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
7
,
No
.
3
,
Ma
r
ch
20
2
5
,
p
p
.
2
0
3
2
~
20
4
3
I
SS
N:
2
502
-
4
7
52
,
DOI
: 1
0
.
1
1
5
9
1
/ijee
cs
.v
3
7
.
i
3
.
pp
203
2
-
20
4
3
2032
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs
.
ia
esco
r
e.
co
m
M
ultif
a
cet
ed
a
pp
ro
a
ch f
o
r anticipa
ting lea
rner
per
forma
nce
using
para
meter
weig
htag
e and en
sem
ble alg
o
rithm
fusio
n
Sh
a
bn
a
m
Ara
S J
a
ha
g
irda
r,
T
a
nu
j
a
Ra
m
a
cha
nd
ra
ia
h
D
e
p
a
r
t
me
n
t
o
f
C
o
mp
u
t
e
r
S
c
i
e
n
c
e
a
n
d
En
g
i
n
e
e
r
i
n
g
,
U
V
C
E,
B
a
n
g
a
l
o
r
e
U
n
i
v
e
r
si
t
y
,
B
a
n
g
a
l
o
r
e
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Feb
29
,
2
0
2
4
R
ev
is
ed
Oct
8
,
202
4
Acc
ep
ted
Oct
28
,
2
0
2
4
An
ti
c
ip
a
ti
n
g
stu
d
e
n
t
p
e
rfo
rm
a
n
c
e
h
a
s
g
a
rn
e
re
d
sig
n
ifi
c
a
n
t
a
tt
e
n
ti
o
n
i
n
e
d
u
c
a
ti
o
n
re
se
a
rc
h
fo
r
o
ffe
rin
g
e
a
rly
in
si
g
h
ts t
h
a
t
e
n
a
b
le t
ime
ly
i
n
t
e
rv
e
n
ti
o
n
s
a
n
d
p
e
rso
n
a
li
z
e
d
su
p
p
o
r
t,
u
l
ti
m
a
tely
imp
ro
v
in
g
stu
d
e
n
t
su
c
c
e
ss
a
n
d
re
ten
ti
o
n
ra
tes
.
Th
is
re
se
a
rc
h
fo
c
u
se
s
o
n
e
n
h
a
n
c
in
g
t
h
e
a
c
c
u
ra
c
y
a
n
d
e
ffi
c
ien
c
y
o
f
stu
d
e
n
t
p
e
rf
o
rm
a
n
c
e
p
re
d
icti
o
n
m
o
d
e
ls
b
y
e
m
p
l
o
y
in
g
a
h
y
b
rid
e
n
se
m
b
le
fra
m
e
wo
rk
th
a
t
in
teg
ra
tes
we
ig
h
ted
fe
a
tu
re
se
lec
ti
o
n
wit
h
m
e
t
a
-
lea
rn
e
r
-
b
a
se
d
a
p
p
ro
a
c
h
e
s.
A
we
ig
h
ted
f
e
a
tu
re
se
lec
ti
o
n
m
e
th
o
d
wa
s
e
m
p
lo
y
e
d
to
p
rio
ri
ti
z
e
th
e
m
o
st
in
f
lu
e
n
t
ial
o
f
t
h
e
2
3
p
a
ra
m
e
ters
in
th
e
d
a
tas
e
t,
e
n
h
a
n
c
i
n
g
p
re
d
ictio
n
a
c
c
u
ra
c
y
wh
il
e
re
d
u
c
in
g
th
e
c
o
m
p
u
tati
o
n
a
l
b
u
rd
e
n
.
Th
e
se
p
a
ra
m
e
ters
we
re
th
e
n
u
se
d
to
b
u
i
ld
a
h
y
b
r
id
e
n
se
m
b
le
m
o
d
e
l
b
y
c
o
m
b
in
i
n
g
b
a
se
lea
rn
e
rs
with
m
e
ta
-
lea
rn
e
rs,
sy
ste
m
a
ti
c
a
ll
y
tu
n
e
d
u
sin
g
h
y
p
e
r
p
a
ra
m
e
ter
o
p
ti
m
iza
ti
o
n
.
Th
is ap
p
ro
a
c
h
a
ime
d
to
fu
rt
h
e
r
imp
r
o
v
e
p
re
d
ictio
n
a
c
c
u
ra
c
y
b
y
fu
sin
g
m
u
lt
i
p
le b
a
se
lea
rn
e
rs,
lev
e
ra
g
in
g
t
h
e
stre
n
g
th
s
o
f
d
iffere
n
t
a
lg
o
rit
h
m
s
fo
r
m
o
re
a
c
c
u
ra
te
p
re
d
ictio
n
s.
Th
e
p
ro
p
o
se
d
h
y
b
rid
m
o
d
e
l
wa
s
v
a
li
d
a
te
d
a
c
ro
ss
d
iffere
n
t
fe
a
tu
re
s
se
le
c
ted
b
a
se
d
o
n
fe
a
tu
re
imp
o
rta
n
c
e
u
sin
g
ra
n
d
o
m
fo
re
st
(RF
).
An
a
c
c
u
ra
c
y
o
f
9
8
.
3
8
%
wa
s
a
c
h
iev
e
d
wh
e
n
a
ll
2
3
fe
a
tu
re
s
we
re
c
o
n
sid
e
re
d
a
n
d
a
n
a
c
c
u
ra
c
y
o
f
9
7
.
1
3
%
wa
s
a
c
h
iev
e
d
wh
e
n
t
h
e
to
p
1
0
fe
a
tu
re
s
we
re
u
se
d
.
Th
e
re
se
a
rc
h
h
ig
h
li
g
h
ts
th
e
si
g
n
ifi
c
a
n
c
e
o
f
e
a
rly
p
re
d
ictio
n
f
o
r
p
ro
m
p
t
i
n
terv
e
n
ti
o
n
a
n
d
d
e
m
o
n
stra
tes
h
o
w
fe
a
tu
re
we
ig
h
ti
n
g
c
a
n
b
o
o
st mo
d
e
l
e
ffica
c
y
.
K
ey
w
o
r
d
s
:
E
n
s
em
b
le
alg
o
r
ith
m
Featu
r
e
im
p
o
r
tan
ce
Me
ta
-
lear
n
er
Par
am
eter
weig
h
tin
g
Stu
d
en
t p
er
f
o
r
m
an
ce
p
r
e
d
ictio
n
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
:
Sh
ab
n
am
Ar
a
S Jah
ag
ir
d
ar
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
an
d
E
n
g
in
ee
r
in
g
,
UVCE,
B
an
g
alo
r
e
Un
iv
er
s
ity
B
an
g
alo
r
e,
I
n
d
ia
E
m
ail:
s
h
ab
n
am
.
jah
ag
i
r
d
ar
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
T
h
er
e
h
as
b
ee
n
a
tr
em
e
n
d
o
u
s
g
r
o
wth
in
t
h
e
ad
o
p
tio
n
o
f
o
n
lin
e
lear
n
in
g
p
latf
o
r
m
s
p
r
i
m
ar
ily
d
r
iv
e
n
b
y
th
e
ea
s
e
o
f
ac
ce
s
s
ib
ilit
y
to
th
ese
co
u
r
s
es
an
d
also
d
u
e
t
o
th
e
C
OVI
D
-
19
p
an
d
em
ic.
W
h
ile
th
is
s
h
if
t
h
as
g
r
ea
tly
b
en
ef
ited
s
tu
d
en
ts
a
n
d
f
ac
u
lty
alik
e
,
it
also
p
r
es
en
ts
ch
allen
g
es
s
u
ch
as
m
ain
tain
in
g
lea
r
n
er
en
g
ag
em
e
n
t
an
d
m
o
tiv
atio
n
t
h
r
o
u
g
h
o
u
t
th
e
c
o
u
r
s
e.
Hen
ce
g
ain
in
g
i
n
s
ig
h
ts
in
to
f
ac
to
r
s
th
at
af
f
ec
t
lear
n
er
s
’
p
er
f
o
r
m
an
ce
a
n
d
p
r
o
v
id
in
g
ea
r
ly
in
ter
v
e
n
tio
n
is
r
eq
u
ir
e
d
b
y
ed
u
ca
to
r
s
t
o
id
e
n
tify
th
e
s
tu
d
en
ts
wh
o
ar
e
a
t
r
is
k
o
f
d
r
o
p
p
in
g
o
u
t o
f
t
h
ese
co
u
r
s
es o
r
u
n
d
e
r
p
er
f
o
r
m
in
g
[
1
]
,
[
2
]
.
T
h
e
liter
atu
r
e
r
ev
iew
ex
p
lo
r
es
th
e
ex
is
tin
g
r
esear
ch
o
n
p
r
ed
ictiv
e
m
o
d
els
in
ed
u
ca
tio
n
,
f
o
cu
s
in
g
o
n
th
e
ef
f
icac
y
o
f
p
ar
am
eter
weig
h
in
g
a
n
d
en
s
em
b
le
alg
o
r
ith
m
s
.
I
t
ex
am
i
n
es
v
ar
io
u
s
s
tu
d
ies
th
at
h
a
v
e
in
v
esti
g
ated
f
ac
to
r
s
in
f
lu
en
cin
g
lear
n
er
p
er
f
o
r
m
an
ce
an
d
th
e
s
u
cc
ess
o
f
ea
r
ly
in
ter
v
en
tio
n
s
tr
ateg
ies
in
o
n
lin
e
lear
n
in
g
en
v
ir
o
n
m
e
n
ts
.
T
h
e
ap
p
licatio
n
o
f
v
ar
i
o
u
s
class
if
ic
atio
n
m
o
d
els
was
em
p
h
asized
to
f
o
r
ec
ast
s
tu
d
en
t
p
er
f
o
r
m
an
ce
,
h
ig
h
lig
h
tin
g
th
e
im
p
o
r
tan
ce
o
f
f
ea
tu
r
e
s
elec
tio
n
an
d
t
h
e
in
f
lu
en
ce
o
f
v
ar
io
u
s
attr
ib
u
tes
o
n
p
r
ed
ictio
n
ac
cu
r
ac
y
[
3
]
.
T
h
e
in
teg
r
atio
n
o
f
ex
p
lain
ab
le
ar
tific
ial
in
tellig
en
ce
(
XAI
)
was
ex
p
lo
r
ed
with
p
r
ed
ictiv
e
m
o
d
els,
o
f
f
e
r
in
g
a
n
o
v
el
p
er
s
p
ec
tiv
e
o
n
h
o
w
th
es
e
tech
n
o
lo
g
ies
ca
n
n
o
t
o
n
ly
p
r
ed
ict
o
u
tc
o
m
es
b
u
t
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
Mu
ltif
a
ce
ted
a
p
p
r
o
a
c
h
fo
r
a
n
t
icip
a
tin
g
lea
r
n
er p
erfo
r
ma
n
ce
…
(
S
h
a
b
n
a
m
A
r
a
S
Ja
h
a
g
ir
d
a
r
)
2033
also
p
r
o
v
id
e
ac
tio
n
ab
le
in
s
ig
h
ts
th
r
o
u
g
h
p
r
escr
ip
tiv
e
a
n
aly
tics
[
4
]
.
A
co
m
p
ar
ativ
e
an
aly
s
is
o
f
d
if
f
er
en
t
p
r
ed
ictiv
e
m
o
d
els
was
p
r
o
v
id
ed
in
th
e
co
n
tex
t
o
f
p
u
b
lic
s
ch
o
o
ls
in
B
r
az
il,
u
n
d
er
s
co
r
in
g
th
e
ef
f
ec
tiv
en
ess
o
f
en
s
em
b
l
e
m
eth
o
d
s
in
im
p
r
o
v
i
n
g
p
r
e
d
ictio
n
ac
c
u
r
ac
y
[
5
]
.
A
s
y
s
tem
atic
r
ev
iew
was
co
n
d
u
cted
o
f
p
r
ed
ictiv
e
m
o
d
elin
g
tech
n
iq
u
es,
id
en
tify
in
g
k
ey
tr
en
d
s
an
d
ch
allen
g
es
in
th
e
f
ield
,
p
ar
ticu
lar
ly
th
e
v
ar
iab
ilit
y
in
m
o
d
el
p
er
f
o
r
m
an
ce
ac
r
o
s
s
d
if
f
er
en
t
ed
u
ca
tio
n
al
co
n
tex
ts
[
6
]
.
L
astl
y
in
v
esti
g
atio
n
o
f
th
e
u
s
e
o
f
m
ac
h
in
e
lear
n
in
g
(
ML
)
alg
o
r
ith
m
s
f
o
r
f
o
r
ec
asti
n
g
s
tu
d
en
t
p
e
r
f
o
r
m
an
ce
,
d
em
o
n
s
tr
atin
g
th
at
m
o
d
els
in
co
r
p
o
r
atin
g
a
b
r
o
ad
e
r
s
et
o
f
f
ea
tu
r
es,
in
clu
d
in
g
b
eh
av
io
r
al
an
d
d
em
o
g
r
ap
h
ic
d
ata,
ten
d
to
y
ield
m
o
r
e
ac
cu
r
at
e
p
r
ed
ict
io
n
s
[
7
]
.
C
o
llectiv
ely
,
th
ese
s
tu
d
ies
u
n
d
er
s
co
r
e
th
e
p
o
ten
tial
o
f
p
r
e
d
i
ctiv
e
m
o
d
elin
g
in
ed
u
ca
tio
n
,
wh
ile
also
p
o
in
tin
g
to
th
e
n
ee
d
f
o
r
m
o
r
e
s
o
p
h
is
ticated
tech
n
iq
u
es,
s
u
ch
as
en
s
em
b
le
lear
n
in
g
an
d
XAI
,
to
en
h
an
ce
m
o
d
el
p
er
f
o
r
m
an
ce
an
d
u
s
ab
ilit
y
in
r
ea
l
-
wo
r
ld
ed
u
ca
tio
n
al
s
ettin
g
s
.
