m
p
ar
ed
t
o
s
p
ec
ialized
ex
is
tin
g
tech
n
iq
u
es.
Ou
r
u
n
if
ied
ap
p
r
o
ac
h
p
r
o
v
id
es
a
n
ad
ap
tab
le
to
o
l
f
o
r
u
n
d
e
r
s
tan
d
in
g
ac
ad
em
ic
ac
h
iev
em
en
t w
ith
in
r
ea
l
-
wo
r
ld
ed
u
ca
tio
n
al
s
ettin
g
s
.
3.
P
RO
P
O
SE
D
AP
P
RO
ACH
No
wad
ay
s
,
s
cien
tific
p
u
b
licatio
n
s
’
ex
p
l
o
s
io
n
m
a
k
es
im
p
o
r
tan
t
wo
r
k
th
e
e
v
alu
atio
n
o
f
t
h
e
im
p
ac
t
an
d
q
u
ality
o
f
r
esear
ch
p
a
p
er
s
.
T
h
e
m
o
s
t
co
m
m
o
n
m
etr
ic
f
o
r
th
e
ev
al
u
atio
n
o
f
s
cien
tific
im
p
ac
t
is
p
r
o
v
i
d
ed
b
y
th
e
n
u
m
b
er
o
f
citatio
n
s
.
B
u
t
it
is
s
till
ch
allen
g
in
g
to
p
r
ed
ict
th
e
l
o
n
g
-
ter
m
citatio
n
im
p
ac
t
o
f
p
a
p
er
s
p
u
b
lis
h
ed
r
ec
en
tly
,
b
ec
a
u
s
e
o
f
th
e
in
t
r
in
s
ically
s
to
ch
asti
c
an
d
u
n
p
r
ed
icta
b
le
n
atu
r
e
o
f
t
h
e
s
cien
tific
en
ter
p
r
is
e.
Mo
r
e
r
ec
en
tly
,
R
NNs
[
2
5
]
h
a
v
e
em
er
g
e
d
as
a
p
o
wer
f
u
l
ap
p
r
o
ac
h
t
o
p
r
e
d
ict
th
e
n
u
m
b
er
o
f
citatio
n
s
.
I
n
t
h
is
s
tu
d
y
,
we
p
r
o
p
o
s
e
a
m
et
h
o
d
b
ased
o
n
R
NNs
f
o
r
p
r
e
d
ictin
g
t
h
e
r
esear
ch
i
m
p
ac
t,
h
er
eb
y
d
ef
in
ed
as
th
e
n
u
m
b
er
o
f
citatio
n
s
p
r
ed
icted
.
I
n
th
e
ev
alu
atio
n
o
f
o
u
r
m
eth
o
d
we
u
s
e
th
r
ee
b
aselin
e
ap
p
r
o
ac
h
e
s
,
n
am
ely
,
r
an
d
o
m
f
o
r
est,
s
u
p
p
o
r
t
v
ec
to
r
r
e
g
r
ess
io
n
an
d
m
u
lti
-
lay
er
p
e
r
ce
p
tr
o
n
.
T
h
e
m
eth
o
d
’
s
p
er
f
o
r
m
an
ce
is
ev
alu
ated
th
r
o
u
g
h
ex
p
er
im
en
ts
o
n
a
d
ataset
co
n
tain
in
g
p
u
b
lis
h
ed
p
a
p
er
s
with
th
eir
citatio
n
s
at
Petr
a
Un
iv
er
s
ity
.
T
h
e
R
NN
ar
ch
itectu
r
e
was
ch
o
s
en
d
u
e
to
its
ab
ilit
y
to
m
o
d
el
a
n
d
ca
p
tu
r
e
s
eq
u
e
n
tial
d
ata,
m
ak
in
g
it
ap
p
r
o
p
r
iate
f
o
r
p
er
f
o
r
m
in
g
tim
e
s
er
ies
p
r
ed
ictio
n
task
s
s
u
ch
as
cita
t
io
n
co
u
n
t
p
r
ed
ictio
n
.
Mo
r
e
s
p
ec
if
ica
lly
,
in
th
is
r
esear
ch
p
r
o
ject,
th
e
en
co
d
er
-
d
ec
o
d
e
r
m
o
d
el
o
f
R
NNs
is
u
s
ed
,
wh
ich
u
s
es
an
R
NN
to
l
ea
r
n
a
c
o
m
p
r
ess
ed
r
ep
r
esen
tatio
n
o
f
t
h
e
in
p
u
t
s
eq
u
en
ce
an
d
b
e
ab
le
to
g
en
er
ate
an
o
u
t
p
u
t
s
eq
u
e
n
ce
g
iv
en
th
e
lea
r
n
ed
r
ep
r
esen
tatio
n
,
w
h
ich
is
v
er
y
s
u
itab
le
f
o
r
lear
n
in
g
th
e
citatio
n
d
ata,
wh
ich
is
b
ased
o
n
th
e
h
is
to
r
ical
citatio
n
n
u
m
b
er
s
,
will
allo
w
th
e
m
o
d
e
l
to
ca
p
tu
r
e
th
e
tim
e
d
e
p
en
d
e
n
cy
as
m
a
n
y
as
th
e
lo
n
g
-
ter
m
d
ep
en
d
en
cy
in
t
h
e
citatio
n
d
ata.
T
h
e
en
co
d
e
r
p
ar
t
an
d
th
e
d
ec
o
d
er
p
ar
t
o
f
th
e
R
NN
ar
e
co
n
s
tr
u
cted
b
y
th
e
lo
n
g
s
h
o
r
t
ter
m
m
em
o
r
y
(
L
STM
)
ce
ll
o
r
th
e
GR
U
ce
ll,
r
esp
ec
tiv
ely
.
T
h
e
r
ea
s
o
n
th
at
we
u
s
ed
th
e
L
STM
ce
ll
o
r
th
e
GR
U
ce
ll
b
u
t
n
o
t
an
y
o
th
er
ty
p
es
o
f
R
NN,
s
u
ch
as
th
e
Simp
le
R
NN
o
r
th
e
E
l
m
an
R
NN,
is
th
at
th
e
L
STM
c
ell
o
r
th
e
GR
U
ce
ll
is
ar
g
u
ed
to
b
e
a
v
er
y
ef
f
ec
tiv
e
m
eth
o
d
to
o
v
e
r
co
m
e
th
e
v
a
n
is
h
in
g
g
r
ad
ien
t p
r
o
b
lem
an
d
c
an
ca
p
tu
r
e
th
e
lo
n
g
-
ter
m
d
ep
e
n
d
en
c
y
th
at
m
ay
ex
is
t
in
th
e
i
n
itial
s
eq
u
en
ce
.
Sin
ce
th
e
in
f
lu
e
n
ce
o
f
th
e
p
ap
er
ca
n
n
o
t
b
e
im
m
ed
iately
av
ailab
le
in
t
h
e
p
r
o
b
l
em
s
ettin
g
o
f
o
u
r
citatio
n
co
u
n
t
p
r
ed
ictio
n
,
it
m
ay
o
c
cu
r
g
r
a
d
u
ally
.
Fo
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
In
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
3
6
,
No
.
2
,
No
v
em
b
er
20
24
:
1
0
7
0
-
1
0
8
2
1074
ex
am
p
le,
o
n
ly
in
t
h
e
f
ir
s
t
f
ew
y
ea
r
s
a
f
ew
p
eo
p
le
h
av
e
f
o
u
n
d
an
d
r
ea
d
t
h
at
p
a
p
er
,
an
d
t
h
en
in
th
e
y
ea
r
y
,
a
g
r
ea
t n
u
m
b
er
o
f
r
esear
ch
e
r
s
ar
e
wo
r
k
in
g
o
n
t
h
e
p
ap
e
r
s
h
ig
h
l
y
d
ep
e
n
d
in
g
o
n
t
h
is
ar
ticle.
I
n
o
u
r
r
esear
c
h
,
a
s
s
u
m
e
th
e
ta
r
g
et
p
ap
e
r
h
as
b
ee
n
cited
0
,
1
,
…
,
tim
es
s
in
ce
it
was
p
u
b
lis
h
ed
.
I
n
o
th
er
wo
r
d
s
,
s
h
o
ws
h
o
w
m
an
y
tim
es
th
e
p
ap
er
was
cited
in
th
e
ith
y
ea
r
af
ter
it
was
p
u
b
lis
h
ed
.
L
et'
s
s
ay
we
k
n
o
w
s0
,
s
1
,
…
,
s
k
an
d
wan
t
to
p
r
ed
ict
+
1
,
+
2
,
…
,
f
o
r
a
p
ap
er
(
)
.
T
h
e
ch
allen
g
e
is
to
p
r
ed
ict
h
o
w
m
an
y
tim
es
an
ar
ticle
will
b
e
cited
b
etwe
en
th
e
y
ea
r
it wa
s
p
u
b
lis
h
ed
an
d
th
e
y
ea
r
b
a
s
ed
o
n
h
o
w
m
an
y
tim
es
it
wa
s
cited
.