T
h
e
in
teg
r
atio
n
o
f
p
r
ed
ictiv
e
m
o
d
elin
g
in
ed
u
ca
tio
n
al
en
v
ir
o
n
m
en
ts
h
as
g
ar
n
er
ed
s
u
b
s
tan
tial
atten
tio
n
in
r
ec
en
t
y
ea
r
s
,
with
a
p
a
r
ticu
lar
f
o
cu
s
o
n
e
n
h
an
ci
n
g
m
o
d
el
ac
c
u
r
ac
y
th
r
o
u
g
h
f
e
atu
r
e
s
elec
tio
n
an
d
p
ar
am
eter
weig
h
tag
e
.
T
h
is
r
ev
i
ew
s
y
n
th
esizes f
in
d
in
g
s
f
r
o
m
k
ey
s
tu
d
ies th
at
h
av
e
em
p
lo
y
e
d
th
ese
m
eth
o
d
s
to
p
r
ed
ict
ac
ad
em
ic
p
er
f
o
r
m
an
c
e.
T
h
e
u
s
e
o
f
m
u
ltis
o
u
r
ce
,
m
u
ltife
atu
r
ed
b
eh
a
v
io
r
al
d
ata
to
p
r
ed
ict
ac
ad
em
ic
p
er
f
o
r
m
an
ce
.
T
h
e
r
esear
ch
er
s
em
p
lo
y
ed
a
n
en
s
em
b
le
o
f
ML
alg
o
r
ith
m
s
to
an
al
y
ze
d
ata
f
r
o
m
v
ar
io
u
s
s
o
u
r
ce
s
,
in
clu
d
in
g
s
tu
d
en
t
d
e
m
o
g
r
a
p
h
i
cs,
o
n
lin
e
ac
tiv
ities
,
an
d
p
h
y
s
ical
b
eh
av
io
r
s
.
T
h
e
s
tu
d
y
f
o
u
n
d
th
at
in
co
r
p
o
r
atin
g
d
iv
er
s
e
f
ea
tu
r
es
f
r
o
m
m
u
ltip
le
s
o
u
r
ce
s
s
ig
n
if
ican
tly
im
p
r
o
v
e
s
th
e
p
r
ed
ictio
n
ac
cu
r
ac
y
o
f
a
ca
d
em
ic
o
u
tco
m
e.
T
h
is
h
ig
h
lig
h
ts
th
e
im
p
o
r
tan
c
e
o
f
c
o
n
s
id
er
in
g
a
wid
e
r
an
g
e
o
f
f
ea
tu
r
es
wh
en
d
ev
elo
p
in
g
p
r
ed
ictiv
e
m
o
d
els
in
ed
u
ca
tio
n
[
8
]
.
I
m
p
r
o
v
in
g
d
r
o
p
o
u
t
f
o
r
ec
asti
n
g
d
u
r
in
g
th
e
C
OVI
D
-
1
9
p
a
n
d
em
ic
u
s
in
g
f
ea
tu
r
e
s
elec
tio
n
tech
n
iq
u
es a
n
d
a
m
u
ltil
ay
er
p
e
r
ce
p
tr
o
n
n
e
u
r
al
n
etwo
r
k
h
as p
r
o
v
en
ef
f
ec
ti
v
e
in
en
h
an
cin
g
m
o
d
el
p
er
f
o
r
m
an
ce
,
p
ar
ticu
lar
ly
i
n
a
ch
allen
g
i
n
g
co
n
tex
t
lik
e
a
p
an
d
e
m
ic.
B
y
ca
r
ef
u
lly
s
elec
tin
g
th
e
m
o
s
t
r
elev
an
t
f
ea
tu
r
es,
th
e
ac
cu
r
ac
y
o
f
d
r
o
p
o
u
t
p
r
e
d
ictio
n
s
was
s
ig
n
if
ican
tly
im
p
r
o
v
ed
,
u
n
d
er
s
co
r
i
n
g
th
e
ess
en
t
ial
r
o
le
o
f
f
ea
t
u
r
e
s
elec
tio
n
in
p
r
ed
ictiv
e
m
o
d
elin
g
[
9
]
.
A
s
y
s
tem
atic
r
ev
iew
o
f
ML
ap
p
licatio
n
s
in
d
ete
r
m
in
in
g
th
e
attr
ib
u
tes
th
at
in
f
lu
en
ce
ac
a
d
em
ic
p
e
r
f
o
r
m
a
n
ce
was
co
n
d
u
cted
.
T
h
e
r
ev
iew
i
d
en
tifie
d
k
e
y
f
ac
to
r
s
s
u
ch
as
s
o
cio
ec
o
n
o
m
ic
s
tatu
s
,
atten
d
a
n
ce
,
an
d
en
g
ag
em
e
n
t
as
cr
itical
p
r
ed
icto
r
s
.
T
h
e
s
tu
d
y
also
e
m
p
h
asized
th
e
n
ee
d
f
o
r
m
o
r
e
r
esear
ch
o
n
th
e
a
p
p
li
ca
tio
n
o
f
f
ea
t
u
r
e
s
elec
tio
n
m
et
h
o
d
s
to
r
ef
i
n
e
th
ese
p
r
e
d
icto
r
s
f
u
r
th
er
,
s
u
g
g
esti
n
g
th
at
a
f
o
cu
s
o
n
p
ar
a
m
eter
wei
g
h
tag
e
co
u
ld
y
ield
ev
en
m
o
r
e
ac
cu
r
ate
m
o
d
els [
1
0
]
.
T
h
e
ch
allen
g
e
o
f
im
p
r
o
v
in
g
tim
e
co
m
p
lex
ity
an
d
ac
c
u
r
ac
y
in
ML
alg
o
r
ith
m
s
was
ad
d
r
ess
ed
b
y
s
elec
tin
g
to
p
'
k
'
f
ea
tu
r
es
wit
h
th
e
h
i
g
h
est
weig
h
ts
f
r
o
m
in
tr
icate
d
atasets
.
T
h
e
s
tu
d
y
d
em
o
n
s
tr
ated
th
at
f
o
cu
s
in
g
o
n
a
s
m
aller
s
u
b
s
et
o
f
h
ig
h
l
y
weig
h
te
d
f
ea
t
u
r
es
ca
n
r
esu
lt
in
n
o
tab
le
en
h
a
n
c
em
en
ts
in
b
o
th
th
e
ef
f
icien
cy
an
d
p
r
ec
is
io
n
o
f
p
r
ed
ictiv
e
m
o
d
els.
T
h
is
ap
p
r
o
ac
h
is
p
ar
ticu
lar
ly
r
elev
an
t
in
e
d
u
ca
tio
n
al
s
ettin
g
s
wh
er
e
d
atasets
ar
e
o
f
ten
c
o
m
p
lex
an
d
f
ea
t
u
r
e
-
r
ic
h
[
1
1
]
.
A
cr
itical
f
ea
tu
r
e
s
elec
tio
n
m
eth
o
d
em
p
l
o
y
in
g
en
s
em
b
l
e
m
eta
-
b
ased
m
o
d
els
was
p
r
o
p
o
s
ed
to
p
r
ed
ict
m
u
lticlas
s
o
u
tco
m
es.
T
h
eir
s
tu
d
y
s
h
o
wed
th
at
co
m
b
in
in
g
m
u
ltip
le
m
o
d
els
an
d
f
o
cu
s
in
g
o
n
cr
itical
f
ea
tu
r
es
ca
n
lead
t
o
m
o
r
e
ac
c
u
r
ate
p
r
ed
ictio
n
s
,
p
ar
ticu
lar
ly
i
n
s
ce
n
ar
io
s
in
v
o
l
v
in
g
m
u
lticlas
s
class
if
icat
io
n
.
T
h
is
f
in
d
i
n
g
is
c
r
u
cial
f
o
r
e
d
u
ca
tio
n
al
c
o
n
tex
ts
wh
er
e
o
u
tco
m
es c
an
b
e
v
ar
ied
an
d
co
m
p
lex
[
1
2
]
.
Var
io
u
s
m
ea
s
u
r
es
o
f
f
ea
tu
r
e
im
p
o
r
tan
ce
wer
e
e
v
alu
ated
to
ex
p
lain
class
if
icatio
n
m
o
d
els
.
T
h
eir
r
esear
ch
h
ig
h
lig
h
ted
th
e
d
if
f
er
en
ce
s
in
h
o
w
d
if
f
e
r
en
t
alg
o
r
ith
m
s
ass
ess
f
ea
tu
r
e
i
m
p
o
r
tan
ce
a
n
d
th
e
im
p
licatio
n
s
th
is
h
as
f
o
r
m
o
d
el
in
ter
p
r
etab
ilit
y
.
Un
d
e
r
s
tan
d
in
g
f
ea
t
u
r
e
im
p
o
r
tan
ce
is
ess
en
tial
f
o
r
ed
u
ca
to
r
s
an
d
p
o
licy
m
ak
e
r
s
wh
o
n
ee
d
to
in
ter
p
r
et
an
d
ac
t
o
n
m
o
d
e
l
p
r
ed
ictio
n
s
[
1
3
]
.
A
f
lex
ib
le
f
ea
tu
r
e
s
elec
tio
n
ap
p
r
o
ac
h
was
d
e
v
elo
p
ed
f
o
r
p
r
ed
ictin
g
s
tu
d
e
n
t
p
er
f
o
r
m
a
n
c
e
in
o
n
lin
e
c
o
u
r
s
es.
T
h
eir
s
tu
d
y
em
p
h
asized
th
e
n
ee
d
f
o
r
a
d
ap
tab
le
m
eth
o
d
s
th
at
ca
n
h
a
n
d
le
t
h
e
u
n
iq
u
e
c
h
allen
g
es
o
f
o
n
lin
e
ed
u
ca
tio
n
.
B
y
em
p
lo
y
in
g
a
f
lex
ib
le
f
ea
tu
r
e
s
elec
tio
n
s
tr
ateg
y
,
th
e
y
wer
e
ab
le
to
im
p
r
o
v
e
th
e
ac
cu
r
ac
y
o
f
t
h
eir
p
r
ed
ictiv
e
m
o
d
els,
d
em
o
n
s
tr
atin
g
t
h
e
im
p
o
r
tan
ce
o
f
ad
ap
ta
b
ilit
y
in
f
ea
tu
r
e
s
elec
tio
n
[
1
4
]
.
A
co
m
p
r
eh
e
n
s
iv
e
s
u
r
v
e
y
o
f
f
ea
tu
r
e
s
elec
tio
n
tech
n
iq
u
es
s
p
an
n
in
g
o
v
er
m
o
r
e
th
an
two
d
ec
ad
es
o
f
r
esear
ch
was
p
r
o
v
id
ed
.
T
h
ey
h
ig
h
lig
h
ted
th
e
e
v
o
lu
tio
n
o
f
th
ese
tech
n
iq
u
es
an
d
th
eir
in
cr
ea
s
in
g
r
elev
an
ce
in
ML
ap
p
licatio
n
s
,
in
cl
u
d
in
g
e
d
u
ca
tio
n
al
d
ata
m
in
in
g
.
T
h
e
s
u
r
v
ey
s
u
g
g
ested
th
at
o
n
g
o
in
g
ad
v
a
n
ce
m
en
ts
in
f
ea
tu
r
e
s
elec
tio
n
m
eth
o
d
s
will c
o
n
tin
u
e
to
p
lay
a
cr
itical
r
o
le
in
en
h
an
cin
g
p
r
ed
ictiv
e
m
o
d
e
llin
g
[
1
5
]
.
An
o
th
er
ex
ten
s
iv
e
r
ev
iew
f
o
cu
s
ed
o
n
f
ea
tu
r
e
s
elec
tio
n
ac
r
o
s
s
v
ar
io
u
s
f
ield
s
o
f
ML
.
T
h
eir
f
i
n
d
in
g
s
u
n
d
er
s
co
r
e
d
th
e
im
p
o
r
tan
ce
o
f
f
ea
tu
r
e
s
elec
tio
n
in
im
p
r
o
v
in
g
m
o
d
el
ac
cu
r
a
cy
an
d
r
e
d
u
cin
g
co
m
p
u
tatio
n
al
co
m
p
lex
ity
.
T
h
e
s
tu
d
y
also
h
ig
h
lig
h
ted
th
e
n
e
ed
f
o
r
d
o
m
ai
n
-
s
p
ec
if
ic
f
ea
tu
r
e
s
elec
tio
n
tech
n
iq
u
es,
p
ar
ticu
lar
ly
in
ed
u
ca
tio
n
,
wh
er
e
th
e
r
elev
a
n
ce
o
f
f
ea
tu
r
e
s
ca
n
v
ar
y
wid
el
y
[
1
6
]
.
T
ab
le
1
s
h
o
ws
co
m
p
ar
is
o
n
o
f
tech
n
iq
u
es
f
o
r
lear
n
er
s
’
p
er
f
o
r
m
an
ce
p
r
ed
ictio
n
.
Ov
er
all,
th
e
s
e
s
tu
d
ies
d
em
o
n
s
tr
ate
th
e
s
ig
n
if
ican
t
i
m
p
ac
t
th
at
f
ea
tu
r
e
s
elec
tio
n
a
n
d
p
a
r
am
eter
weig
h
tag
e
ca
n
h
av
e
o
n
th
e
ac
cu
r
ac
y
f
ea
tu
r
es,
r
esear
ch
e
r
s
ca
n
d
e
v
elo
p
m
o
r
e
ac
cu
r
ate
an
d
in
t
er
p
r
etab
le
m
o
d
els,
u
ltima
tely
lead
in
g
to
b
etter
ed
u
ca
tio
n
al
o
u
tco
m
es.
T
h
is
r
ev
iew
u
n
d
er
s
co
r
es
th
e
c
r
itical
r
o
le
th
at
f
ea
tu
r
e
s
elec
tio
n
an
d
p
a
r
am
ete
r
weig
h
tag
e
p
la
y
in
en
h
an
cin
g
p
r
ed
ictiv
e
m
o
d
elin
g
ac
cu
r
ac
y
i
n
ed
u
ca
tio
n
al
c
o
n
tex
ts
.
I
t
also
h
ig
h
lig
h
ts
th
e
n
ee
d
f
o
r
co
n
tin
u
ed
r
esear
ch
in
th
is
ar
ea
to
r
ef
in
e
th
ese
tech
n
iq
u
es
f
u
r
th
er
an
d
ap
p
l
y
th
em
to
a
b
r
o
ad
e
r
r
an
g
e
o
f
ed
u
ca
tio
n
al
s
ce
n
ar
io
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
7
,
No
.
3
,
Ma
r
ch
20
2
5
:
2
0
3
2
-
20
4
3
2034
T
ab
le
1
.
C
o
m
p
a
r
is
o
n
o
f
tech
n
i
q
u
es f
o
r
lea
r
n
er
s
’
p
e
r
f
o
r
m
an
c
e
p
r
ed
ictio
n
A
u
t
h
o
r
M
e
t
h
o
d
o
l
o
g
y
u
se
d
F
o
c
u
s
o
f
r
e
s
e
a
r
c
h
K
e
y
f
i
n
d
i
n
g
s
Li
mi
t
a
t
i
o
n
s
R
a
y
e
t
a
l
.
[
1
7
]
En
se
mb
l
e
met
h
o
d
s
a
n
d
d
e
e
p
l
e
a
r
n
i
n
g
mo
d
e
l
s
A
c
a
d
e
mi
c
p
e
r
f
o
r
m
a
n
c
e
p
r
e
d
i
c
t
i
o
n
En
se
mb
l
e
met
h
o
d
s
i
mp
r
o
v
e
a
c
c
u
r
a
c
y
i
n
p
r
e
d
i
c
t
i
n
g
a
c
a
d
e
m
i
c
p
e
r
f
o
r
m
a
n
c
e
.
M
o
r
e
d
i
v
e
r
se
d
a
t
a
se
t
s
a
r
e
n
e
e
d
e
d
f
o
r
b
r
o
a
d
e
r
a
p
p
l
i
c
a
b
i
l
i
t
y
.
A
f
sh
a
r
e
t
a
l
.
[
1
8
]
En
se
mb
l
e
o
f
m
e
t
a
-
l
e
a
r
n
e
r
s
R
a
n
k
i
n
g
a
n
d
c
o
m
b
i
n
i
n
g
p
r
e
d
i
c
t
i
v
e
mo
d
e
l
s
En
se
mb
l
e
t
e
c
h
n
i
q
u
e
s
r
e
d
u
c
e
p
r
e
d
i
c
t
i
o
n
v
a
r
i
a
n
c
e
,
i
mp
r
o
v
i
n
g
o
v
e
r
a
l
l
p
e
r
f
o
r
ma
n
c
e
.
La
c
k
o
f
l
a
b
e
l
e
d
d
a
t
a
a
f
f
e
c
t
s
t
h
e
r
o
b
u
s
t
n
e
ss
o
f
t
h
e
p
r
e
d
i
c
t
i
o
n
s.
Li
e
t
a
l
.
[
1
9
]
I
n
t
e
l
l
i
g
e
n
t
f
u
z
z
y
r
e
g
r
e
ss
i
o
n
c
l
a
ss
i
f
i
c
a
t
i
o
n
(
I
F
R
C
)
mo
d
e
l
w
i
t
h
e
n
sem
b
l
e
t
e
c
h
n
i
q
u
e
s
P
r
e
d
i
c
t
i
n
g
s
t
u
d
e
n
t
p
e
r
f
o
r
m
a
n
c
e
i
n
m
u
si
c
e
d
u
c
a
t
i
o
n
R
o
b
u
st
f
r
a
mew
o
r
k
p
r
o
v
i
d
e
d
b
y
e
n
s
e
mb
l
e
t
e
c
h
n
i
q
u
e
s s
i
g
n
i
f
i
c
a
n
t
l
y
e
n
h
a
n
c
e
s
p
r
e
d
i
c
t
i
o
n
a
c
c
u
r
a
c
y
.
F
o
c
u
ses
sp
e
c
i
f
i
c
a
l
l
y
o
n
mu
si
c
e
d
u
c
a
t
i
o
n
,
l
i
m
i
t
i
n
g
i
t
s
g
e
n
e
r
a
l
i
z
a
b
i
l
i
t
y
.
N
i
y
o
g
i
s
u
b
i
z
o
e
t
a
l
.
[
2
0
]
Tw
o
-
l
a
y
e
r
e
n
sem
b
l
e
a
p
p
r
o
a
c
h
u
si
n
g
S
t
a
c
k
e
d
G
e
n
e
r
a
l
i
z
a
t
i
o
n
P
r
e
d
i
c
t
i
n
g
s
t
u
d
e
n
t
d
r
o
p
o
u
t
i
n
u
n
i
v
e
r
s
i
t
y
c
l
a
ss
e
s
Th
e
t
w
o
-
l
a
y
e
r
e
n
sem
b
l
e
mo
d
e
l
w
i
t
h
st
a
c
k
e
d
g
e
n
e
r
a
l
i
z
a
t
i
o
n
a
c
c
u
r
a
t
e
l
y
p
r
e
d
i
c
t
s s
t
u
d
e
n
t
d
r
o
p
o
u
t
.
Th
e
mo
d
e
l
's
c
o
mp
l
e
x
i
t
y
mi
g
h
t
l
i
mi
t
i
t
s
sca
l
a
b
i
l
i
t
y
t
o
l
a
r
g
e
r
d
a
t
a
s
e
t
s
o
r
d
i
f
f
e
r
e
n
t
e
d
u
c
a
t
i
o
n
a
l
c
o
n
t
e
x
t
s.
A
l
su
l
a
m
i
e
t
a
l
.
[
2
1
]
En
se
mb
l
e
t
e
c
h
n
i
q
u
e
s
i
n
c
l
u
d
i
n
g
R
F
,
X
G
B
o
o
st
,
a
n
d
l
i
g
h
t
g
r
a
d
i
e
n
t
-
b
o
o
s
t
i
n
g
mac
h
i
n
e
(
Li
g
h
t
G
B
M
)
mo
d
e
l
s
En
h
a
n
c
i
n
g
t
h
e
p
r
e
d
i
c
t
i
o
n
a
c
c
u
r
a
c
y
o
f
e
-
l
e
a
r
n
i
n
g
s
t
u
d
e
n
t
p
e
r
f
o
r
m
a
n
c
e
Li
g
h
t
G
B
M
s
h
o
w
s
t
h
e
b
e
s
t
p
e
r
f
o
r
m
a
n
c
e
i
n
p
r
e
d
i
c
t
i
n
g
e
-
l
e
a
r
n
i
n
g
st
u
d
e
n
t
o
u
t
c
o
mes
.
Th
e
st
u
d
y
i
s
l
i
m
i
t
e
d
t
o
a
sp
e
c
i
f
i
c
e
-
l
e
a
r
n
i
n
g
p
l
a
t
f
o
r
m
,
w
h
i
c
h
ma
y
a
f
f
e
c
t
t
h
e
g
e
n
e
r
a
l
i
z
a
b
i
l
i
t
y
o
f
t
h
e
r
e
su
l
t
s.
Al
-
A
meri
e
t
a
l
.
[
2
2
]
En
se
mb
l
e
mo
d
e
l
w
i
t
h
mu
l
t
i
m
e
d
i
a
-
b
a
s
e
d
b
i
g
d
a
t
a
A
c
a
d
e
mi
c
s
u
c
c
e
ss
p
r
e
d
i
c
t
i
o
n
u
si
n
g
mu
l
t
i
m
e
d
i
a
d
a
t
a
f
r
o
m
l
e
a
r
n
i
n
g
ma
n
a
g
e
me
n
t
sy
st
e
m (L
M
S
)
En
se
mb
l
e
mo
d
e
l
s
h
o
w
s
h
i
g
h
a
c
c
u
r
a
c
y
i
n
p
r
e
d
i
c
t
i
n
g
s
t
u
d
e
n
t
su
c
c
e
ss.
R
e
q
u
i
r
e
s
a
l
a
r
g
e
a
mo
u
n
t
o
f
m
u
l
t
i
m
e
d
i
a
d
a
t
a
,
w
h
i
c
h
ma
y
n
o
t
a
l
w
a
y
s
b
e
a
v
a
i
l
a
b
l
e
.
P
e
k
e
t
a
l
.
[
2
3
]
M
L
m
o
d
e
l
s
i
n
c
l
u
d
i
n
g
R
F
,
S
V
M
,
a
n
d
d
e
c
i
si
o
n
t
r
e
e
(
D
T)
I
d
e
n
t
i
f
y
i
n
g
a
t
-
r
i
s
k
st
u
d
e
n
t
s
a
n
d
mi
n
i
mi
z
i
n
g
f
a
i
l
u
r
e
M
L
t
e
c
h
n
i
q
u
e
s
e
f
f
e
c
t
i
v
e
l
y
r
e
d
u
c
e
st
u
d
e
n
t
f
a
i
l
u
r
e
r
a
t
e
s
b
y
e
a
r
l
y
i
d
e
n
t
i
f
i
c
a
t
i
o
n
,
w
i
t
h
R
F
b
e
i
n
g
t
h
e
m
o
st
su
c
c
e
ssf
u
l
.
Th
e
st
u
d
y
mi
g
h
t
m
i
ss
f
a
c
t
o
r
s,
c
a
u
si
n
g
i
n
a
c
c
u
r
a
t
e
p
r
e
d
i
c
t
i
o
n
s
.
Zh
a
n
g
e
t
a
l
.
[
2
4
]
En
se
mb
l
e
mo
d
e
l
c
o
m
b
i
n
i
n
g
D
T
,
R
F
,
a
n
d
G
B
P
r
e
d
i
c
t
i
n
g
s
t
u
d
e
n
t
a
c
h
i
e
v
e
me
n
t
s
i
n
l
e
a
r
n
i
n
g
P
o
r
t
u
g
u
e
s
e
En
se
mb
l
e
mo
d
e
l
s,
e
sp
e
c
i
a
l
l
y
g
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
,
e
x
c
e
l
i
n
p
r
e
d
i
c
t
i
n
g
a
c
h
i
e
v
e
me
n
t
s
.
Th
e
st
u
d
y
i
s
l
i
m
i
t
e
d
t
o
l
a
n
g
u
a
g
e
l
e
a
r
n
i
n
g
,
w
h
i
c
h
may
l
i
mi
t
t
h
e
g
e
n
e
r
a
l
i
z
a
b
i
l
i
t
y
o
f
t
h
e
f
i
n
d
i
n
g
s
t
o
o
t
h
e
r
su
b
j
e
c
t
s
.
B
a
k
y
a
l
a
k
s
h
m
i
e
t
a
l
.
[
2
5
]
M
u
l
t
i
-
v
i
e
w
d
e
e
p
l
e
a
r
n
i
n
g
w
i
t
h
e
n
s
e
mb
l
e
t
e
c
h
n
i
q
u
e
s
P
r
e
d
i
c
t
i
n
g
s
t
u
d
e
n
t
p
e
r
f
o
r
m
a
n
c
e
a
n
d
i
d
e
n
t
i
f
y
i
n
g
h
i
g
h
-
r
i
s
k
st
u
d
e
n
t
s
M
u
l
t
i
-
v
i
e
w
d
e
e
p
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
i
n
c
r
e
a
s
e
s a
c
c
u
r
a
c
y
i
n
e
a
r
l
y
i
d
e
n
t
i
f
i
c
a
t
i
o
n
o
f
h
i
g
h
-
r
i
sk
st
u
d
e
n
t
s.
Li
mi
t
e
d
t
o
e
d
u
c
a
t
i
o
n
a
l
set
t
i
n
g
s
,
ma
y
n
o
t
a
p
p
l
y
t
o
o
t
h
e
r
d
o
m
a
i
n
s.
Te
o
h
e
t
a
l
.
[
2
6
]
En
se
mb
l
e
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s
i
n
c
l
u
d
i
n
g
b
a
g
g
i
n
g
,
b
o
o
st
i
n
g
,
a
n
d
s
t
a
c
k
i
n
g
P
r
e
d
i
c
t
i
n
g
s
t
u
d
e
n
t
p
e
r
f
o
r
m
a
n
c
e
i
n
v
i
d
e
o
-
b
a
s
e
d
l
e
a
r
n
i
n
g
e
n
v
i
r
o
n
m
e
n
t
En
se
mb
l
e
l
e
a
r
n
i
n
g
w
i
t
h
B
o
o
st
i
n
g
i
n
c
r
e
a
ses
a
c
c
u
r
a
c
y
i
n
v
i
d
e
o
-
b
a
se
d
p
e
r
f
o
r
m
a
n
c
e
p
r
e
d
i
c
t
i
o
n
.
Th
e
st
u
d
y
f
o
c
u
s
e
s s
o
l
e
l
y
o
n
v
i
d
e
o
-
b
a
s
e
d
l
e
a
r
n
i
n
g
,
w
h
i
c
h
ma
y
n
o
t
r
e
f
l
e
c
t
o
t
h
e
r
f
o
r
ms
o
f
l
e
a
r
n
i
n
g
.
T
h
e
liter
atu
r
e
r
e
v
iew
r
ev
ea
ls
s
ev
er
al
cr
itical
g
ap
s
th
at
r
em
ai
n
u
n
a
d
d
r
ess
ed
in
c
u
r
r
e
n
t
r
esea
r
ch
.
T
h
ese
g
ap
s
h
ig
h
lig
h
t
o
p
p
o
r
tu
n
ities
f
o
r
f
u
r
t
h
er
ex
p
lo
r
atio
n
t
o
en
h
a
n
ce
th
e
ac
cu
r
ac
y
an
d
a
p
p
lica
b
ilit
y
o
f
p
r
e
d
ictiv
e
m
o
d
els
in
e
d
u
ca
tio
n
.