As
d
ef
in
ed
,
th
e
citatio
n
co
u
n
t
p
r
e
d
ictio
n
p
r
o
b
lem
'
s
in
p
u
ts
an
d
o
u
tp
u
ts
ar
e
b
o
th
s
eq
u
en
ce
s
o
f
v
al
u
es.
R
NNs
ar
e
g
o
o
d
a
t
lear
n
in
g
th
e
o
r
d
e
r
o
f
v
alu
e
s
in
th
ese
p
r
o
b
lem
s
,
s
o
we
u
s
ed
R
NNs
in
o
u
r
p
r
o
p
o
s
ed
m
eth
o
d
.
R
NNs
ca
n
l
ea
r
n
to
d
o
task
s
wh
e
r
e
th
e
in
p
u
ts
f
o
llo
w
a
n
at
u
r
al
o
r
d
er
.
T
h
e
in
p
u
t
s
eq
u
en
c
e
f
o
r
th
e
s
p
ec
if
ied
citatio
n
p
r
ed
icti
o
n
p
r
o
b
lem
is
(
0
,
…
,
)
h
as
+
1
v
alu
es,
a
n
d
th
e
o
u
tp
u
t
s
eq
u
en
ce
(
+
1
,
…
,
)
h
as
v
alu
es.
So
,
in
o
u
r
p
r
o
p
o
s
ed
m
eth
o
d
,
we
d
ev
is
ed
a
"m
an
y
-
to
-
m
a
n
y
"
R
NN
ar
ch
itectu
r
e
(
R
NNs
ar
e
d
iv
id
ed
in
to
f
o
u
r
g
r
o
u
p
s
b
ased
o
n
th
e
len
g
t
h
o
f
t
h
e
d
ata
th
ey
tak
e
in
a
n
d
g
iv
e
o
u
t:
1
-
to
-
o
n
e,
2
-
to
-
m
an
y
,
3
-
to
-
o
n
e,
a
n
d
4
-
to
-
m
a
n
y
)
.
T
ab
le
1
.
S
u
m
m
a
r
izes th
e
r
elat
ed
r
esear
ch
wo
r
k
R
e
f
e
r
e
n
c
e
M
e
t
h
o
d
D
a
t
a
s
e
t
K
e
y
f
i
n
d
i
n
g
s
M
a
n
s
o
u
r
e
t
a
l
.
[
1
6
]
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
ms
t
o
p
r
e
d
i
c
t
f
u
t
u
r
e
c
i
t
a
t
i
o
n
s
o
f
su
b
m
i
t
t
e
d
r
e
sea
r
c
h
p
a
p
e
r
s
I
n
t
e
r
n
a
t
i
o
n
a
l
A
r
a
b
Jo
u
r
n
a
l
o
f
I
n
f
o
r
ma
t
i
o
n
Te
c
h
n
o
l
o
g
y
d
a
t
a
se
t
s
p
a
n
n
i
n
g
a
d
e
c
a
d
e
Th
e
n
u
m
b
e
r
o
f
r
e
f
e
r
e
n
c
e
s
w
a
s
t
h
e
m
o
st
i
mp
o
r
t
a
n
t
f
e
a
t
u
r
e
,
w
h
i
l
e
t
h
e
n
u
m
b
e
r
o
f
a
u
t
h
o
r
s was
t
h
e
l
e
a
st
si
g
n
i
f
i
c
a
n
t
f
e
a
t
u
r
e
;
n
e
u
r
a
l
n
e
t
w
o
r
k
a
n
d
v
o
t
i
n
g
c
l
a
ssi
f
i
e
r
1
t
e
c
h
n
i
q
u
e
s
o
u
t
p
e
r
f
o
r
m
e
d
o
t
h
e
r
t
e
c
h
n
i
q
u
e
s.
A
b
r
i
s
h
a
m
i
a
n
d
A
l
i
a
k
b
a
r
y
[
1
7
]
A
r
t
i
f
i
c
i
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
t
o
p
r
e
d
i
c
t
t
h
e
l
o
n
g
-
t
e
r
m
c
i
t
a
t
i
o
n
c
o
u
n
t
o
f
a
p
a
p
e
r
b
a
s
e
d
o
n
i
t
s
c
i
t
a
t
i
o
n
s
d
u
r
i
n
g
t
h
e
i
n
i
t
i
a
l
y
e
a
r
s
a
f
t
e
r
p
u
b
l
i
c
a
t
i
o
n
N
/
A
Th
e
p
r
o
p
o
s
e
d
me
t
h
o
d
o
u
t
p
e
r
f
o
r
me
d
e
x
i
s
t
i
n
g
st
a
t
e
-
of
-
t
h
e
-
a
r
t
m
e
t
h
o
d
s
i
n
p
r
e
d
i
c
t
i
n
g
y
e
a
r
l
y
a
n
d
t
o
t
a
l
c
i
t
a
t
i
o
n
c
o
u
n
t
s
.
M
a
t
su
i
e
t
a
l
.
[
1
8
]
R
e
g
r
e
ssi
o
n
a
n
a
l
y
si
s
-
b
a
s
e
d
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
t
o
p
r
e
d
i
c
t
f
u
t
u
r
e
c
i
t
a
t
i
o
n
c
o
u
n
t
o
f
a
r
e
s
e
a
r
c
h
a
r
t
i
c
l
e
N
/
A
D
e
l
i
v
e
r
e
d
c
i
t
a
t
i
o
n
c
o
u
n
t
p
r
e
d
i
c
t
i
o
n
u
s
i
n
g
t
h
e
p
r
o
p
o
se
d
me
t
h
o
d
A
h
u
j
a
[
1
9
]
Tw
o
t
y
p
e
s
o
f
a
n
a
l
y
s
i
s
t
o
p
r
e
d
i
c
t
t
h
e
g
r
o
w
t
h
o
f
u
n
i
v
e
r
si
t
i
e
s
a
b
o
v
e
a
n
d
b
e
l
o
w
t
h
e
a
v
e
r
a
g
e
c
o
n
c
e
r
n
i
n
g
t
h
e
t
o
t
a
l
n
u
m
b
e
r
o
f
u
n
i
v
e
r
s
i
t
i
e
s
D
a
t
a
s
e
t
s
p
a
n
n
i
n
g
f
r
o
m
2
0
1
1
t
o
2
0
1
6
,
i
n
c
l
u
d
i
n
g
a
l
l
u
n
i
v
e
r
si
t
i
e
s'
p
u
b
l
i
c
a
t
i
o
n
a
n
d
c
i
t
a
t
i
o
n
d
e
t
a
i
l
s
P
r
e
d
i
c
t
e
d
a
b
o
v
e
-
a
v
e
r
a
g
e
g
r
o
w
t
h
i
n
u
n
i
v
e
r
si
t
y
p
u
b
l
i
c
a
t
i
o
n
s
a
n
d
c
i
t
a
t
i
o
n
s fo
r
t
h
e
y
e
a
r
s 2
0
1
7
an
d
o
n
w
a
r
d
s
S
u
[
2
0
]
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
m
o
d
e
l
s
a
n
d
a
n
e
u
r
a
l
n
e
t
w
o
r
k
t
o
p
r
e
d
i
c
t
f
u
t
u
r
e
c
i
t
a
t
i
o
n
c
o
u
n
t
s
b
a
se
d
o
n
e
i
g
h
t
b
i
b
l
i
o
met
r
i
c
f
e
a
t
u
r
e
s
o
f
c
i
t
i
n
g
p
a
p
e
r
s
2
6
0
0
p
a
p
e
r
s
o
f
p
h
y
s
i
o
l
o
g
y
e
x
t
r
a
c
t
e
d
f
r
o
m
t
h
e
W
e
b
o
f
S
c
i
e
n
c
e
S
e
l
e
c
t
e
d
f
e
a
t
u
r
e
s wer
e
v
a
l
u
a
b
l
e
i
n
p
r
e
d
i
c
t
i
n
g
l
o
n
g
-
t
e
r
m
c
i
t
a
t
i
o
n
c
o
u
n
t
s,
a
n
d
t
h
e
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
n
d
n
e
u
r
a
l
n
e
t
w
o
r
k
mo
d
e
l
s
h
e
l
p
e
d
p
r
e
d
i
c
t
f
u
t
u
r
e
c
i
t
a
t
i
o
n
c
o
u
n
t
s
D
u
[
2
1
]
S
e
v
e
r
a
l
mac
h
i
n
e
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s
t
o
r
a
n
k
r
e
se
a
r
c
h
i
n
s
t
i
t
u
t
i
o
n
s
b
a
se
d
o
n
p
r
e
d
i
c
t
i
n
g
t
h
e
n
u
m
b
e
r
o
f
a
c
c
e
p
t
e
d
p
a
p
e
r
s a
t
u
p
c
o
mi
n
g
t
o
p
c
o
n
f
e
r
e
n
c
e
s
N
/
A
P
r
o
p
o
se
d
a
t
h
r
e
e
-
p
h
a
s
e
e
x
p
e
r
i
me
n
t
,
st
a
r
t
i
n
g
w
i
t
h
a
si
m
p
l
e
a
v
e
r
a
g
e
m
e
t
h
o
d
a
n
d
i
m
p
r
o
v
i
n
g
p
r
e
d
i
c
t
i
o
n
s wi
t
h
l
i
n
e
a
r
r
e
g
r
e
ssi
o
n
,
r
a
n
k
S
V
M
,
a
n
d
e
n
s
e
mb
l
e
mo
d
e
l
s
W
e
n
e
t
a
l
.
[
2
2
]
R
N
N
me
t
h
o
d
w
i
t
h
t
h
e
G
R
U
-
C
P
M
t
o
p
r
e
d
i
c
t
c
i
t
a
t
i
o
n
n
u
mb
e
r
s
R
e
a
l
d
a
t
a
s
e
t
s
G
R
U
-
C
P
M
h
a
s
h
i
g
h
e
r
p
r
e
d
i
c
t
i
o
n
a
c
c
u
r
a
c
y
a
n
d
f
a
s
t
e
r
c
o
n
v
e
r
g
e
n
c
e
s
p
e
e
d
a
n
d
o
u
t
p
e
r
f
o
r
me
d
e
x
i
s
t
i
n
g
me
t
h
o
d
s
i
n
t
h
e
t
i
m
e
seri
e
s
p
r
e
d
i
c
t
i
o
n
o
f
c
i
t
a
t
i
o
n
c
o
u
n
t
C
r
o
f
t
a
n
d
S
a
c
k
[
2
3
]
N
e
u
r
a
l
n
e
t
w
o
r
k
m
o
d
e
l
s
t
o
p
r
e
d
i
c
t
f
u
t
u
r
e
p
e
r
f
o
r
m
a
n
c
e
o
f
j
o
u
r
n
a
l
s
H
i
st
o
r
i
c
a
l
b
i
b
l
i
o
me
t
r
i
c
d
a
t
a
f
o
r
j
o
u
r
n
a
l
s
i
n
d
e
x
e
d
i
n
S
c
o
p
u
s
Th
e
p
r
o
p
o
s
e
d
mo
d
e
l
s f
o
r
p
r
e
d
i
c
t
i
n
g
f
u
t
u
r
e
c
i
t
a
t
i
o
n
s
a
n
d
C
i
t
e
S
c
o
r
e
v
a
l
u
e
s
o
u
t
p
e
r
f
o
r
me
d
h
e
u
r
i
s
t
i
c
p
r
e
d
i
c
t
i
o
n
b
a
se
l
i
n
e
s
a
n
d
c
l
a
s
si
c
mac
h
i
n
e
-
le
a
r
n
i
n
g
m
o
d
e
l
s
R
u
a
n
e
t
a
l
.
[
2
4
]
F
o
u
r
-
l
a
y
e
r
BP
n
e
u
r
a
l
n
e
t
w
o
r
k
mo
d
e
l
t
o
p
r
e
d
i
c
t
f
i
v
e
-
y
e
a
r
c
i
t
a
t
i
o
n
s
o
f
p
a
p
e
r
s
P
a
p
e
r
s
i
n
t
h
e
l
i
b
r
a
r
y
,
i
n
f
o
r
mat
i
o
n
,
a
n
d
d
o
c
u
me
n
t
a
t
i
o
n
f
i
e
l
d
i
n
d
e
x
e
d
b
y
t
h
e
C
S
S
C
I
d
a
t
a
b
a
se
f
r
o
m
2
0
0
0
t
o
2
0
1
3
Th
e
B
a
c
k
p
r
o
p
a
g
a
t
i
o
n
(
B
P
)
n
e
u
r
a
l
n
e
t
w
o
r
k
mo
d
e
l
o
u
t
p
e
r
f
o
r
me
d
s
i
x
o
t
h
e
r
b
a
s
e
l
i
n
e
mo
d
e
l
s
,
s
h
o
w
i
n
g
a
h
i
g
h
e
r
a
c
c
u
r
a
c
y
i
n
p
r
e
d
i
c
t
i
n
g
c
i
t
a
t
i
o
n
s f
o
r
l
e
ss fr
e
q
u
e
n
t
l
y
c
i
t
e
d
p
a
p
e
r
s
t
h
a
n
t
h
o
se
o
f
t
e
n
c
i
t
e
d
.
T
h
i
s
p
e
r
f
o
r
m
a
n
c
e
w
a
s
i
n
f
l
u
e
n
c
e
d
s
i
g
n
i
f
i
c
a
n
t
l
y
b
y
f
i
v
e
k
e
y
f
e
a
t
u
r
e
s,
w
h
i
c
h
p
l
a
y
e
d
a
s
i
g
n
i
f
i
c
a
n
t
r
o
l
e
i
n
t
h
e
m
o
d
e
l
's
p
r
e
d
i
c
t
i
v
e
su
c
c
e
ss.
Alg
o
r
ith
m
1
s
h
o
ws
th
e
p
r
o
p
o
s
ed
m
eth
o
d
f
o
r
f
o
r
ec
asti
n
g
f
u
t
u
r
e
citatio
n
co
u
n
ts
u
s
in
g
a
R
NN
m
o
d
el.
T
h
e
alg
o
r
ith
m
tak
es
as
in
p
u
t
t
h
e
h
is
to
r
ical
citatio
n
c
o
u
n
ts
(
s
0
,
s
1
,
.
.
.
,
s
k
)
o
f
a
g
iv
e
n
p
a
p
er
,
th
e
y
ea
r
(
k
)
wh
en
th
e
citatio
n
co
u
n
ts
wer
e
last
o
b
s
er
v
ed
,
a
n
d
th
e
y
ea
r
(
n
)
u
p
to
wh
ich
th
e
citatio
n
c
o
u
n
ts
ar
e
r
eq
u
ir
e
d
to
b
e
f
o
r
ec
asted
.
T
h
e
o
u
tp
u
t
o
f
th
e
alg
o
r
ith
m
is
th
e
f
o
r
ec
asted
citatio
n
co
u
n
ts
(
s
(
k
+1
)
,
s
(
k
+2
)
,
.
.
.
,
s
n
)
f
o
r
th
e
y
ea
r
s
k
+1
to
n
.
T
h
e
alg
o
r
ith
m
c
o
n
s
i
s
ts
o
f
s
i
x
s
tep
s
:
d
ata
p
r
ep
r
o
ce
s
s
in
g
,
s
eq
u
en
ce
co
n
s
tr
u
ctio
n
,
m
o
d
el
c
o
n
s
tr
u
ctio
n
,
m
o
d
el
tr
ain
in
g
,
h
y
p
er
p
a
r
am
e
ter
tu
n
in
g
,
a
n
d
citatio
n
c
o
u
n
t
p
r
ed
ictio
n
.
I
n
t
h
e
d
ata
p
r
ep
r
o
ce
s
s
in
g
s
tep
,
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:
2
5
0
2
-
4
7
52
F
o
r
ec
a
s
tin
g
r
esea
r
ch
in
flu
en
ce
:
a
r
ec
u
r
r
en
t n
eu
r
a
l n
etw
o
r
k
…
(
N
a
s
er J
a
ma
l
)
1075
s
tatis
t
ical
n
o
r
m
aliza
tio
n
an
d
lo
g
-
m
o
d
u
lu
s
c
o
n
v
e
r
s
io
n
ar
e
ap
p
lied
to
th
e
r
aw
citati
o
n
c
o
u
n
ts
to
en
s
u
r
e
a
co
n
s
is
ten
t
an
d
s
tan
d
ar
d
ized
r
an
g
e
o
f
v
alu
es.
T
h
e
s
eq
u
en
ce
co
n
s
tr
u
ctio
n
s
tep
d
i
v
id
es
th
e
p
r
ep
r
o
ce
s
s
ed
d
ata
in
to
in
p
u
t a
n
d
o
u
tp
u
t seq
u
en
ce
s
.
T
h
e
m
o
d
el
-
b
u
ild
in
g
s
tep
im
p
lem
en
ts
a
s
eq
u
en
ce
-
to
-
s
eq
u
en
ce
r
ec
u
r
r
en
t
en
c
o
d
er
-
d
ec
o
d
er
n
eu
r
al
n
etw
o
r
k
a
r
ch
itectu
r
e
with
a
m
an
y
-
to
-
m
an
y
to
p
o
lo
g
y
.
T
h
e
en
co
d
e
r
p
r
o
ce
s
s
es
th
e
in
p
u
t
s
eq
u
en
ce
wh
ile
th
e
d
ec
o
d
er
g
en
e
r
ates
th
e
o
u
tp
u
t
s
eq
u
en
ce
.
I
n
th
e
m
o
d
el
tr
ain
i
n
g
s
tep
,
t
h
e
m
o
d
el
is
tr
ain
e
d
u
s
in
g
th
e
s
eq
2
s
eq
ap
p
r
o
ac
h
,
a
n
d
h
y
p
er
p
ar
a
m
eter
tu
n
in
g
is
p
e
r
f
o
r
m
ed
to
o
p
tim
ize
th
e
m
o
d
el'
s
p
er
f
o
r
m
an
ce
.