Alth
o
u
g
h
th
er
e
h
as
b
ee
n
s
o
m
e
wo
r
k
t
h
at
h
as
b
ee
n
d
o
n
e
o
n
th
e
im
p
a
ct
o
f
f
ac
to
r
s
s
u
ch
as
s
tu
d
en
t
d
em
o
g
r
ap
h
ics,
o
n
lin
e
ac
tiv
ities
an
d
lear
n
in
g
s
ty
les
o
n
s
tu
d
en
t
p
e
r
f
o
r
m
an
ce
p
r
e
d
i
ctio
n
b
u
t
th
er
e
is
a
lack
o
f
in
-
d
ep
th
an
aly
s
is
o
n
h
o
w
th
ese
f
ac
to
r
s
im
p
ac
t
th
e
p
r
ed
ictio
n
o
u
tco
m
e
o
n
a
n
in
d
iv
id
u
al
b
asis
.
T
h
is
em
p
h
asizes
th
e
n
ee
d
f
o
r
m
o
r
e
r
esear
ch
o
n
th
e
ap
p
licatio
n
o
f
f
ea
tu
r
e
s
elec
tio
n
m
eth
o
d
s
to
r
e
f
in
e
th
ese
p
r
ed
icto
r
s
f
u
r
th
er
,
s
u
g
g
esti
n
g
t
h
at
a
f
o
cu
s
o
n
p
a
r
am
eter
weig
h
tag
e
co
u
l
d
y
ield
e
v
en
m
o
r
e
a
cc
u
r
ate
m
o
d
els.
W
h
ile
en
s
em
b
le
tech
n
iq
u
es
h
a
v
e
p
r
o
v
en
ef
f
e
ctiv
e
in
p
r
e
d
ictin
g
s
tu
d
en
t
p
e
r
f
o
r
m
an
ce
,
th
e
r
e
is
lim
ited
o
r
n
o
r
esear
ch
o
n
t
h
e
im
p
ac
t
o
f
co
m
b
in
in
g
th
ese
tech
n
iq
u
es
with
m
eta
-
lear
n
er
s
.
T
h
e
p
o
ten
tial
f
o
r
m
eta
-
lear
n
er
s
to
f
u
r
th
er
en
h
an
ce
th
e
p
r
ed
ictiv
e
p
o
wer
an
d
r
o
b
u
s
tn
ess
o
f
en
s
em
b
le
m
o
d
els
r
em
ain
s
u
n
d
er
ex
p
l
o
r
ed
,
r
ep
r
esen
tin
g
a
s
ig
n
if
ican
t
g
ap
in
th
e
cu
r
r
en
t
liter
atu
r
e.
T
o
ad
d
r
ess
th
e
id
en
tifie
d
g
ap
s
in
th
e
liter
atu
r
e,
o
u
r
r
esear
ch
f
o
cu
s
ed
o
n
a
two
-
f
o
l
d
ap
p
r
o
ac
h
.
Firstl
y
,
we
co
n
d
u
cted
a
co
m
p
r
e
h
en
s
iv
e
an
aly
s
is
o
f
th
e
im
p
ac
t
o
f
u
s
in
g
weig
h
ted
p
ar
am
eter
s
o
n
s
tu
d
en
t
p
er
f
o
r
m
a
n
ce
with
in
e
n
s
em
b
le
p
r
ed
ictio
n
m
o
d
els.
B
y
in
teg
r
atin
g
th
ese
v
ar
iab
les
in
to
th
e
p
r
ed
ictiv
e
m
o
d
els,
we
aim
e
d
to
en
h
a
n
ce
th
e
in
ter
p
r
etab
ilit
y
o
f
t
h
e
p
r
e
d
ictio
n
s
an
d
p
r
o
v
id
e
ac
tio
n
ab
le
in
s
ig
h
ts
th
at
ca
n
g
u
id
e
m
o
r
e
e
f
f
ec
tiv
e
ed
u
ca
tio
n
al
in
ter
v
en
tio
n
s
.
Seco
n
d
ly
,
we
ex
p
lo
r
ed
t
h
e
n
o
v
el
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
Mu
ltif
a
ce
ted
a
p
p
r
o
a
c
h
fo
r
a
n
t
icip
a
tin
g
lea
r
n
er p
erfo
r
ma
n
ce
…
(
S
h
a
b
n
a
m
A
r
a
S
Ja
h
a
g
ir
d
a
r
)
2035
in
teg
r
atio
n
o
f
m
eta
-
lea
r
n
er
s
with
en
s
em
b
le
tech
n
iq
u
es
to
in
v
esti
g
ate
th
eir
p
o
ten
tial
in
im
p
r
o
v
i
n
g
p
r
e
d
ictio
n
ac
cu
r
ac
y
an
d
r
o
b
u
s
tn
ess
.
T
h
is
in
v
o
lv
es
th
e
d
ev
elo
p
m
e
n
t
an
d
test
in
g
o
f
h
y
b
r
id
m
o
d
els
th
at
co
m
b
in
e
v
ar
io
u
s
en
s
em
b
le
m
eth
o
d
s
with
m
eta
-
lear
n
er
s
,
s
u
ch
as
s
tack
in
g
an
d
b
len
d
in
g
.
Ou
r
r
esear
ch
will
em
p
lo
y
a
r
ig
o
r
o
u
s
ex
p
er
im
en
tal
d
esig
n
,
u
tili
zin
g
cr
o
s
s
-
v
alid
atio
n
tech
n
iq
u
es a
n
d
p
er
f
o
r
m
an
ce
m
etr
ics to
ass
ess
th
e
ef
f
ec
tiv
en
ess
o
f
th
ese
h
y
b
r
id
m
o
d
els
co
m
p
ar
ed
to
tr
ad
itio
n
al
en
s
em
b
le
m
eth
o
d
s
.
T
h
is
ap
p
r
o
ac
h
n
o
t
o
n
ly
f
ills
th
e
ex
is
tin
g
g
ap
s
b
u
t
also
co
n
tr
ib
u
tes
to
th
e
ad
v
an
ce
m
en
t
o
f
p
r
ed
ic
tiv
e
an
aly
tics
in
ed
u
ca
tio
n
b
y
o
f
f
er
in
g
m
o
r
e
co
m
p
r
eh
e
n
s
iv
e
an
d
p
o
wer
f
u
l
m
o
d
els f
o
r
p
r
ed
ictin
g
s
tu
d
en
t
p
er
f
o
r
m
an
ce
.
2.
M
E
T
H
O
D
2
.
1
.
Resea
rc
h
d
esig
n
T
h
is
r
esear
ch
u
tili
ze
s
a
q
u
an
ti
tativ
e
ap
p
r
o
ac
h
t
o
ex
am
i
n
e
th
e
ef
f
ec
tiv
en
ess
o
f
p
a
r
am
eter
weig
h
tag
e
an
d
e
n
s
em
b
le
alg
o
r
ith
m
f
u
s
i
o
n
in
p
r
ed
ictin
g
lear
n
er
p
er
f
o
r
m
an
ce
.
T
h
e
r
esear
ch
m
eth
o
d
in
v
o
lv
es
u
tili
zin
g
h
is
to
r
ical
lear
n
er
d
ata
to
d
ev
elo
p
an
d
v
alid
ate
p
r
e
d
ictiv
e
m
o
d
els.
E
ac
h
s
tep
is
m
eticu
l
o
u
s
ly
d
esig
n
e
d
to
en
s
u
r
e
r
o
b
u
s
tn
ess
an
d
ac
cu
r
a
cy
in
th
e
p
r
ed
ictiv
e
m
o
d
el.
T
h
e
p
r
o
ce
s
s
wo
r
k
f
lo
w
is
d
e
p
icted
in
Fig
u
r
e
1
,
illu
s
tr
atin
g
th
e
s
eq
u
en
tial
s
te
p
s
f
r
o
m
d
ata
co
llectio
n
to
th
e
f
in
al
p
r
ed
ictio
n
an
d
ev
a
lu
at
io
n
.
T
h
e
wo
r
k
f
lo
w
b
eg
in
s
with
d
ataset
ac
q
u
is
itio
n
an
d
p
r
ep
r
o
ce
s
s
in
g
,
e
n
s
u
r
in
g
th
e
d
ata
is
clea
n
a
n
d
ap
p
r
o
p
r
i
ately
s
tr
u
ctu
r
ed
f
o
r
an
aly
s
is
.
T
h
e
s
tu
d
y
f
o
cu
s
es
o
n
th
e
id
en
tific
atio
n
an
d
ev
alu
atio
n
o
f
k
ey
f
ea
tu
r
es
in
f
lu
en
cin
g
lear
n
er
p
er
f
o
r
m
an
ce
,
em
p
l
o
y
in
g
s
tat
is
tical
tech
n
iq
u
es
an
d
ML
a
lg
o
r
ith
m
s
.
T
o
en
s
u
r
e
r
o
b
u
s
t
m
o
d
el
ev
alu
atio
n
,
s
tr
atif
ied
1
0
-
f
o
l
d
cr
o
s
s
-
v
alid
at
io
n
was
ap
p
lied
,
an
d
h
y
p
er
p
ar
am
eter
tu
n
in
g
was
co
n
d
u
cted
to
o
p
tim
ize
m
o
d
el
p
er
f
o
r
m
an
ce
.
T
h
e
en
s
em
b
le
m
o
d
el,
b
u
ilt
u
s
in
g
a
v
o
tin
g
te
ch
n
iq
u
e,
i
n
teg
r
ates
p
r
e
d
ictio
n
s
f
r
o
m
m
u
ltip
le
b
ase
lear
n
er
s
,
aim
in
g
to
im
p
r
o
v
e
a
cc
u
r
ac
y
an
d
r
eliab
ilit
y
.
W
e
p
r
o
v
id
e
a
d
etailed
e
x
p
lan
atio
n
o
f
ea
ch
p
h
ase
o
f
th
e
m
eth
o
d
o
l
o
g
y
,
e
n
s
u
r
in
g
th
at
ev
er
y
asp
ec
t
is
th
o
r
o
u
g
h
ly
c
o
v
er
ed
to
f
ac
ilit
ate
r
ep
licatio
n
a
n
d
v
alid
atio
n
o
f
o
u
r
a
p
p
r
o
ac
h
.
Fig
u
r
e
1
.
Pro
ce
s
s
f
lo
wch
a
r
t
2
.
2
.
Da
t
a
Co
llect
io
n
T
h
e
d
ataset
u
tili
ze
d
in
th
is
s
tu
d
y
was
co
llected
f
r
o
m
th
e
G
o
v
er
n
m
en
t
Po
ly
tech
n
ic
o
f
Kar
n
atak
a
s
tate
an
d
en
co
m
p
ass
es
a
co
m
p
r
e
h
en
s
iv
e
s
et
o
f
f
ea
tu
r
es
ai
m
ed
at
ca
p
tu
r
i
n
g
v
a
r
io
u
s
asp
ec
ts
o
f
lear
n
er
s
'
d
em
o
g
r
a
p
h
ics,
b
eh
a
v
io
r
,
an
d
p
ar
ticip
atio
n
o
n
th
e
o
n
lin
e
p
latf
o
r
m
.
A
s
tr
atif
ied
s
am
p
lin
g
m
eth
o
d
was
em
p
lo
y
ed
t
o
en
s
u
r
e
d
iv
er
s
e
r
ep
r
esen
tatio
n
ac
r
o
s
s
d
if
f
er
en
t
p
er
f
o
r
m
an
ce
lev
els,
in
co
r
p
o
r
atin
g
d
ata
co
llected
f
r
o
m
b
o
th
L
MS
an
d
o
f
f
lin
e
s
o
u
r
ce
s
,
r
esu
ltin
g
i
n
a
co
m
p
r
e
h
en
s
iv
e
d
ataset
co
m
p
r
is
in
g
1
,
0
0
2
s
tu
d
e
n
t
r
ec
o
r
d
s
with
2
3
d
is
tin
ct
ch
ar
ac
ter
is
tics
.
T
h
is
r
esear
ch
ad
o
p
ts
a
cr
o
s
s
-
s
ec
tio
n
al
ap
p
r
o
ac
h
,
u
tili
zin
g
d
ata
co
llected
o
v
er
a
s
in
g
le
ac
ad
em
ic
y
ea
r
to
b
u
il
d
an
d
v
alid
ate
p
r
e
d
ictiv
e
m
o
d
els.
T
h
is
r
ig
o
r
o
u
s
d
ata
co
llecti
o
n
p
r
o
ce
s
s
p
r
o
v
id
es
a
s
o
lid
f
o
u
n
d
atio
n
f
o
r
t
h
e
s
u
b
s
eq
u
en
t a
n
al
y
s
is
.
T
h
e
f
ea
tu
r
es
as in
T
ab
le
2
wer
e
c
o
n
s
id
er
ed
f
o
r
an
al
y
s
is
.
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
7
,
No
.
3
,
Ma
r
ch
20
2
5
:
2
0
3
2
-
20
4
3
2036
T
ab
le
2
.
Data
s
et
d
escr
ip
tio
n
A
t
t
r
i
b
u
t
e
D
e
scri
p
t
i
o
n
S
S
LC
m
e
d
i
u
m
o
f
st
u
d
y
M
e
d
i
u
m
o
f
s
t
u
d
y
i
n
t
h
e
S
e
c
o
n
d
a
r
y
S
c
h
o
o
l
.
W
e
e
k
l
y
l
o
g
i
n
f
r
e
q
u
e
n
c
y
Th
e
n
u
m
b
e
r
o
f
t
i
m
e
s a
st
u
d
e
n
t
l
o
g
s
i
n
p
e
r
w
e
e
k
.
D
u
r
a
t
i
o
n
o
f
o
n
l
i
n
e
l
e
c
t
u
r
e
v
i
e
w
i
n
g
Th
e
a
v
e
r
a
g
e
t
i
me
sp
e
n
t
w
a
t
c
h
i
n
g
o
n
l
i
n
e
l
e
c
t
u
r
e
s.
Ti
me
S
p
e
n
t
o
n
o
n
l
i
n
e
a
ssi
g
n
me
n
t
s
To
t
a
l
t
i
me
sp
e
n
t
o
n
c
o
m
p
l
e
t
i
n
g
o
n
l
i
n
e
a
ssi
g
n
me
n
t
s
.