Fin
ally
,
th
e
tr
ai
n
ed
R
NN
m
o
d
el
is
u
s
ed
to
p
r
ed
ict
th
e
citatio
n
co
u
n
ts
f
o
r
k
+1
to
n
b
y
in
p
u
ttin
g
th
e
o
b
s
er
v
ed
c
itatio
n
co
u
n
ts
an
d
allo
win
g
th
e
m
o
d
el
to
g
e
n
er
at
e
th
e
ex
p
ec
ted
co
u
n
ts
.
Alg
o
r
ith
m
1
.
C
itatio
n
co
u
n
t
p
r
ed
ictio
n
u
s
in
g
R
NN
Inputs:
-
Historical citation counts: s0, s1, ..., sk
-
k: the year the citation counts were last observed
-
n: the year to predict citation counts
Output:
-
Predicted citation counts for years k+1 to n
: s(k+1), s(k+2), ..., sn
Steps
:
1. Data Preprocessing:
1.1. Apply statistical normalization to the raw citation counts.
1.2. Perform log
-
modulus conversion to re
-
scale the integer counts and reduce skewness.
2. Sequence Construction:
2.1. Split t
he preprocessed citation count data into input and output sequences.
2.2. Input sequence: (s0, s1, ..., sk)
2.3. Output sequence: (s(k+1), s(k+2), ..., sn)
3. Model Construction:
3.1. Implement a sequence
-
to
-
sequence recurrent encoder
-
decoder neur
al network
architecture.
3.2. Use a many
-
to
-
many network topology.
3.3. The encoder processes the input sequence of length k.
3.4. The decoder generates the output sequence of length n
-
k.
3.5. Utilize Long Short
-
Term Memory (LSTM) or Gated Recu
rrent Unit (GRU) cells in the
encoder and decoder components.
4. Model Training:
4.1. Train the Recurrent Neural Network (RNN) using the sequence
-
to
-
sequence (seq2seq)
approach.
4.2. The encoder is responsible for processing the input sequence.
4.3. The decoder is tasked with generating the output sequence.
5. Hyperparameter Tuning:
5.1. Perform hyperparameter tuning to select the optimal values for the RNN
hyperparameter
s.
5.2. Tune the number of LSTM cells, layers, activation function, learning rate, dropout
rate, batch size, and epochs.
6. Citation Count Prediction:
6.1. Input the observed citation counts (s0, s1, ..., sk) into the trained RNN model.
6.2. Let t
he RNN generate the predicted citation counts (s(k+1), s(k+2), ..., sn) for
years k+1 to n.
7. Return the predicted citation counts (s(k+1), s(k+2), ..., sn).
T
h
e
m
ain
e
q
u
atio
n
u
s
ed
i
n
th
is
ap
p
r
o
ac
h
is
th
e
s
eq
u
e
n
ce
-
to
-
s
eq
u
e
n
ce
m
o
d
el,
wh
i
ch
ca
n
b
e
r
ep
r
esen
ted
as:
1
,
2
,
…
,
−
=
(
0
,
1
,
…
,
)
(
1
)
W
h
er
e
0
,
1
,
…
,
ar
e
th
e
o
b
s
er
v
ed
citatio
n
co
u
n
ts
,
an
d
1
,
2
,
…
,
−
ar
e
th
e
p
r
ed
icted
citatio
n
co
u
n
ts
.
T
h
e
R
NN
is
tr
ain
ed
u
s
in
g
th
e
in
p
u
t
s
eq
u
en
ce
(
0
,
1
,
…
,
)
an
d
th
e
o
u
tp
u
t
s
eq
u
en
ce
(
+
1
,
+
2
,
…
,
)
u
s
in
g
th
e
s
eq
u
en
ce
-
to
-
s
eq
u
e
n
c
e
m
o
d
el
tech
n
i
q
u
e.
T
h
e
o
u
t
p
u
t
o
f
th
e
R
NN
is
th
e
p
r
ed
icted
s
eq
u
en
ce
o
f
citatio
n
co
u
n
ts
(
1
,
2
,
…
,
−
)
.
3
.
1
.
I
m
ple
m
ent
a
t
io
n
d
et
a
ils
T
o
m
ak
e
th
e
r
esear
c
h
f
in
d
in
g
s
r
ep
licab
le,
we
b
r
ief
ly
p
r
o
v
i
d
e
th
e
s
p
ec
if
ic
im
p
lem
en
tatio
n
s
o
f
o
u
r
s
tu
d
ies
in
th
is
s
ec
tio
n
.
W
e
u
s
ed
th
e
"Ke
r
as"
f
r
am
ewo
r
k
[
2
6
]
,
a
well
-
k
n
o
wn
an
d
p
o
p
u
lar
d
ee
p
lear
n
i
n
g
an
d
ar
tific
ial
n
eu
r
al
n
etwo
r
k
im
p
l
em
en
tatio
n
(
h
ttp
s
://k
er
as
.
io
)
.
W
ith
a
lear
n
in
g
r
ate
o
f
0
.
0
0
0
1
,
we
ap
p
lied
th
e
R
MSPr
o
p
o
p
tim
izatio
n
alg
o
r
i
th
m
[
2
7
]
,
a
v
alu
ab
le
tech
n
i
q
u
e
f
o
r
tr
ain
in
g
n
eu
r
al
n
etwo
r
k
s
.
I
t
is
im
p
o
r
tan
t
to
n
o
te
th
at
th
ese
ar
e
ju
s
t
s
o
m
e
p
o
s
s
ib
le
v
alu
es
an
d
th
at
th
e
o
p
tim
al
v
alu
es
f
o
r
th
ese
p
ar
am
eter
s
w
ill
d
ep
en
d
o
n
th
e
s
p
ec
if
ic
ch
a
r
ac
ter
is
tics
o
f
th
e
tr
ain
in
g
d
ataset
an
d
th
e
r
e
s
ea
r
ch
q
u
esti
o
n
b
ein
g
ad
d
r
ess
ed
.
Ad
d
itio
n
ally
,
it
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
In
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
3
6
,
No
.
2
,
No
v
em
b
er
20
24
:
1
0
7
0
-
1
0
8
2
1076
is
wo
r
th
n
o
tin
g
th
at
th
e
c
h
o
ic
e
o
f
h
y
p
er
p
ar
am
eter
s
ca
n
s
ig
n
if
ican
tly
im
p
ac
t
th
e
p
er
f
o
r
m
a
n
ce
o
f
t
h
e
R
NN
an
d
th
at
s
elec
tin
g
o
p
tim
al
v
alu
es c
an
r
eq
u
i
r
e
s
ig
n
if
ican
t e
x
p
er
im
en
tatio
n
an
d
tu
n
in
g
.
Selectin
g
o
p
tim
al
m
eta
-
p
ar
am
eter
s
co
n
s
titu
tes
a
p
iv
o
tal
co
n
s
titu
en
t
in
co
n
s
tr
u
ctin
g
r
o
b
u
s
t
p
r
ed
ictiv
e
m
o
d
els
b
ased
o
n
R
NNs
.
I
t
is
im
p
er
ativ
e
to
s
y
s
tem
atica
lly
e
x
p
lo
r
e
th
e
h
y
p
e
r
p
ar
a
m
eter
s
p
a
ce
an
d
id
en
tif
y
th
e
id
ea
l
co
n
f
ig
u
r
atio
n
s
th
r
o
u
g
h
r
ig
o
r
o
u
s
ex
p
er
im
e
n
tatio
n
-
d
r
iv
en
v
alid
atio
n
h
ar
n
ess
in
g
th
e
v
alid
atio
n
d
ataset.
T
h
e
cu
lm
in
ate
d
v
alu
es
r
ep
r
esen
t
th
e
b
est
-
p
er
f
o
r
m
in
g
ar
c
h
itectu
r
al
s
ettin
g
s
tailo
r
ed
t
o
th
e
p
r
o
b
lem
c
o
n
tex
t
th
at
p
o
ten
tiate
m
ax
im
izin
g
g
e
n
er
aliza
b
le
p
r
ed
ictiv
e
p
er
f
o
r
m
an
ce
o
n
u
n
s
ee
n
d
ata.
T
ab
le
2
s
h
o
ws
p
ar
a
m
eter
v
alu
es tested
f
o
r
o
p
tim
izin
g
th
e
R
NN
m
o
d
el
f
o
r
citatio
n
p
r
ed
ictio
n
.
T
ab
le
2
.
Op
tim
ized
p
ar
am
eter
v
alu
es f
o
r
th
e
R
NN
m
o
d
el
P
a
r
a
me
t
e
r
V
a
l
u
e
O
u
t
p
u
t
l
a
y
e
r
D
e
n
se
l
a
y
e
r
A
c
t
i
v
a
t
i
o
n
f
u
n
c
t
i
o
n
R
e
LU
Ep
o
c
h
s
2
0
0
O
p
t
i
mi
z
a
t
i
o
n
a
l
g
o
r
i
t
h
m
A
d
a
m
Le
a
r
n
i
n
g
r
a
t
e
1
0
−
3
1
0
−
3
B
a
t
c
h
s
i
z
e
1
2
8
3
.
2
.