B
a
si
c
c
o
m
p
u
t
e
r
sk
i
l
l
s
P
r
o
f
i
c
i
e
n
c
y
i
n
b
a
si
c
c
o
m
p
u
t
e
r
s
k
i
l
l
s.
P
h
y
s
i
c
a
l
a
c
t
i
v
i
t
y
F
r
e
q
u
e
n
c
y
o
f
e
n
g
a
g
e
me
n
t
i
n
p
h
y
s
i
c
a
l
a
c
t
i
v
i
t
i
e
s
.
R
e
si
d
i
n
g
r
e
g
i
o
n
t
y
p
e
Ty
p
e
o
f
r
e
g
i
o
n
w
h
e
r
e
t
h
e
st
u
d
e
n
t
r
e
s
i
d
e
s.
F
a
mi
l
y
y
e
a
r
l
y
i
n
c
o
me
A
n
n
u
a
l
i
n
c
o
m
e
o
f
t
h
e
st
u
d
e
n
t
's
f
a
mi
l
y
.
D
i
e
t
D
i
e
t
a
r
y
h
a
b
i
t
s
o
f
t
h
e
s
t
u
d
e
n
t
s.
S
l
e
e
p
d
u
r
a
t
i
o
n
A
v
e
r
a
g
e
d
u
r
a
t
i
o
n
o
f
sl
e
e
p
p
e
r
d
a
y
.
Li
b
r
a
r
y
v
i
s
i
t
s
F
r
e
q
u
e
n
c
y
o
f
v
i
si
t
s
t
o
t
h
e
l
i
b
r
a
r
y
.
S
martp
h
o
n
e
u
sa
g
e
f
o
r
s
t
u
d
y
p
u
r
p
o
ses
F
r
e
q
u
e
n
c
y
o
f
u
s
i
n
g
a
s
mart
p
h
o
n
e
f
o
r
st
u
d
y
p
u
r
p
o
s
e
s.
P
a
r
t
i
c
i
p
a
t
i
o
n
i
n
t
e
a
m a
c
t
i
v
i
t
i
e
s
o
n
l
i
n
e
I
n
v
o
l
v
e
m
e
n
t
i
n
c
o
l
l
a
b
o
r
a
t
i
v
e
t
e
a
m a
c
t
i
v
i
t
i
e
s
o
n
l
i
n
e
.
V
i
d
e
o
p
a
u
s
e
f
r
e
q
u
e
n
c
y
Th
e
n
u
m
b
e
r
o
f
t
i
m
e
s a
st
u
d
e
n
t
p
a
u
s
e
s
v
i
d
e
o
w
h
i
l
e
w
a
t
c
h
i
n
g
a
l
e
c
t
u
r
e
.
F
r
e
q
u
e
n
c
y
o
f
v
i
d
e
o
r
e
p
e
t
i
t
i
o
n
H
o
w
o
f
t
e
n
s
t
u
d
e
n
t
s re
p
e
a
t
t
h
e
v
i
d
e
o
l
e
c
t
u
r
e
.
A
c
t
i
v
i
t
i
e
s
c
o
m
p
l
e
t
e
d
P
r
o
p
o
r
t
i
o
n
o
f
a
ss
i
g
n
e
d
a
c
t
i
v
i
t
i
e
s
c
o
m
p
l
e
t
e
d
.
Le
c
t
u
r
e
s
w
a
t
c
h
e
d
Th
e
p
e
r
c
e
n
t
a
g
e
o
f
t
o
t
a
l
l
e
c
t
u
r
e
s wa
t
c
h
e
d
.
A
c
c
e
ss
t
o
in
t
e
r
n
e
t
A
v
a
i
l
a
b
i
l
i
t
y
o
f
i
n
t
e
r
n
e
t
c
o
n
n
e
c
t
i
v
i
t
y
.
Ea
se
o
f
o
n
l
i
n
e
c
o
mm
u
n
i
c
a
t
i
o
n
w
i
t
h
t
e
a
c
h
e
r
s
P
e
r
c
e
i
v
e
d
e
a
s
e
o
f
c
o
mm
u
n
i
c
a
t
i
o
n
w
i
t
h
t
e
a
c
h
e
r
s
o
n
l
i
n
e
.
Lo
g
i
n
t
o
po
r
t
a
l
D
o
e
s s
t
u
d
e
n
t
s
l
o
g
i
n
t
o
p
o
r
t
a
l
d
a
i
l
y
.
En
g
a
g
e
m
e
n
t
i
n
o
n
l
i
n
e
f
o
r
u
ms
A
c
t
i
v
e
p
a
r
t
i
c
i
p
a
t
i
o
n
i
n
o
n
l
i
n
e
d
i
s
c
u
s
si
o
n
f
o
r
u
ms.
P
e
r
c
e
i
v
e
d
Ea
s
e
o
f
o
n
l
i
n
e
l
e
a
r
n
i
n
g
S
t
u
d
e
n
t
p
e
r
c
e
p
t
i
o
n
o
f
t
h
e
e
a
s
e
o
f
o
n
l
i
n
e
l
e
a
r
n
i
n
g
.
I
n
t
e
r
a
c
t
i
o
n
w
i
t
h
p
e
e
r
s
i
n
o
n
l
i
n
e
f
o
r
u
m
s
Le
v
e
l
o
f
e
n
g
a
g
e
me
n
t
w
i
t
h
p
e
e
r
s
i
n
o
n
l
i
n
e
f
o
r
u
ms
.
2
.
3
.
Da
t
a
p
re
pro
ce
s
s
ing
Data
p
r
ep
r
o
ce
s
s
in
g
is
a
p
iv
o
tal
p
h
ase
in
th
e
ML
p
ip
elin
e,
cr
u
cial
f
o
r
en
s
u
r
in
g
th
e
d
ataset's
q
u
ality
an
d
r
eliab
ilit
y
b
e
f
o
r
e
m
o
d
el
d
ev
elo
p
m
en
t.
I
n
th
is
s
tu
d
y
,
a
r
i
g
o
r
o
u
s
p
r
ep
r
o
ce
s
s
in
g
ap
p
r
o
ac
h
was
u
n
d
er
tak
e
n
to
ad
d
r
ess
ch
allen
g
es
s
u
ch
as
d
at
a
in
co
n
s
is
ten
cy
,
m
is
s
in
g
v
alu
e
s
,
an
d
class
im
b
alan
ce
s
.
T
o
h
a
n
d
le
m
is
s
in
g
d
ata,
eig
h
t
r
ec
o
r
d
s
with
s
ig
n
if
ican
t
m
is
s
in
g
v
alu
es
wer
e
eli
m
in
a
ted
,
wh
ile
f
o
u
r
r
ec
o
r
d
s
with
p
ar
tial
m
is
s
in
g
d
ata
wer
e
co
r
r
ec
ted
u
s
in
g
m
ea
n
im
p
u
tatio
n
.
Fo
r
ca
teg
o
r
ical
v
ar
iab
les,
b
o
th
o
r
d
in
al
an
d
lab
el
en
c
o
d
in
g
tech
n
iq
u
es
wer
e
em
p
lo
y
e
d
.
Or
d
in
al
en
co
d
in
g
was
ap
p
lied
to
f
ea
tu
r
es
with
an
in
h
er
e
n
t
o
r
d
er
,
s
u
c
h
as
'
Fam
il
y
Yea
r
ly
I
n
co
m
e,
'
'
Acc
ess
to
I
n
ter
n
et,
'
an
d
'
Per
ce
iv
ed
E
ase
o
f
On
l
in
e
L
ea
r
n
in
g
,
'
th
er
eb
y
p
r
ese
r
v
in
g
t
h
e
n
at
u
r
al
r
a
n
k
in
g
o
f
th
ese
ca
teg
o
r
ies.
C
o
n
v
er
s
ely
,
lab
el
en
co
d
in
g
w
as
u
s
ed
f
o
r
n
o
m
in
al
v
ar
ia
b
les
with
o
u
t
in
tr
in
s
ic
o
r
d
er
,
in
cl
u
d
in
g
'
SS
L
C
Me
d
iu
m
o
f
Stu
d
y
,
'
'
R
e
s
id
in
g
R
eg
io
n
T
y
p
e,
'
'B
asic
C
o
m
p
u
ter
Sk
ills
,
'
an
d
'
Diet.
'
Ad
d
itio
n
ally
,
th
e
tar
g
et
v
ar
iab
le
'
Gr
ad
e'
was o
r
d
in
ally
en
c
o
d
ed
t
o
f
ac
il
itate
th
e
p
r
ed
ictio
n
task
b
y
tr
a
n
s
f
o
r
m
in
g
class
lab
els in
to
n
u
m
er
ical
v
alu
es.
T
o
ad
d
r
ess
class
im
b
alan
ce
,
p
ar
ticu
lar
ly
wi
th
in
th
e
'
Gr
ad
e'
clas
s
es,
r
an
d
o
m
o
v
er
s
am
p
lin
g
was
im
p
lem
en
ted
.
T
h
is
m
eth
o
d
i
n
v
o
lv
e
d
r
ep
licatin
g
i
n
s
tan
ce
s
f
r
o
m
t
h
e
m
in
o
r
ity
class
es
t
o
eq
u
alize
th
e
class
d
is
tr
ib
u
tio
n
,
en
s
u
r
in
g
ea
c
h
cl
ass
co
n
tain
ed
3
9
8
in
s
tan
ce
s
.
T
h
is
s
tep
was
cr
itical
in
p
r
e
v
en
tin
g
m
o
d
el
b
ias
to
war
d
s
th
e
m
ajo
r
ity
class
an
d
p
r
o
m
o
tin
g
f
air
n
ess
in
th
e
tr
ain
in
g
p
r
o
ce
s
s
.
T
h
e
f
in
al
p
r
e
-
p
r
o
ce
s
s
ed
d
ataset,
in
co
r
p
o
r
atin
g
b
o
th
o
r
d
in
al
an
d
lab
el
en
co
d
in
g
,
p
r
o
v
id
ed
a
r
o
b
u
s
t f
o
u
n
d
atio
n
f
o
r
b
u
ild
i
n
g
a
cc
u
r
ate
an
d
r
eliab
le
ML
m
o
d
els.
2
.
4
.
Ana
ly
s
is
2
.
4
.
1
.
P
a
ra
m
et
er
weig
ht
a
g
e
c
a
lcula
t
io
n
I
n
th
is
r
esear
ch
,
th
e
R
F
alg
o
r
i
th
m
was
em
p
lo
y
ed
to
ass
ig
n
weig
h
ts
to
th
e
2
3
f
ea
tu
r
es
co
n
s
id
er
ed
f
o
r
p
r
ed
ictin
g
lear
n
er
p
e
r
f
o
r
m
an
c
e.
R
F
was
ch
o
s
en
f
o
r
its
s
tr
o
n
g
p
r
e
d
ictiv
e
ca
p
ab
ilit
ies
an
d
its
ab
ilit
y
to
ev
alu
ate
th
e
r
ele
v
an
ce
o
f
ea
ch
f
ea
tu
r
e
in
r
elatio
n
to
t
h
e
tar
g
et
v
ar
i
ab
le,
wh
ich
was
th
e
lear
n
e
r
s
'
‘G
r
ad
e
’
.
T
h
is
d
u
al
f
u
n
ctio
n
ality
o
f
p
r
ed
ictio
n
a
n
d
f
ea
tu
r
e
im
p
o
r
tan
ce
m
ad
e
it a
v
alu
ab
le
to
o
l i
n
th
e
s
tu
d
y
.
-
T
r
ain
in
g
R
F
cla
s
s
if
ier
:
T
h
e
p
r
o
ce
s
s
b
eg
an
with
th
e
co
n
s
tr
u
ctio
n
o
f
a
R
F
m
o
d
el
u
s
in
g
th
e
p
r
e
-
p
r
o
ce
s
s
ed
d
ataset.
T
h
e
m
o
d
el
co
m
p
r
is
es
m
u
ltip
le
DT
,
ea
c
h
tr
ai
n
ed
o
n
a
d
if
f
er
e
n
t
s
u
b
s
et
o
f
th
e
d
ata
th
r
o
u
g
h
b
o
o
ts
tr
ap
p
in
g
;
wh
ile
r
an
d
o
m
l
y
s
elec
tin
g
at
ea
ch
s
p
lit,
a
s
u
b
s
et
o
f
f
ea
tu
r
es
is
s
e
lecte
d
to
p
r
o
m
o
te
d
iv
er
s
ity
am
o
n
g
th
e
tr
ee
s
.
T
h
is
ap
p
r
o
ac
h
r
ed
u
ce
s
th
e
lik
elih
o
o
d
o
f
o
v
er
f
itti
n
g
an
d
in
cr
ea
s
es
th
e
g
en
er
aliza
b
ilit
y
o
f
th
e
m
o
d
el.
Giv
en
a
d
ataset
=
{(
₁
,
₁)
,
(
₂
,
₂
)
,
…,
(
n
,
n
)}
r
ep
r
esen
ted
in
s
tan
ce
s
wh
er
e
ᵢ
d
e
n
o
ted
f
ea
tu
r
e
v
ec
to
r
s
an
d
ᵢ
s
ig
n
if
ie
d
co
r
r
esp
o
n
d
in
g
lab
els.,
we
f
ir
s
t c
o
n
s
tr
u
ct
a
R
F m
o
d
el
F.
-
Featu
r
e
im
p
o
r
tan
ce
ca
lcu
latio
n
: Fo
r
ea
ch
f
ea
tu
r
e
ⱼ
(
f
o
r
j=1
,
2
….
2
3
)
,
its
im
p
o
r
tan
ce
s
co
r
e
ⱼ
d
eter
m
in
ed
b
ased
o
n
th
e
r
ed
u
ctio
n
in
im
p
u
r
ity
it c
o
n
tr
i
b
u
tes ac
r
o
s
s
al
l tr
ee
s
in
th
e
f
o
r
est as in
(
1
)
.
T
h
e
f
ea
tu
r
e
im
p
o
r
tan
ce
s
co
r
e
o
f
t
h
e
in
d
iv
i
d
u
al
p
ar
am
ete
r
s
is
as sh
o
wn
in
Fig
u
r
e
2
.
=
1
∑
∑
(
)
∈
=
1
(
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:
2
5
0
2
-
4
7
52
Mu
ltif
a
ce
ted
a
p
p
r
o
a
c
h
fo
r
a
n
t
icip
a
tin
g
lea
r
n
er p
erfo
r
ma
n
ce
…
(
S
h
a
b
n
a
m
A
r
a
S
Ja
h
a
g
ir
d
a
r
)
2037
W
h
er
e
Δ
I
m
p
u
r
ity
(
t)
is
th
e
d
ec
r
ea
s
e
in
im
p
u
r
ity
at
n
o
d
e
t w
h
en
s
p
litt
in
g
o
n
f
ea
tu
r
e
x
j
an
d
B
is
th
e
n
u
m
b
er
o
f
tr
ee
s
in
th
e
f
o
r
est.
-
Ass
ig
n
in
g
W
eig
h
ts
:
T
h
e
co
m
p
u
ted
weig
h
ts
w
j
ar
e
s
u
b
s
eq
u
en
tly
in
co
r
p
o
r
ate
d
in
to
th
e
e
n
s
em
b
le
m
o
d
el,
wh
er
e
th
ey
in
f
lu
e
n
ce
th
e
p
r
e
d
i
ctio
n
p
r
o
ce
s
s
b
y
p
r
i
o
r
itizin
g
m
o
r
e
in
f
lu
e
n
tial f
ea
tu
r
es.
T
h
e
weig
h
ted
f
ea
tu
r
e
v
ec
to
r
x
w
f
o
r
an
o
b
s
er
v
atio
n
is
d
ef
in
ed
as
in
(
2
)
.
T
h
e
f
in
al
p
r
ed
ictiv
e
m
o
d
el
th
e
n
o
p
er
ates
o
n
th
ese
weig
h
ted
f
ea
tu
r
es,
e
n
h
an
cin
g
b
o
th
ac
cu
r
ac
y
an
d
in
ter
p
r
etab
i
lity
b
y
ac
co
u
n
tin
g
f
o
r
t
h
e
v
ar
y
in
g
im
p
o
r
tan
ce
o
f
ea
ch
f
ea
tu
r
e.
x
w
= {w
1
⋅
x
1
,
w
2
⋅
x
2
, …, w
p
⋅
x
p
}
(
2
)
W
h
er
e
p
is
th
e
n
u
m
b
er
o
f
p
ar
a
m
eter
s
.
Fig
u
r
e
2
.
Featu
r
e
I
m
p
o
r
tan
ce
Sco
r
e
2
.
4
.
2
.
H
y
brid
ens
em
ble m
o
d
el
I
n
th
is
s
tu
d
y
,
we
im
p
le
m
en
te
d
a
m
u
lti
-
lev
el
en
s
em
b
le
lear
n
in
g
ap
p
r
o
ac
h
as
d
ep
icted
in
Fig
u
r
e
3
,
wh
er
e
b
ase
lear
n
er
s
wer
e
f
ir
s
t
tr
ain
ed
in
d
ep
e
n
d
en
tly
,
an
d
th
e
ir
p
r
ed
ictio
n
s
wer
e
s
u
b
s
eq
u
en
t
ly
co
m
b
in
ed
u
s
in
g
a
m
eta
-
lear
n
er
.
T
h
is
h
ier
ar
ch
i
ca
l
m
o
d
el
s
tr
u
ctu
r
e
is
d
esig
n
e
d
to
lev
er
ag
e
th
e
s
tr
en
g
th
s
o
f
d
if
f
er
en
t
alg
o
r
ith
m
s
,
r
ed
u
cin
g
th
e
v
a
r
ian
ce
an
d
b
ia
s
in
h
er
en
t
in
in
d
iv
id
u
al
m
o
d
e
ls
,
an
d
th
er
eb
y
im
p
r
o
v
in
g
th
e
o
v
er
all
p
r
ed
ictiv
e
p
er
f
o
r
m
an
ce
.
T
h
e
en
s
em
b
le
p
r
ed
ictio
n
p
r
o
ce
s
s
b
eg
a
n
b
y
s
elec
tin
g
a
d
i
v
er
s
e
s
et
o
f
b
ase
lear
n
er
s
,
ea
ch
o
f
wh
ich
ca
p
tu
r
es
d
if
f
e
r
en
t
asp
e
cts
o
f
th
e
d
ata
d
is
tr
ib
u
tio
n
.
T
h
e
b
ase
lear
n
er
s
u
s
ed
in
th
is
s
tu
d
y
in
clu
d
ed
K
-
n
e
a
re
st n
e
ig
h
b
o
rs
(
KNN)
,
R
F,
SVM,
g
r
ad
ien
t b
o
o
s
tin
g
(
GB
)
,
DT
,
an
d
eXtr
em
e
g
r
ad
ien
t
b
o
o
s
t
(
XGB).
Fig
u
r
e
3
.
E
n
s
em
b
le
m
o
d
el
with
v
o
tin
g
ap
p
r
o
ac
h
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
7
,
No
.
3
,
Ma
r
ch
20
2
5
:
2
0
3
2
-
20
4
3
2038
As
o
u
tlin
ed
in
Alg
o
r
ith
m
1
,
we
s
elec
ted
s
ix
d
iv
er
s
e
b
ase
lear
n
er
s
to
ca
p
tu
r
e
d
if
f
er
en
t
as
p
ec
ts
o
f
th
e
d
ata
d
is
tr
ib
u
tio
n
.
E
ac
h
b
ase
lear
n
er
was
tr
ain
ed
o
n
th
e
tr
ain
in
g
d
ataset
D
train
an
d
o
p
tim
ized
th
r
o
u
g
h
h
y
p
er
p
ar
am
eter
t
u
n
in
g
u
s
in
g
cr
o
s
s
-
v
alid
atio
n
to
m
ax
im
iz
e
in
d
iv
id
u
al
p
er
f
o
r
m
an
ce
.
T
h
e
p
r
ed
ictio
n
s
f
r
o
m
ea
ch
b
ase
lear
n
e
r
o
n
a
h
o
ld
-
o
u
t
v
alid
atio
n
s
et
D
val
wer
e
s
to
r
ed
,
f
o
r
m
in
g
th
e
b
asis
o
f
a
n
ew
d
ataset
D
meta
,
wh
er
e
ea
ch
p
r
ed
ictio
n
v
ec
to
r
f
r
o
m
th
e
b
ase
lear
n
er
s
b
ec
o
m
es a
f
ea
tu
r
e.
Alg
o
r
ith
m
1
.
B
ase
lear
n
er
tr
ai
n
in
g
Fo
r
ea
ch
ba
se
le
ar
ne
r
L
i
where
i
ra
ng
es
fr
om
1
to
6
(c
or
re
sp
on
di
ng
to
R
F,
SV
M,
KN
N,
GB
,
XGB, and DT):
-
Train
L
i
on the training dataset
D
train
.
-
Optimize
L
i
by
tu
ni
ng
it
s
hy
pe
r
parameters
θ
i
us
in
g
cr
os
s
-
validation
to
maxi
mize
its
individual performance.
-
St
or
e
th
e
tr
ai
ne
d
mo
de
l
M
i
al
on
g
wi
th
it
s
pr
ed
ic
t
io
ns
on
a
ho
ld
-
ou
t
va
li
d
at
io
n
se
t
D
val.
Collect
predictions
from
all
base
learners
on
D
val
to
cr
ea
te
a
ne
w
da
ta
se
t
D
meta
,
wh
er
e
ea
ch
prediction vector from the base learners becomes a feature in
D
meta
.
T
h
e
s
ec
o
n
d
le
v
el
o
f
th
e
en
s
em
b
le
s
tr
u
ctu
r
e
,
d
escr
ib
ed
in
Alg
o
r
ith
m
2
,
in
v
o
lv
es
a
m
e
ta
-
lear
n
er
,
wh
ich
is
tr
ain
ed
o
n
D
meta
to
le
ar
n
th
e
o
p
tim
al
co
m
b
in
atio
n
o
f
th
e
b
ase
lear
n
er
s
’
p
r
e
d
ictio
n
s
.
I
n
th
is
s
tu
d
y
,
we
ex
p
lo
r
ed
m
u
ltip
le
alg
o
r
ith
m
s
as
m
eta
-
lear
n
er
s
,
in
clu
d
in
g
R
F,
SVM,
K
NN,
G
B
,
XG
B
,
an
d
DT
,
to
d
eter
m
in
e
wh
ich
co
m
b
i
n
atio
n
wo
u
ld
y
iel
d
th
e
b
est p
r
e
d
ictiv
e
p
er
f
o
r
m
a
n
ce
.
Alg
o
r
ith
m
2
.
Me
ta
-
lear
n
er
t
r
ain
in
g
Construct
the
meta
da
ta
set
D
meta
as
de
sc
ri
be
d
in
al
go
ri
th
m
1,
wi
th
e
ac
h
ob
se
rv
at
io
n
representing the output from all base learners for a given input.
For each meta learner
M
meta
(corresponding to RF, SVM, KNN, GB, XGB, DT):
-
Train
M
meta
on
D
meta
using the selected algorithm.
-
Th
e
me
ta
le
ar
ne
r
le
ar
ns
a
fu
nc
ti
on
f
meta
(x)
=
∑
6
=
1
L
i
(x)
,
where
α
i
ar
e
th
e
coefficients (weights) assigned to each base learner’s prediction
(
)
.
-
Optimize
α
i
to
mi
ni
mi
ze
pr
ed
ic
ti
on
er
ro
r,
ef
f
ec
ti
ve
ly
fu
si
ng
th
e
ba
se
learners' predictions in a way that maximizes overall accuracy.
Hy
p
er
p
ar
a
m
eter
tu
n
in
g
was
s
y
s
tem
atica
lly
ap
p
lied
to
b
o
t
h
t
h
e
b
ase
lear
n
e
r
s
an
d
th
e
m
eta
lear
n
er
s
to
en
s
u
r
e
o
p
tim
al
p
er
f
o
r
m
a
n
ce
a
cr
o
s
s
th
e
en
s
em
b
le.
Du
r
in
g
t
h
e
h
y
p
e
r
p
ar
am
eter
t
u
n
in
g
f
o
r
th
e
b
ase
lear
n
er
a
g
r
id
s
ea
r
ch
co
m
b
in
ed
with
k
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
was
em
p
lo
y
ed
f
o
r
ea
ch
b
ase
lear
n
e
r
to
ex
p
lo
r
e
a
r
a
n
g
e
o
f
h
y
p
er
p
ar
am
eter
v
alu
es,
aim
in
g
to
m
in
im
ize
cr
o
s
s
-
v
alid
atio
n
er
r
o
r
an
d
e
n
s
u
r
e
t
h
at
ea
ch
m
o
d
el
was
n
eith
er
o
v
er
f
itti
n
g
n
o
r
u
n
d
er
f
itti
n
g
t
h
e
d
ata.
W
h
er
e
as
in
t
h
e
ca
s
e
o
f
m
eta
lear
n
er
s
,
h
y
p
er
p
a
r
am
eter
s
wer
e
tu
n
ed
t
o
o
p
tim
ize
th
eir
ab
ilit
y
to
co
m
b
in
e
th
e
b
ase
lear
n
er
s
'
o
u
tp
u
ts
ef
f
ec
tiv
ely
.
T
h
is
in
clu
d
ed
tu
n
i
n
g
p
ar
am
ete
r
s
s
u
ch
as
th
e
d
e
p
th
o
f
t
r
ee
s
in
R
F
an
d
DT
,
th
e
m
a
r
g
in
p
ar
a
m
eter
‘
C
’
in
SVM,
an
d
th
e
lear
n
i
n
g
r
ate
in
GB
an
d
XGB.
T
h
e
s
p
ec
if
ic
h
y
p
er
p
ar
am
eter
s
u
s
ed
f
o
r
ea
ch
class
if
ier
ar
e
d
etailed
in
T
ab
le
3
,
r
ef
lectin
g
th
e
ca
r
ef
u
l
tu
n
i
n
g
p
r
o
ce
s
s
u
n
d
er
ta
k
en
to
ac
h
iev
e
o
p
tim
al
m
o
d
el
p
er
f
o
r
m
an
ce
.
T
ab
le
3
.
Hy
p
er
p
ar
a
m
eter
c
o
n
f
ig
u
r
atio
n
C
l
a
s
si
f
i
e
r
s
H
y
p
e
r
p
a
r
a
me
t
e
r
s
RF
(
n
_
e
st
i
ma
t
o
r
s=
1
0
0
,
m
a
x
_
d
e
p
t
h
=
2
0
,
m
i
n
_
sam
p
l
e
s
_
l
e
a
f
=
1
,
mi
n
_
s
a
m
p
l
e
s_
s
p
l
i
t
=
2
,
r
a
n
d
o
m_
s
t
a
t
e
=
4
2
,
m
a
x
_
f
e
a
t
u
r
e
s='s
q
r
t
'
,
c
r
i
t
e
r
i
o
n
=
'
g
i
n
i
')
VM
(
C
=
5
0
0
,
k
e
r
n
e
l
=
'r
b
f
',
g
a
mm
a
=
1
0
0
,
c
l
a
ss_
w
e
i
g
h
t
=
'b
a
l
a
n
c
e
d
',
p
r
o
b
a
b
i
l
i
t
y
=
Tr
u
e
)
DT
(
max
_
d
e
p
t
h
=
2
0
,
m
i
n
_
sam
p
l
e
s
_
sp
l
i
t
=
2
,
m
i
n
_
s
a
m
p
l
e
s
_
l
e
a
f
=
1
,
c
r
i
t
e
r
i
o
n
=
'g
i
n
i
',
r
a
n
d
o
m
_
s
t
a
t
e
=
4
2
)
K
N
N
(
l
e
a
f
_
si
z
e
=
1
0
,
n
_
n
e
i
g
h
b
o
r
s=
3
,
w
e
i
g
h
t
s='d
i
st
a
n
c
e
')
GB
(
max
_
d
e
p
t
h
=
4
,
n
_
e
s
t
i
m
a
t
o
r
s=
2
0
0
,
su
b
sam
p
l
e
=
0
.