Da
t
a
s
et
T
o
co
n
s
titu
te
an
ap
p
r
o
p
r
iate
co
r
p
u
s
f
o
r
tr
ain
in
g
a
n
d
test
in
g
th
e
p
r
o
p
o
s
ed
p
r
e
d
ictiv
e
m
e
th
o
d
o
lo
g
y
,
p
u
b
licatio
n
r
ec
o
r
d
s
an
d
citatio
n
co
u
n
ts
wer
e
s
y
s
tem
atica
lly
co
llated
f
r
o
m
Go
o
g
le
Sch
o
lar
,
en
co
m
p
ass
in
g
th
e
s
ch
o
lar
ly
o
u
tp
u
t
f
r
o
m
Petr
a
Un
iv
er
s
ity
b
etwe
en
2
0
1
5
an
d
2
0
2
2
.
Stra
tifie
d
r
an
d
o
m
p
ar
titi
o
n
in
g
was
u
n
d
er
tak
e
n
with
d
o
cu
m
e
n
ts
p
u
b
lis
h
ed
d
u
r
in
g
2
0
1
5
-
2
0
2
0
ass
ig
n
ed
f
o
r
m
o
d
el
d
ev
el
o
p
m
en
t,
wh
ile
ar
ticles
f
r
o
m
2
0
2
1
-
20
2
2
wer
e
h
eld
o
u
t
f
o
r
u
n
f
o
r
eseen
ev
alu
atio
n
.
T
h
is
tem
p
o
r
al
d
ata
s
p
litt
in
g
ap
p
r
o
ac
h
en
s
u
r
es
m
o
d
el
g
en
e
r
aliza
tio
n
b
y
p
r
e
clu
d
in
g
o
v
er
f
itti
n
g
o
n
tem
p
o
r
ally
co
r
r
elate
d
tr
ain
in
g
in
s
tan
ce
s
th
at
co
u
ld
p
o
s
itiv
ely
s
k
ew
p
er
f
o
r
m
an
ce
m
etr
ics.
B
ased
o
n
liter
atu
r
e
an
d
d
o
m
ain
ex
p
er
tis
e,
d
iv
er
s
e
ex
p
lan
ato
r
y
f
ea
tu
r
es
wer
e
ju
d
icio
u
s
ly
id
en
tifie
d
to
ch
ar
ac
ter
ize
th
e
m
u
ltid
im
e
n
s
io
n
al
attr
ib
u
tes
h
y
p
o
th
esized
t
o
in
f
lu
en
ce
f
u
tu
r
e
citatio
n
co
u
n
ts
.
Sp
ec
if
ically
,
th
e
f
ea
tu
r
e
s
p
ac
e
s
p
an
n
ed
:
i)
in
tr
in
s
ic
p
r
o
p
er
ties
o
f
th
e
ar
ticle
en
co
m
p
ass
in
g
au
th
o
r
co
u
n
t,
titl
e,
an
d
ab
s
tr
ac
t
len
g
th
s
;
ii)
v
en
u
e
-
s
p
ec
if
ic
m
etr
ics
in
clu
d
in
g
jo
u
r
n
al
im
p
a
ct
f
ac
to
r
,
Hir
s
ch
in
d
ex
q
u
a
n
tify
in
g
jo
u
r
n
al
lev
el
p
r
o
d
u
ctiv
ity
an
d
citatio
n
im
p
ac
t
;
iii)
co
llab
o
r
atio
n
s
tatu
s
;
i
v
)
f
u
n
d
in
g
in
f
o
r
m
atio
n
;
v
)
b
i
b
lio
g
r
ap
h
ic
f
ea
tu
r
es
s
u
ch
as
r
ef
er
e
n
ce
s
cited
;
v
i)
p
u
b
lis
h
er
r
ep
u
tatio
n
;
an
d
v
i
i)
tem
p
o
r
al
ag
e.
C
o
llectiv
ely
,
th
ese
d
escr
ip
tiv
e
f
ac
to
r
s
en
ca
p
s
u
late
th
e
s
ch
o
lar
ly
im
p
ac
t,
q
u
ality
,
ex
p
o
s
u
r
e,
an
d
tem
p
o
r
al
m
atu
r
ity
th
at
ca
n
in
f
o
r
m
p
r
ed
ictiv
e
m
o
d
els
to
f
o
r
etell
f
u
t
u
r
e
cita
tio
n
s
m
o
r
e
ac
cu
r
ately
.
T
h
e
d
ev
elo
p
ed
p
r
e
d
ictiv
e
s
o
lu
tio
n
ca
n
ass
is
t
cr
itical
s
tak
eh
o
ld
er
s
,
in
clu
d
in
g
ac
ad
e
m
ics,
in
s
titu
tio
n
al
d
ec
is
io
n
m
ar
k
er
s
,
an
d
s
cien
ce
p
o
licy
ag
en
cies,
in
in
f
o
r
m
e
d
r
esear
ch
ev
alu
atio
n
an
d
s
tr
a
te
g
ically
allo
ca
tin
g
s
ch
o
lar
ly
r
eso
u
r
ce
s
to
m
a
x
im
ize
s
cien
tific
p
r
o
g
r
ess
.
W
e
ca
r
ef
u
lly
s
elec
ted
f
ea
tu
r
es
b
ased
o
n
liter
atu
r
e
an
d
d
o
m
ain
e
x
p
er
tis
e
to
ch
a
r
a
cter
ize
th
e
h
y
p
o
th
esized
m
u
ltid
im
en
s
io
n
al
attr
ib
u
tes
in
f
lu
en
ci
n
g
f
u
tu
r
e
citatio
n
co
u
n
ts
.
T
ab
le
3
p
r
e
s
e
n
ts
th
e
f
ea
tu
r
e
s
et
u
s
ed
in
o
u
r
s
tu
d
y
a
n
d
b
r
ief
ly
d
escr
ib
es
ea
ch
f
ea
tu
r
e.
T
h
es
e
f
ea
tu
r
es
en
c
o
m
p
ass
in
tr
in
s
i
c
p
r
o
p
er
ties
o
f
t
h
e
ar
ticle
(
e.
g
.
,
a
u
th
o
r
c
o
u
n
t,
titl
e
len
g
th
)
,
v
e
n
u
e
-
s
p
ec
if
ic
m
etr
ics
(
e.
g
.
,
jo
u
r
n
al
im
p
ac
t
f
ac
to
r
,
h
-
in
d
ex
)
,
co
llab
o
r
atio
n
s
tatu
s
,
f
u
n
d
in
g
in
f
o
r
m
atio
n
,
b
ib
lio
g
r
ap
h
ic
f
ea
t
u
r
es
(
e.
g
.
,
r
ef
er
en
c
es
cited
)
,
p
u
b
lis
h
er
r
ep
u
tatio
n
,
an
d
tem
p
o
r
al
ag
e.
B
y
in
c
o
r
p
o
r
atin
g
th
ese
d
iv
e
r
s
e
f
ac
to
r
s
,
we
aim
to
ca
p
tu
r
e
th
e
s
ch
o
lar
ly
im
p
ac
t,
q
u
ality
,
ex
p
o
s
u
r
e,
a
n
d
tem
p
o
r
al
m
at
u
r
i
ty
th
at
ca
n
in
f
o
r
m
p
r
ed
ictiv
e
m
o
d
els to
f
o
r
ec
ast f
u
tu
r
e
citati
o
n
s
ac
cu
r
ately
.
4.
E
XP
E
R
I
M
E
N
T
AND
R
E
SU
L
T
S
T
ab
le
4
d
elin
ea
tes
th
e
s
et
o
f
q
u
an
titativ
e
p
er
f
o
r
m
an
ce
m
etr
i
cs
u
tili
ze
d
f
o
r
v
alid
atin
g
th
e
ef
f
icac
y
o
f
o
u
r
p
r
o
p
o
s
ed
p
r
e
d
ictiv
e
m
eth
o
d
o
lo
g
y
.
Fo
r
r
eg
r
ess
io
n
ass
es
s
m
en
t,
m
ea
n
ab
s
o
lu
te
er
r
o
r
(
MA
E
)
,
m
ea
n
s
q
u
ar
e
d
er
r
o
r
(
MSE
)
,
a
n
d
r
o
o
t
m
ea
n
s
q
u
ar
ed
e
r
r
o
r
(
R
MSE
)
ar
e
e
m
p
lo
y
ed
to
q
u
a
n
tify
t
h
e
d
e
v
i
atio
n
s
b
etwe
en
t
h
e
f
o
r
ec
asted
an
d
g
r
o
u
n
d
tr
u
th
ci
tatio
n
co
u
n
ts
.
L
o
wer
er
r
o
r
s
s
i
g
n
if
y
en
h
a
n
ce
d
p
r
ed
ictiv
e
ac
cu
r
ac
y
.
Ad
d
itio
n
ally
,
class
if
icatio
n
m
etr
i
cs,
in
clu
d
in
g
ar
ea
u
n
d
e
r
th
e
r
ec
eiv
er
o
p
er
atin
g
ch
ar
ac
te
r
is
tic
(
R
OC
)
cu
r
v
e
(
AUC),
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
a
n
d
F1
s
co
r
e,
ev
alu
ate
th
e
m
o
d
el’
s
ab
ilit
y
to
d
is
cr
im
in
ate
b
etwe
en
h
ig
h
l
y
an
d
p
o
o
r
ly
cited
p
u
b
licatio
n
s
.
Sp
e
cif
ically
,
AUC
ev
alu
ates
th
e
class
if
i
er
’
s
o
v
er
all
ab
ilit
y
to
c
ateg
o
r
ize
co
r
r
ec
tly
ac
r
o
s
s
d
if
f
er
en
t
th
r
esh
o
ld
lev
e
ls
.
Acc
u
r
ac
y
co
m
p
u
tes
th
e
r
atio
o
f
ac
cu
r
ate
p
r
e
d
ictio
n
s
to
th
e
to
tal
p
o
p
u
latio
n
.
Pre
cisi
o
n
r
ep
r
esen
ts
th
e
p
r
o
p
o
r
tio
n
o
f
c
o
r
r
ec
tly
p
r
ed
icted
h
ig
h
ly
cited
ar
ticles
to
all
th
o
s
e
f
o
r
ec
asted
to
b
e
h
ig
h
ly
cited
.
R
ec
all
q
u
an
tifie
s
th
e
f
r
ac
tio
n
o
f
h
ig
h
ly
cited
p
ap
er
s
co
r
r
ec
tly
id
e
n
tifie
d
b
y
t
h
e
m
o
d
el
f
r
o
m
all
h
ig
h
ly
cited
ar
ticles
in
th
e
d
ataset.
T
h
e
F1
-
s
co
r
e
c
o
n
s
titu
tes
th
e
h
ar
m
o
n
ic
av
e
r
ag
e
b
et
wee
n
p
r
ec
is
io
n
a
n
d
r
ec
all,
im
p
ar
tin
g
eq
u
al
wei
g
h
t
ag
e
to
b
o
th
m
etr
ics.
C
o
llectiv
ely
,
th
ese
m
etr
ics
f
ac
ilit
ate
a
h
o
lis
tic
ass
e
s
s
m
en
t
o
f
th
e
g
e
n
er
aliza
b
ilit
y
,
r
o
b
u
s
tn
ess
,
an
d
r
ea
l
-
wo
r
ld
v
iab
ilit
y
o
f
th
e
d
ev
elo
p
ed
p
r
ed
ictiv
e
s
o
lu
t
io
n
.
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:
2
5
0
2
-
4
7
52
F
o
r
ec
a
s
tin
g
r
esea
r
ch
in
flu
en
ce
:
a
r
ec
u
r
r
en
t n
eu
r
a
l n
etw
o
r
k
…
(
N
a
s
er J
a
ma
l
)
1077
T
ab
le
3
.
Featu
r
es set
F
e
a
t
u
r
e
s
D
e
scri
p
t
i
o
n
P
u
b
l
i
c
a
t
i
o
n
y
e
a
r
Th
e
y
e
a
r
i
n
w
h
i
c
h
t
h
e
p
a
p
e
r
w
a
s
p
u
b
l
i
sh
e
d
N
u
mb
e
r
o
f
a
u
t
h
o
r
s
To
t
a
l
n
u
m
b
e
r
o
f
a
u
t
h
o
r
s
o
f
t
h
e
p
u
b
l
i
c
a
t
i
o
n
C
o
l
l
a
b
o
r
a
t
i
o
n
W
h
e
t
h
e
r
t
h
e
p
a
p
e
r
i
s a
r
e
s
u
l
t
o
f
a
c
o
l
l
a
b
o
r
a
t
i
o
n
o
r
n
o
t
Jo
u
r
n
a
l
i
mp
a
c
t
f
a
c
t
o
r
A
mea
s
u
r
e
o
f
t
h
e
a
v
e
r
a
g
e
n
u
mb
e
r
o
f
c
i
t
a
t
i
o
n
s
t
o
a
r
t
i
c
l
e
s
p
u
b
l
i
s
h
e
d
i
n
a
j
o
u
r
n
a
l
Jo
u
r
n
a
l
c
i
t
e
-
sc
o
r
e
A
mea
s
u
r
e
o
f
t
h
e
a
v
e
r
a
g
e
c
i
t
a
t
i
o
n
s
r
e
c
e
i
v
e
d
p
e
r
d
o
c
u
me
n
t
p
u
b
l
i
sh
e
d
i
n
a
j
o
u
r
n
a
l
Jo
u
r
n
a
l
h
-
i
n
d
e
x
A
mea
s
u
r
e
o
f
t
h
e
p
r
o
d
u
c
t
i
v
i
t
y
a
n
d
c
i
t
a
t
i
o
n
i
mp
a
c
t
o
f
t
h
e
p
u
b
l
i
c
a
t
i
o
n
s
i
n
a
j
o
u
r
n
a
l
S
c
o
p
u
s
q
u
a
r
t
i
l
e
S
c
o
p
u
s
q
u
a
r
t
i
l
e
o
f
t
h
e
j
o
u
r
n
a
l
i
n
w
h
i
c
h
t
h
e
p
a
p
e
r
w
a
s
p
u
b
l
i
sh
e
d
Ti
t
l
e
l
e
n
g
t
h
N
u
mb
e
r
o
f
w
o
r
d
s i
n
t
h
e
p
u
b
l
i
c
a
t
i
o
n
t
i
t
l
e
A
b
st
r
a
c
t
l
e
n
g
t
h
N
u
mb
e
r
o
f
w
o
r
d
s i
n
t
h
e
p
u
b
l
i
c
a
t
i
o
n
a
b
st
r
a
c
t
N
u
mb
e
r
o
f
p
a
g
e
s
To
t
a
l
n
u
m
b
e
r
o
f
p
a
g
e
s
o
f
t
h
e
p
u
b
l
i
c
a
t
i
o
n
K
e
y
w
o
r
d
c
o
u
n
t
To
t
a
l
n
u
m
b
e
r
o
f
k
e
y
w
o
r
d
s
i
n
t
h
e
p
u
b
l
i
c
a
t
i
o
n
C
i
t
a
t
i
o
n
s
i
n
t
h
e
f
i
r
s
t
y
e
a
r
Th
e
n
u
m
b
e
r
o
f
c
i
t
a
t
i
o
n
s
t
h
e
p
a
p
e
r
r
e
c
e
i
v
e
d
i
n
i
t
s
f
i
r
st
y
e
a
r
o
f
p
u
b
l
i
c
a
t
i
o
n
P
u
b
l
i
c
a
t
i
o
n
t
y
p
e
W
h
e
t
h
e
r
t
h
e
p
a
p
e
r
i
s a
c
o
n
f
e
r
e
n
c
e
p
r
o
c
e
e
d
i
n
g
,
j
o
u
r
n
a
l
a
r
t
i
c
l
e
,
e
t
c
.
F
u
n
d
i
n
g
W
h
e
t
h
e
r
t
h
e
p
a
p
e
r
r
e
c
e
i
v
e
d
f
u
n
d
i
n
g
o
r
n
o
t
R
e
f
e
r
e
n
c
e
s
N
u
mb
e
r
o
f
r
e
f
e
r
e
n
c
e
s
c
i
t
e
d
i
n
t
h
e
p
a
p
e
r
P
u
b
l
i
s
h
e
r
Th
e
p
u
b
l
i
sh
e
r
o
f
t
h
e
j
o
u
r
n
a
l
F
u
t
u
r
e
c
i
t
a
t
i
o
n
s
Ex
p
e
c
t
e
d
c
i
t
a
t
i
o
n
T
ab
le
4
.
E
v
alu
atio
n
m
et
r
ics
M
e
a
su
r
e
D
e
scri
p
t
i
o
n
Eq
u
a
t
i
o
n
M
A
E
M
e
a
su
r
e
s t
h
e
a
v
e
r
a
g
e
ma
g
n
i
t
u
d
e
o
f
e
r
r
o
r
s i
n
t
h
e
p
r
e
d
i
c
t
i
o
n
s
=
1
∑
=
1
M
S
E
M
e
a
su
r
e
s t
h
e
a
v
e
r
a
g
e
o
f
t
h
e
sq
u
a
r
e
d
d
i
f
f
e
r
e
n
c
e
s
b
e
t
w
e
e
n
t
h
e
p
r
e
d
i
c
t
e
d
a
n
d
a
c
t
u
a
l
v
a
l
u
e
s
=
1
∑
(
−
̂
)
2
=
1
R
M
S
E
M
e
a
su
r
e
s t
h
e
sq
u
a
r
e
r
o
o
t
o
f
t
h
e
a
v
e
r
a
g
e
o
f
t
h
e
sq
u
a
r
e
d
d
i
f
f
e
r
e
n
c
e
s
b
e
t
w
e
e
n
t
h
e
p
r
e
d
i
c
t
e
d
a
n
d
a
c
t
u
a
l
v
a
l
u
e
s
=
√
1
∑
(
−
̂
)
2
=
1
AUC
-
R
O
C
M
e
a
su
r
e
s t
h
e
a
b
i
l
i
t
y
o
f
t
h
e
mo
d
e
l
t
o
d
i
st
i
n
g
u
i
s
h
b
e
t
w
e
e
n
p
o
si
t
i
v
e
a
n
d
n
e
g
a
t
i
v
e
c
l
a
sses
A
c
c
u
r
a
c
y
M
e
a
su
r
e
s t
h
e
p
r
o
p
o
r
t
i
o
n
o
f
c
o
r
r
e
c
t
l
y
c
l
a
ss
i
f
i
e
d
i
n
s
t
a
n
c
e
s
=
N
um
b
e
r
o
f
C
o
r
r
e
c
t
P
r
e
di
c
t
i
o
n
s
T
o
t
a
l N
u
m
b
e
r
o
f
Pr
e
dic
t
i
o
n
s
P
r
e
c
i
s
i
o
n
Th
e
p
r
o
p
o
r
t
i
o
n
o
f
t
r
u
e
p
o
s
i
t
i
v
e
s
a
mo
n
g
a
l
l
p
r
e
d
i
c
t
e
d
p
o
si
t
i
v
e
s
=
+
R
e
c
a
l
l
Th
e
p
r
o
p
o
r
t
i
o
n
o
f
t
r
u
e
p
o
s
i
t
i
v
e
s
a
mo
n
g
a
l
l
a
c
t
u
a
l
p
o
s
i
t
i
v
e
s
=
+
F1
-
sc
o
r
e
Th
e
h
a
r
mo
n
i
c
m
e
a
n
o
f
p
r
e
c
i
si
o
n
a
n
d
R
e
c
a
l
l
1
=
2
⋅
⋅
+
T
ab
le
5
g
iv
es
a
co
m
p
ar
ativ
e
p
r
ed
ictiv
e
p
er
f
o
r
m
an
ce
b
e
twee
n
th
e
ex
am
in
ed
m
o
d
els
o
n
k
ey
ev
alu
atio
n
m
etr
ics.