8
)
XGB
(
max
_
d
e
p
t
h
=
1
0
,
c
o
l
sam
p
l
e
_
b
y
t
r
e
e
=
0
.
6
,
g
a
mm
a
=
0
.
1
,
l
e
a
r
n
i
n
g
_
r
a
t
e
=
0
.
1
,
n
_
e
st
i
m
a
t
o
r
s=
2
0
0
,
su
b
s
a
m
p
l
e
=
1
.
0
)
M
e
t
a
-
Le
a
r
n
e
r
(
e
st
i
m
a
t
o
r
s=
b
a
s
e
_
l
e
a
r
n
e
r
s,
v
o
t
i
n
g
=
'
h
a
r
d
')
{b
a
s
e
_
l
e
a
r
n
e
r
s=
(
R
F
,
S
V
M
,
D
T
,
K
N
N
,
G
B
,
X
G
B
)
,
met
a
_
l
e
a
r
n
e
r
=
(
b
a
s
e
_
mo
d
e
l
)
}
2
.
5
.
P
er
f
o
rma
nce
e
v
a
lua
t
i
o
n
T
h
e
ef
f
ec
tiv
e
n
ess
o
f
th
e
f
u
s
ed
en
s
em
b
le
m
o
d
el
was
ass
ess
e
d
b
y
c
o
m
p
a
r
in
g
its
p
er
f
o
r
m
a
n
ce
ag
ain
s
t
in
d
iv
id
u
al
b
ase
lear
n
er
s
an
d
s
im
p
ler
en
s
em
b
le
m
eth
o
d
s
.
E
v
alu
atio
n
m
etr
ics,
in
cl
u
d
in
g
a
cc
u
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
s
co
r
e
as
d
e
f
in
e
d
in
(
3
)
,
(
4
)
,
(
5
)
,
an
d
(
6
)
,
we
r
e
u
s
ed
to
p
r
o
v
i
d
e
a
th
o
r
o
u
g
h
ass
es
s
m
en
t
o
f
th
e
m
o
d
el'
s
p
er
f
o
r
m
a
n
ce
.
E
ac
h
c
lass
if
ier
was
tr
ain
ed
,
test
ed
o
n
th
e
test
s
et,
an
d
ev
alu
ate
d
u
s
in
g
t
h
ese
k
ey
p
er
f
o
r
m
an
ce
m
etr
ics.
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
Mu
ltif
a
ce
ted
a
p
p
r
o
a
c
h
fo
r
a
n
t
icip
a
tin
g
lea
r
n
er p
erfo
r
ma
n
ce
…
(
S
h
a
b
n
a
m
A
r
a
S
Ja
h
a
g
ir
d
a
r
)
2039
A
cc
u
r
a
cy
=
+
(
3
)
P
r
ec
is
io
n
=
+
(
4
)
R
ec
a
ll=
+
(
5
)
F1
-
S
co
r
e
=
2
×
+
(
6
)
2
.
6
.
I
m
ple
m
ent
a
t
io
n
d
et
a
ils
T
h
e
ex
p
er
im
en
ts
f
o
r
th
is
r
ese
ar
ch
wer
e
co
n
d
u
cte
d
u
s
in
g
G
o
o
g
le
C
o
lab
,
a
n
in
te
r
ac
tiv
e
e
n
v
ir
o
n
m
e
n
t
th
at
p
r
o
v
id
es
ac
ce
s
s
to
p
o
we
r
f
u
l
co
m
p
u
tatio
n
al
r
eso
u
r
ce
s
.
T
h
e
im
p
lem
e
n
tatio
n
was
ca
r
r
ied
o
u
t
in
Py
th
o
n
3
.
1
0
.
1
2
(
6
4
-
b
it),
lev
e
r
ag
in
g
a
r
an
g
e
o
f
well
-
estab
lis
h
ed
ML
lib
r
ar
ies,
in
clu
d
in
g
s
cik
it
-
lear
n
,
T
en
s
o
r
Flo
w,
an
d
XGBo
o
s
t.
T
h
ese
lib
r
ar
ies
wer
e
u
tili
ze
d
f
o
r
task
s
s
u
ch
as
d
a
ta
p
r
ep
r
o
ce
s
s
i
n
g
,
m
o
d
el
t
r
ain
i
n
g
,
h
y
p
er
p
ar
am
eter
tu
n
in
g
,
an
d
ev
alu
atio
n
.
T
h
e
clo
u
d
-
b
ased
en
v
ir
o
n
m
en
t
o
f
C
o
lab
f
ac
ilit
ated
s
ea
m
less
ex
ec
u
tio
n
o
f
co
m
p
u
tatio
n
ally
in
ten
s
iv
e
p
r
o
ce
s
s
es,
allo
win
g
f
o
r
ef
f
icien
t
h
an
d
lin
g
o
f
lar
g
e
d
atasets
an
d
co
m
p
lex
m
o
d
els.
Fu
r
th
er
m
o
r
e
,
th
e
in
teg
r
atio
n
o
f
C
o
lab
with
Go
o
g
le
Dr
iv
e
en
s
u
r
ed
th
e
s
ec
u
r
e
s
to
r
ag
e
an
d
ea
s
y
ac
ce
s
s
o
f
ex
p
er
im
en
tal
d
ata
,
en
ab
lin
g
r
e
p
r
o
d
u
cib
ilit
y
an
d
c
o
llab
o
r
ativ
e
r
esear
ch
ef
f
o
r
ts
.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
r
esu
lts
o
f
th
is
s
tu
d
y
ar
e
p
r
esen
ted
to
ass
ess
th
e
ef
f
ec
tiv
en
ess
o
f
t
h
e
weig
h
ted
p
ar
am
ete
r
ap
p
r
o
ac
h
an
d
th
e
m
eta
-
lear
n
er
-
en
h
an
ce
d
en
s
em
b
le
f
r
am
ewo
r
k
in
p
r
e
d
ictin
g
lea
r
n
er
p
er
f
o
r
m
an
ce
.
T
h
e
f
o
llo
win
g
s
ec
tio
n
d
is
cu
s
s
es
th
e
p
er
f
o
r
m
a
n
ce
im
p
r
o
v
e
m
en
t
s
,
co
m
p
ar
e
s
t
h
em
with
ex
is
tin
g
m
o
d
els
.
I
t
also
h
ig
h
lig
h
ts
th
e
im
p
o
r
tan
ce
o
f
t
h
ese
f
in
d
in
g
s
in
t
h
e
b
r
o
ad
er
f
i
eld
o
f
ed
u
ca
tio
n
al
d
ata
m
in
in
g
.
3
.
1
.
Resul
t
s
T
h
e
in
tr
o
d
u
ctio
n
o
f
p
ar
am
et
er
weig
h
ts
s
ig
n
if
ican
tly
im
p
r
o
v
ed
th
e
p
e
r
f
o
r
m
an
ce
o
f
th
e
en
s
em
b
le
m
o
d
el.
As
ca
n
b
e
s
ee
n
in
T
ab
le
4
,
t
h
e
e
n
s
em
b
le
with
o
u
t
p
a
r
am
eter
weig
h
ts
,
t
h
e
m
o
d
el
ac
h
iev
ed
a
n
ac
c
u
r
ac
y
o
f
8
6
.
6
8
%,
with
a
p
r
ec
is
io
n
o
f
8
6
.
5
5
%
an
d
a
r
ec
all
o
f
8
6
.
4
5
%.
W
h
ile
th
ese
m
etr
ics
wer
e
ac
ce
p
tab
le,
ce
r
tain
f
ea
tu
r
es
co
n
tr
i
b
u
ted
litt
le
to
t
h
e
o
v
e
r
all
p
r
e
d
ictio
n
ac
cu
r
ac
y
,
lead
in
g
to
an
in
co
n
s
is
ten
t
p
er
f
o
r
m
an
ce
ac
r
o
s
s
d
if
f
er
en
t
s
tu
d
e
n
t
g
r
o
u
p
s
.
Af
te
r
in
tr
o
d
u
cin
g
p
ar
am
eter
weig
h
ts
,
d
er
iv
ed
u
s
in
g
R
F
f
ea
tu
r
e
im
p
o
r
tan
ce
s
co
r
es,
th
e
en
s
em
b
le
m
o
d
el'
s
ac
cu
r
ac
y
im
p
r
o
v
ed
to
9
2
.
9
1
%
a
r
ela
tiv
e
in
cr
ea
s
e
o
f
7
.
2
%.
Pre
cisi
o
n
an
d
r
ec
all
a
ls
o
in
cr
ea
s
ed
to
9
2
.
5
9
%
an
d
9
2
.
5
0
%,
r
esp
ec
tiv
ely
,
s
h
o
wca
s
in
g
th
e
m
o
d
el'
s
en
h
an
ce
d
a
b
ilit
y
to
co
r
r
ec
tly
id
en
tif
y
at
-
r
is
k
s
tu
d
en
ts
with
o
u
t
s
ac
r
if
icin
g
r
ec
all.
T
h
e
weig
h
te
d
en
s
em
b
le'
s
F1
-
s
co
r
e
f
u
r
th
er
d
em
o
n
s
tr
ated
a
s
ig
n
if
ican
t
im
p
r
o
v
em
en
t,
in
c
r
ea
s
in
g
f
r
o
m
8
6
.
3
0
%
to
9
2
.
1
6
%,
in
d
icatin
g
th
at
t
h
e
m
o
d
el
b
etter
b
alan
ce
d
p
r
ec
is
io
n
an
d
r
ec
all.
T
h
e
r
esu
lts
o
f
th
e
f
ea
tu
r
e
s
elec
tio
n
an
d
m
eta
-
lear
n
er
c
o
m
p
a
r
is
o
n
p
r
o
v
id
e
v
alu
ab
le
in
s
ig
h
t
s
in
to
th
e
im
p
ac
t
o
f
r
e
d
u
cin
g
th
e
n
u
m
b
e
r
o
f
f
ea
t
u
r
es
an
d
in
te
g
r
atin
g
m
eta
-
lear
n
er
s
with
in
an
en
s
e
m
b
le
f
r
a
m
ewo
r
k
.
Fo
r
th
is
we
d
iv
id
ed
th
e
to
tal
s
et
o
f
f
ea
tu
r
es
in
to
g
r
o
u
p
s
o
f
5
,
1
0
,
1
5
,
2
0
,
a
n
d
all
2
3
,
to
o
b
s
er
v
e
h
o
w
co
m
p
u
tatio
n
al
ef
f
icien
cy
a
n
d
m
o
d
el
p
er
f
o
r
m
an
ce
ar
e
a
f
f
ec
ted
.
Fu
r
th
er
m
o
r
e,
we
ev
alu
ated
ea
ch
b
ase
al
g
o
r
ith
m
R
F,
SVM,
KNN,
GB
,
XGB,
an
d
DT
as
p
o
ten
tial
m
eta
-
lear
n
e
r
s
,
r
ev
ea
l
in
g
wh
ich
c
o
m
b
in
atio
n
o
f
b
as
e
lear
n
er
a
n
d
m
eta
-
lear
n
er
d
eliv
er
e
d
th
e
m
o
s
t
ac
cu
r
ate
p
r
ed
ictio
n
s
wh
ile
m
ain
tain
in
g
co
m
p
u
tatio
n
al
ef
f
icie
n
cy
.
T
h
e
f
o
llo
win
g
r
esu
lts
d
em
o
n
s
tr
ate
th
e
s
u
p
er
i
o
r
ity
o
f
h
y
b
r
id
en
s
em
b
le
with
m
eta
lear
n
er
s
o
v
e
r
th
e
b
asic e
n
s
em
b
le
ap
p
r
o
ac
h
.
Fig
u
r
e
4
illu
s
tr
ates
a
c
o
m
p
ar
is
o
n
b
etwe
en
m
eta
an
d
b
asic
en
s
em
b
le
m
o
d
el
f
o
r
t
o
p
‘
5
’
an
d
‘
1
0
’
f
ea
t
u
r
es.
As
s
h
o
wn
in
Fig
u
r
e
4
(
a
)
th
e
b
asic
en
s
em
b
le
wh
en
co
m
p
a
r
ed
with
th
e
h
y
b
r
id
m
eta
-
lear
n
er
e
n
s
em
b
le
f
o
r
t
h
e
to
p
5
p
ar
am
eter
s
r
e
v
ea
ls
a
s
ig
n
if
ica
n
t
ac
cu
r
ac
y
im
p
r
o
v
e
m
en
t.
T
h
e
R
F
an
d
K
N
N
h
y
b
r
id
s
s
h
o
w
th
e
h
ig
h
est
g
ai
n
s
,
ac
h
iev
in
g
9
2
.
7
3
%
ac
cu
r
ac
y
v
er
s
u
s
7
3
.
7
1
%
f
o
r
th
e
b
asic
en
s
em
b
le.
Similar
ly
,
we
ca
n
s
ee
in
F
ig
u
r
e
4
(
b
)
,
f
o
r
th
e
to
p
1
0
p
a
r
am
eter
s
,
th
e
h
y
b
r
id
a
p
p
r
o
ac
h
co
n
tin
u
es
to
o
u
tp
er
f
o
r
m
,
with
XGB
an
d
S
VM
h
y
b
r
id
s
d
eliv
e
r
in
g
th
e
m
o
s
t
n
o
t
ab
le
im
p
r
o
v
em
e
n
ts
with
ac
cu
r
ac
ies
o
f
9
7
.
1
3
%
as
co
m
p
ar
ed
to
7
9
.
7
6
%
o
f
th
e
b
asic
en
s
em
b
le
ap
p
r
o
ac
h
.
T
ab
le
4
.
Per
f
o
r
m
an
ce
co
m
p
a
r
is
o
n
b
etwe
en
weig
h
ted
an
d
n
o
n
-
weig
h
ted
e
n
s
em
b
le
m
o
d
els
M
e
t
r
i
c
N
o
n
-
w
e
i
g
h
t
e
d
e
n
se
mb
l
e
W
e
i
g
h
t
e
d
e
n
s
e
m
b
l
e
I
mp
r
o
v
e
m
e
n
t
(
%)
A
c
c
u
r
a
c
y
8
6
.