T
h
e
p
r
o
p
o
s
ed
R
NN
ap
p
r
o
ac
h
ac
h
iev
es
a
s
u
p
er
io
r
MA
E
o
f
1
.
5
(
±
0
.
1
)
co
m
p
ar
ed
to
th
e
m
u
lti
-
lay
er
p
er
ce
p
tr
o
n
,
r
an
d
o
m
f
o
r
est,
an
d
SVM
m
o
d
els.
Ad
d
itio
n
ally
,
t
h
e
R
NN
m
eth
o
d
o
lo
g
y
attain
s
th
e
lo
west
MSE
o
f
5
.
0
(
±
0
.
8
)
,
c
o
n
tr
asted
b
y
h
i
g
h
er
er
r
o
r
s
f
r
o
m
th
e
o
th
er
tech
n
i
q
u
es.
R
eg
ar
d
in
g
th
e
AUC
m
etr
ic
th
at
ev
alu
ates
d
is
cr
im
in
ate
ab
ilit
y
,
th
e
R
NN
o
b
tain
s
0
.
9
2
(
±
0
.
0
1
)
,
s
u
r
p
ass
in
g
th
e
alter
n
a
tiv
es.
Fin
ally
,
o
u
r
in
tr
o
d
u
ce
d
R
NN
p
r
ed
icto
r
r
e
alize
s
th
e
h
ig
h
est
ac
cu
r
ac
y
o
f
0
.
8
6
(
±
0
.
0
2
)
f
o
r
co
r
r
ec
tly
e
s
tim
atin
g
citatio
n
s
,
ex
ce
ed
ed
o
n
ly
m
a
r
g
in
ally
b
y
t
h
e
ML
P.
T
ab
le
5
.
co
m
p
ar
is
o
n
o
f
e
x
p
er
i
m
en
tal
r
esu
lts
(
ex
p
er
im
en
t
1
,
8
0
:2
0
)
M
o
d
e
l
M
A
E
M
A
E
S
TD
M
S
E
M
S
E
S
TD
AUC
A
U
C
S
TD
A
c
c
u
r
a
c
y
A
c
c
u
r
a
c
y
S
TD
R
a
n
d
o
m
f
o
r
e
s
t
2
.
5
0
.
3
10
1
.
5
0
.
8
0
0
.
0
5
0
.
7
5
0
.
0
1
S
V
M
3
.
0
0
.
4
12
1
.
7
0
.
7
8
0
.
0
6
0
.
6
5
0
.
0
1
M
LP
2
.
1
0
.
2
8
.
5
1
.
0
0
.
8
8
0
.
0
2
0
.
8
0
0
.
0
3
P
r
o
p
o
se
d
R
N
N
1
.
5
0
.
1
5
.
0
0
.
8
0
.
9
2
0
.
0
1
0
.
8
6
0
.
0
2
As
s
h
o
wn
in
Fig
u
r
e
1
,
th
e
p
r
o
p
o
s
ed
d
ee
p
r
ec
u
r
r
en
t
n
etwo
r
k
ar
c
h
itectu
r
e
c
o
n
s
is
ten
tly
m
a
n
if
ests
th
e
o
p
tim
al
p
er
f
o
r
m
an
c
e
o
v
e
r
all
o
n
th
e
MA
E
,
MSE
,
AUC
-
R
OC
,
an
d
ac
cu
r
ac
y
m
etr
ics
with
tig
h
t
s
tan
d
ar
d
d
ev
iatio
n
s
.
T
h
e
r
o
b
u
s
t r
esu
lts
f
u
r
th
er
r
ein
f
o
r
ce
o
u
r
ap
p
r
o
ac
h
'
s
g
en
er
aliza
b
ilit
y
an
d
ef
f
icac
y
.
Ou
r
m
eth
o
d
o
lo
g
y
of
f
er
s
d
em
o
n
s
tr
ab
le
v
ia
b
ilit
y
f
o
r
r
ea
l
-
wo
r
l
d
ap
p
licatio
n
in
ac
cu
r
ately
f
o
r
ec
asti
n
g
citatio
n
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
52
In
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
3
6
,
No
.
2
,
No
v
em
b
er
20
24
:
1
0
7
0
-
1
0
8
2
1078
Fig
u
r
e
1
.
E
x
p
er
im
e
n
t 1
(
8
0
:2
0
s
p
lit)
m
o
d
el
p
er
f
o
r
m
a
n
ce
co
m
p
ar
is
o
n
T
ab
le
6
d
elin
ea
tes
th
e
p
r
ed
ic
tiv
e
p
er
f
o
r
m
an
ce
o
f
th
e
f
o
u
r
ex
am
in
ed
m
o
d
els
v
ia
1
0
-
f
o
ld
s
tr
atif
ied
cr
o
s
s
-
v
alid
atio
n
.
T
h
e
ass
ess
m
en
t
m
etr
ics
en
c
o
m
p
ass
R
MSE
,
ac
cu
r
ac
y
,
th
e
ar
ea
u
n
d
er
th
e
R
OC
cu
r
v
e
(
AUC),
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
s
co
r
e.
An
R
MSE
o
f
4
.
5
6
was
o
b
t
ain
ed
f
o
r
t
h
e
r
a
n
d
o
m
f
o
r
est
ap
p
r
o
ac
h
,
d
en
o
tin
g
s
u
b
s
tan
tial
d
ev
iatio
n
b
etwe
en
th
e
p
r
ed
icted
a
n
d
ac
tu
al
citati
o
n
co
u
n
ts
.
An
ac
cu
r
ac
y
o
f
0
.
7
8
im
p
lies
ac
cu
r
ate
class
if
icatio
n
b
y
th
e
r
an
d
o
m
f
o
r
est
m
o
d
el
f
o
r
7
8
%
o
f
th
e
p
u
b
licatio
n
s
.
T
h
e
AUC
o
f
0
.
8
4
in
d
icate
s
r
ea
s
o
n
ab
ly
g
o
o
d
d
is
cr
im
in
atio
n
b
etwe
en
h
ig
h
ly
an
d
p
o
o
r
ly
cited
ar
ticles.
Pre
cisi
o
n
,
r
ec
all,
an
d
F1
s
co
r
e
v
alu
es
o
f
0
.
7
5
,
0
.
8
0
,
a
n
d
0
.
7
8
,
r
esp
ec
tiv
ely
,
s
i
g
n
if
y
s
u
p
e
r
io
r
id
e
n
tific
atio
n
o
f
lo
w
-
im
p
ac
t
p
u
b
licatio
n
s
r
ela
tiv
e
to
h
ig
h
-
im
p
ac
t
o
n
es b
y
th
e
m
o
d
el.
T
ab
l
e
6
.
C
o
m
p
a
r
is
o
n
o
f
ex
p
e
r
im
en
tal
r
esu
lts
(
ex
p
er
im
en
t
2
,
c
r
o
s
s
-
v
alid
atio
n
)
M
o
d
e
l
R
M
S
E
A
c
c
u
r
a
c
y
AUC
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
-
sc
o
r
e
R
a
n
d
o
m
f
o
r
e
s
t
4
.
5
6
0
.
7
8
0
.
8
4
0
.
7
5
0
.
8
0
0
.
7
8
S
u
p
p
o
r
t
v
e
c
t
o
r
r
e
g
r
e
ss
i
o
n
3
.
9
4
0
.
8
1
0
.
8
8
0
.
8
1
0
.
8
3
0
.
8
2
M
u
l
t
i
-
l
a
y
e
r
p
e
r
c
e
p
t
r
o
n
3
.
7
2
0
.
8
4
0
.
9
1
0
.
8
3
0
.
8
5
0
.
8
4
P
r
o
p
o
se
d
a
p
p
r
o
a
c
h
(
R
N
N
)
1
.
8
4
0
.
9
1
0
.
9
6
0
.
9
1
0
.
9
3
0
.
9
2
T
h
e
s
u
p
p
o
r
t
v
ec
to
r
r
e
g
r
ess
io
n
m
o
d
el
d
em
o
n
s
tr
ates
e
n
h
an
ce
d
p
r
ed
ictiv
e
ca
p
ab
ilit
ies
o
v
er
r
an
d
o
m
f
o
r
est
with
lo
wer
ed
R
MSE
o
f
3
.
9
4
an
d
s
u
p
er
io
r
ac
cu
r
ac
y
o
f
0
.
8
1
,
AUC,
p
r
ec
is
io
n
,
r
ec
all
,
an
d
F1
s
co
r
es
o
f
0
.
8
8
,
0
.
8
1
,
0
.
8
3
,
an
d
0
.
8
2
,
r
es
p
ec
tiv
ely
.
Fu
r
th
er
im
p
r
o
v
em
e
n
ts
in
p
r
o
g
n
o
s
tic
p
er
f
o
r
m
an
ce
ar
e
ex
h
ib
ited
b
y
th
e
m
u
lti
-
lay
er
p
e
r
ce
p
tr
o
n
m
o
d
el,
attain
in
g
a
r
ed
u
ce
d
R
MSE
o
f
3
.
7
2
an
d
an
ac
cu
r
ac
y
o
f
0
.
8
4
.
T
h
e
h
ig
h
est
AUC
o
f
0
.
9
1
im
p
lies
its
s
tr
o
n
g
est
d
is
cr
im
in
atio
n
b
etwe
en
h
ig
h
an
d
lo
w
-
im
p
ac
t
ar
ticles,
co
m
p
lem
en
ted
b
y
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
o
f
0
.
8
3
,
0
.
8
5
,
an
d
0
.
8
4
.
As
s
h
o
wn
in
Fig
u
r
e
2
,
f
in
ally
,
th
e
p
r
o
p
o
s
ed
R
NN
ap
p
r
o
ac
h
s
u
r
p
ass
es
all
p
r
ec
ed
in
g
m
o
d
els
o
n
all
m
etr
ics
-
lo
west
R
M
SE
o
f
2
.
8
4
,
th
e
h
ig
h
est
ac
cu
r
ac
y
o
f
0
.
9
1
,
AUC
o
f
0
.
9
6
,
an
d
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
o
f
0
.
9
1
,
0
.
9
3
,
an
d
0
.
9
2
,
r
esp
ec
t
iv
ely
.
Ou
r
in
tr
o
d
u
ce
d
m
eth
o
d
o
lo
g
y
d
eliv
e
r
ed
th
e
m
o
s
t
a
cc
u
r
ate
an
d
r
o
b
u
s
t
cita
tio
n
co
u
n
t
f
o
r
ec
asts
co
n
f
ir
m
ed
v
ia
r
ig
o
r
o
u
s
k
-
f
o
l
d
cr
o
s
s
-
v
alid
atio
n
.
Ou
r
m
o
d
el'
s
g
en
er
aliza
b
le
d
ee
p
lear
n
in
g
ar
ch
itectu
r
e
an
d
o
p
t
im
izatio
n
p
r
o
ce
s
s
en
ab
le
r
eli
ab
le,
r
ea
l
-
wo
r
ld
ap
p
licatio
n
t
o
u
n
s
ee
n
s
ch
o
lar
ly
p
u
b
licatio
n
s
.
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:
2
5
0
2
-
4
7
52
F
o
r
ec
a
s
tin
g
r
esea
r
ch
in
flu
en
ce
:
a
r
ec
u
r
r
en
t n
eu
r
a
l n
etw
o
r
k
…
(
N
a
s
er J
a
ma
l
)
1079
Fig
u
r
e
2
.
E
x
p
er
im
e
n
t 2
(
c
r
o
s
s
-
v
alid
atio
n
)
m
o
d
el
p
e
r
f
o
r
m
an
c
e
co
m
p
ar
is
o
n
T
h
e
W
ilco
x
o
n
s
ig
n
ed
-
r
an
k
te
s
t
is
a
n
o
n
-
p
ar
am
etr
ic
s
tatis
ti
ca
l
h
y
p
o
th
esis
test
u
s
ed
to
co
m
p
ar
e
two
r
elate
d
s
am
p
les.
T
a
b
le
7
s
h
o
w
s
th
e
p
-
v
alu
es
f
o
r
ea
ch
co
m
p
a
r
is
o
n
b
etwe
en
th
e
two
m
o
d
els.
T
h
e
p
-
v
alu
e
is
t
h
e
p
r
o
b
a
b
ilit
y
o
f
o
b
s
er
v
i
n
g
a
te
s
t
s
tati
s
tic
as
ex
tr
em
e
as
th
e
o
n
e
c
o
m
p
u
ted
,
ass
u
m
in
g
th
e
n
u
ll
h
y
p
o
t
h
esis
is
ac
tu
al.
B
ased
o
n
th
e
p
-
v
alu
es,
we
ca
n
m
ak
e
th
e
f
o
llo
win
g
co
n
clu
s
io
n
s
:
−
No
s
ig
n
if
ican
t
d
if
f
e
r
en
ce
e
x
is
ts
b
etwe
en
r
an
d
o
m
f
o
r
est
a
n
d
s
u
p
p
o
r
t
v
ec
t
o
r
r
e
g
r
ess
io
n
m
o
d
els
in
all
ev
alu
atio
n
m
etr
ics s
in
ce
th
e
p
-
v
alu
e
is
m
o
r
e
s
ig
n
if
ica
n
t th
an
0
.
0
5
(
th
e
u
s
u
al
s
ig
n
if
ica
n
ce
le
v
el)
.
−
T
h
er
e
is
a
s
ig
n
if
ican
t
d
if
f
er
e
n
ce
b
etwe
en
r
an
d
o
m
f
o
r
est
a
n
d
m
u
lti
-
lay
er
p
er
ce
p
tr
o
n
m
o
d
els
r
eg
ar
d
in
g
R
MSE
an
d
a
cc
u
r
ac
y
ev
alu
at
io
n
m
etr
ics
s
in
ce
th
e
p
-
v
alu
e
is
les
s
th
an
0
.
0
5
.
Ho
wev
er
,
th
er
e
is
n
o
s
ig
n
if
ican
t d
if
f
er
e
n
ce
b
etwe
en
AUC,
p
r
ec
is
io
n
,
r
ec
all
,
an
d
F1
-
s
co
r
e.
−
T
h
er
e
is
a
s
ig
n
if
ican
t
d
if
f
er
en
ce
b
etwe
en
th
e
r
an
d
o
m
f
o
r
est
an
d
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
(
R
NN)
m
o
d
els
in
all
ev
alu
atio
n
m
etr
ics s
in
ce
th
e
p
-
v
alu
e
is
less
th
an
0
.
0
5
.
−
T
h
er
e
is
n
o
s
ig
n
if
ican
t
d
if
f
er
e
n
ce
b
etwe
en
s
u
p
p
o
r
t
v
ec
to
r
r
eg
r
ess
io
n
an
d
m
u
lti
-
lay
er
p
er
ce
p
tr
o
n
m
o
d
els
in
all
ev
alu
atio
n
m
etr
ics.
−
T
h
er
e
is
a
s
ig
n
if
ican
t
d
if
f
er
e
n
ce
b
etwe
en
s
u
p
p
o
r
t
v
ec
to
r
r
eg
r
ess
io
n
an
d
p
r
o
p
o
s
ed
a
p
p
r
o
ac
h
(
R
NN)
m
o
d
els
r
eg
ar
d
in
g
R
MSE
an
d
AUC
ev
alu
atio
n
m
etr
ics
s
in
ce
th
e
p
-
v
alu
e
is
less
th
an
0
.
0
5
.
Ho
we
v
er
,
th
er
e
is
n
o
s
ig
n
if
ican
t d
if
f
er
e
n
ce
in
ac
c
u
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all
,
an
d
F1
s
co
r
e.
−
T
h
er
e
is
a
s
ig
n
if
ican
t
d
i
f
f
er
en
ce
b
etwe
en
th
e
m
u
lti
-
la
y
e
r
p
er
ce
p
tr
o
n
an
d
p
r
o
p
o
s
ed
a
p
p
r
o
ac
h
(
R
NN)
m
o
d
els
r
eg
ar
d
i
n
g
R
MSE
an
d
AUC
ev
alu
atio
n
m
etr
ics
s
in
ce
th
e
p
-
v
alu
e
is
less
th
an
0
.
0
5
.
Ho
wev
er
,
th
er
e
is
n
o
s
ig
n
if
ican
t d
if
f
e
r
en
ce
in
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all
,
a
n
d
F1
s
co
r
e.
Evaluation Warning : The document was created with Spire.PDF for Python.