6
8
%
9
2
.
9
1
%
7
.
2
%
P
r
e
c
i
s
i
o
n
8
6
.
5
5
%
9
2
.
5
9
%
6
.
9
%
R
e
c
a
l
l
8
6
.
4
5
%
9
2
.
5
0
%
7
.
0
%
F1
-
S
c
o
r
e
8
6
.
3
0
%
9
2
.
1
6
%
6
.
8
%
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
7
,
No
.
3
,
Ma
r
ch
20
2
5
:
2
0
3
2
-
20
4
3
2040
(
a)
(
b
)
Fig
u
r
e
4
.
Me
ta
v
s
b
asic e
n
s
em
b
le
co
m
p
ar
is
o
n
(
a)
c
o
m
p
ar
is
o
n
o
f
to
p
5
f
ea
tu
r
es
an
d
(
b
)
co
m
p
ar
is
o
n
o
f
to
p
1
0
f
ea
tu
r
es
Fig
u
r
e
5
d
ep
icts
a
co
m
p
ar
is
o
n
b
etwe
en
m
eta
an
d
b
asic
en
s
em
b
le
m
o
d
el
f
o
r
to
p
‘
1
5
’
an
d
‘
2
0
’
f
ea
tu
r
es.
Fig
u
r
e
5
(
a)
s
h
o
ws
th
e
ca
s
e
f
o
r
to
p
1
5
p
ar
am
ete
r
s
,
th
e
h
y
b
r
id
e
n
s
em
b
le
ap
p
r
o
ac
h
m
ain
tain
s
its
ad
v
an
tag
e,
with
th
e
G
B
an
d
R
F
h
y
b
r
id
s
ac
h
iev
in
g
th
e
m
o
s
t
s
ig
n
if
ican
t
ac
cu
r
ac
y
g
ai
n
s
,
r
ea
ch
in
g
9
6
.
7
7
%
an
d
9
6
.
5
9
%
,
co
m
p
ar
ed
to
7
7
.
8
9
%
f
o
r
t
h
e
b
asic
e
n
s
em
b
le.
As
th
e
p
ar
am
eter
s
et
in
cr
ea
s
es
to
2
0
as
s
ee
n
in
F
ig
u
r
e
5
(
b
)
,
th
e
tr
e
n
d
p
er
s
is
ts
,
with
GB
an
d
S
VM
h
y
b
r
id
s
co
n
tin
u
in
g
to
o
u
tp
e
r
f
o
r
m
th
e
b
ase
en
s
em
b
le,
d
eliv
er
in
g
ac
c
u
r
ac
ies o
f
9
7
.
0
4
% a
n
d
9
7
.
0
1
%
v
er
s
u
s
7
9
.
6
0
%.
Fig
u
r
e
6
d
is
p
lay
s
two
c
o
m
p
ar
is
o
n
s
b
etwe
en
m
eta
a
n
d
b
asic
en
s
em
b
le
m
o
d
el
f
o
r
to
p
‘
2
3
’
an
d
th
eir
ex
ec
u
tio
n
tim
e.
As
ca
n
b
e
s
ee
n
in
Fig
u
r
e
6
(
a
)
f
o
r
t
h
e
f
u
ll
s
et
o
f
2
3
p
ar
am
eter
s
,
th
e
h
y
b
r
id
m
eta
-
lear
n
er
m
o
d
els
ag
ain
o
u
ts
h
in
e
t
h
e
b
asic
en
s
em
b
le
a
p
p
r
o
ac
h
.
T
h
e
G
B
an
d
S
VM
h
y
b
r
i
d
s
s
tan
d
o
u
t,
ac
h
i
ev
in
g
ac
cu
r
ac
ies
o
f
9
8
.
38
%,
co
m
p
a
r
ed
to
8
6
.
6
8
%
f
o
r
th
e
b
asic
en
s
em
b
le.
Ov
er
all,
ac
r
o
s
s
all
p
ar
am
eter
s
ets,
th
e
h
y
b
r
id
m
eta
-
m
o
d
el
c
o
n
s
is
ten
tly
d
eliv
e
r
ed
f
ar
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
,
h
ig
h
li
g
h
tin
g
its
ef
f
ec
tiv
en
ess
in
en
h
an
cin
g
p
r
ed
ictio
n
ac
cu
r
ac
y
co
m
p
ar
ed
to
th
e
b
asic
en
s
em
b
le
m
o
d
el.
T
h
e
g
r
a
p
h
in
Fig
u
r
e
6
(
b
)
illu
s
tr
ates
h
o
w
r
ed
u
cin
g
th
e
f
ea
tu
r
e
co
u
n
t th
r
o
u
g
h
weig
h
te
d
f
ea
tu
r
e
s
elec
tio
n
s
ig
n
if
ican
tly
d
ec
r
ea
s
es c
o
m
p
u
tatio
n
al
tim
e
f
o
r
all
alg
o
r
ith
m
s
.
Fo
r
in
s
tan
c
e,
th
e
tim
e
d
r
o
p
s
f
r
o
m
3
7
.
6
5
s
ec
o
n
d
s
with
2
3
f
ea
tu
r
es
t
o
2
9
.
3
5
s
ec
o
n
d
s
with
1
0
f
ea
tu
r
es,
h
ig
h
lig
h
tin
g
th
e
ef
f
i
cien
cy
o
f
u
s
in
g
f
ewe
r
,
m
o
r
e
r
elev
an
t
p
ar
am
eter
s
in
ac
ce
le
r
atin
g
th
e
p
r
e
d
ictio
n
p
r
o
ce
s
s
.
(
a)
(
b
)
Fig
u
r
e
5
.
Me
ta
v
s
b
asic e
n
s
em
b
le
co
m
p
ar
is
o
n
(
a)
c
o
m
p
ar
is
o
n
o
f
to
p
1
5
f
ea
tu
r
es
a
n
d
(
b
)
co
m
p
ar
is
o
n
o
f
to
p
2
0
f
ea
tu
r
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:
2
5
0
2
-
4
7
52
Mu
ltif
a
ce
ted
a
p
p
r
o
a
c
h
fo
r
a
n
t
icip
a
tin
g
lea
r
n
er p
erfo
r
ma
n
ce
…
(
S
h
a
b
n
a
m
A
r
a
S
Ja
h
a
g
ir
d
a
r
)
2041
(
a)
(
b
)
Fig
u
r
e
6
.
Per
f
o
r
m
an
c
e
an
d
ex
ec
u
tio
n
tim
e
co
m
p
ar
is
o
n
o
f
m
eta
an
d
b
asic e
n
s
em
b
le
m
o
d
el
s
(
a)
m
eta
v
s
b
asic
en
s
em
b
le
co
m
p
a
r
is
o
n
f
o
r
2
3
f
ea
tu
r
es
an
d
(
b
)
co
m
p
ar
is
o
n
o
f
ex
ec
u
tio
n
tim
e
3
.
2
.
Dis
cus
s
io
ns
T
h
e
weig
h
ted
p
a
r
am
eter
ap
p
r
o
ac
h
em
p
lo
y
ed
in
th
is
s
tu
d
y
p
r
o
v
id
ed
a
s
u
b
s
tan
tial
im
p
r
o
v
e
m
en
t
o
v
er
th
e
tr
ad
itio
n
al
n
o
n
-
weig
h
ted
en
s
em
b
le
m
o
d
els.
T
h
is
is
c
lear
ly
ev
id
e
n
t
with
th
e
n
o
n
-
weig
h
ted
en
s
em
b
le
m
o
d
el'
s
ac
cu
r
ac
y
o
f
8
6
.
6
8
%
c
o
m
p
ar
ed
to
with
th
e
weig
h
ted
en
s
em
b
le'
s
ac
cu
r
ac
y
o
f
9
2
.
9
1
%.
B
u
ild
in
g
o
n
th
e
f
o
u
n
d
atio
n
o
f
weig
h
ted
p
ar
a
m
eter
s
,
th
e
in
teg
r
atio
n
o
f
h
y
b
r
id
m
eta
-
lear
n
e
r
s
f
u
r
th
er
b
o
o
s
ted
th
e
m
o
d
el'
s
p
er
f
o
r
m
an
ce
as
co
m
p
a
r
ed
to
t
h
e
b
asic
en
s
em
b
le
ap
p
r
o
ac
h
.
Fo
r
in
s
tan
ce
,
wh
en
u
tili
zin
g
t
h
e
to
p
5
p
ar
am
eter
s
,
th
e
h
y
b
r
id
m
o
d
el
s
h
o
wed
a
c
o
n
s
id
er
ab
le
ac
c
u
r
ac
y
in
cr
ea
s
e
,
with
R
F
an
d
K
NN
h
y
b
r
id
s
ac
h
iev
in
g
9
2
.
7
3
%,
co
m
p
ar
ed
to
th
e
b
asic
en
s
em
b
le'
s
7
3
.
71
%.
T
h
is
tr
en
d
p
e
r
s
is
ts
ev
en
with
lar
g
er
p
ar
am
et
er
s
ets,
s
u
ch
as
2
3
p
ar
am
eter
s
,
wh
e
r
e
h
y
b
r
id
m
o
d
els
r
ea
ch
e
d
a
n
im
p
r
ess
iv
e
9
8
.
38
%,
f
u
r
th
e
r
r
ein
f
o
r
cin
g
th
e
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
o
f
th
e
h
y
b
r
id
e
n
s
em
b
le
.
An
o
th
er
n
o
tab
le
b
en
e
f
i
t
o
f
ass
ig
n
in
g
weig
h
ts
to
f
ea
t
u
r
es
is
th
e
ab
ilit
y
to
r
ed
u
ce
th
e
ex
ec
u
tio
n
tim
e
b
y
p
r
io
r
itizin
g
th
e
k
ey
f
ea
t
u
r
es.
Fo
r
in
s
tan
ce
,
as
s
ee
n
in
F
ig
u
r
e
6
(
b
)
,
SVM
t
ak
es
2
9
.
3
5
s
ec
o
n
d
s
f
o
r
ex
ec
u
tio
n
u
n
d
er
1
0
f
ea
tu
r
e
s
ets
,
wh
er
ea
s
th
e
s
am
e
m
o
d
el
tak
es
3
7
.
6
5
s
ec
o
n
d
s
f
o
r
a
f
u
ll
f
ea
tu
r
e
s
et
o
f
2
3
p
ar
am
eter
s
.
W
h
en
we
lo
o
k
at
th
e
ac
cu
r
ac
y
p
o
in
t
o
f
v
iew,
th
e
SVM
m
o
d
el
g
iv
es
an
ac
cu
r
ac
y
o
f
9
7
.
1
3
%
f
o
r
th
e
1
0
-
f
ea
tu
r
e
s
et
wh
ich
m
ar
g
in
ally
in
cr
ea
s
es
to
9
8
.
3
8
%
f
o
r
2
3
-
f
ea
tu
r
e
s
et.
B
u
t
th
er
e
i
s
2
1
.
2
4
% in
c
r
ea
s
e
in
ex
ec
u
tio
n
tim
e
with
n
o
s
ig
n
if
ican
t
g
ain
i
n
ac
cu
r
ac
y
lev
els.
W
h
en
co
m
p
a
r
ed
t
o
p
r
io
r
s
tu
d
ies,
th
is
r
esear
ch
p
r
esen
ts
n
o
tab
le
im
p
r
o
v
em
e
n
ts
.
Fo
r
ex
am
p
le,
r
esear
ch
er
s
in
[
1
7
]
,
[2
1]
e
x
p
lo
r
ed
e
n
s
em
b
le
tech
n
iq
u
es
f
o
r
ac
ad
em
ic
p
er
f
o
r
m
an
ce
p
r
ed
ictio
n
,
y
et
th
ei
r
ap
p
r
o
ac
h
es
d
id
n
o
t
f
u
lly
in
te
g
r
ate
th
e
m
eta
-
lear
n
i
n
g
lay
er
.
T
h
is
s
tu
d
y
'
s
h
y
b
r
id
m
eth
o
d
with
b
o
th
b
ase
an
d
m
eta
lear
n
er
s
en
h
an
ce
s
m
o
d
el
ac
cu
r
ac
y
b
ey
o
n
d
th
ese
tr
ad
itio
n
al
en
s
em
b
le
s
.
Mo
r
eo
v
er
,
N
i
y
o
g
i
s
u
b
i
z
o
e
t
a
l
.
[
2
0
]
in
tr
o
d
u
ce
d
a
two
-
lay
e
r
e
n
s
em
b
le
f
o
r
d
r
o
p
o
u
t
p
r
e
d
ictio
n
,
wh
ich
alig
n
s
with
th
e
s
tack
ed
g
e
n
er
aliza
tio
n
u
s
ed
in
th
is
s
tu
d
y
.
Ho
wev
er
,
th
e
in
cl
u
s
io
n
o
f
p
a
r
am
eter
weig
h
tin
g
,
as
in
th
is
r
esear
ch
,
ad
d
s
an
ad
d
itio
n
al
lay
er
o
f
p
r
ec
is
io
n
,
r
e
d
u
cin
g
co
m
p
u
tati
o
n
al
co
s
ts
wh
ile
m
ain
tain
in
g
h
ig
h
ac
cu
r
ac
y
,
m
a
k
in
g
it
a
m
o
r
e
ef
f
icien
t
an
d
s
ca
lab
le
s
o
lu
tio
n
.
Desp
ite
th
e
s
e
s
tr
en
g
th
s
,
a
lim
itatio
n
o
f
th
e
s
tu
d
y
lies
in
th
e
ab
s
en
ce
o
f
lo
n
g
itu
d
i
n
al
d
ata,
wh
ich
co
u
ld
f
u
r
t
h
er
v
alid
ate
t
h
e
m
o
d
el’
s
p
er
f
o
r
m
an
ce
o
v
er
m
u
ltip
le
ac
ad
em
ic
y
ea
r
s
.
T
h
e
p
u
r
p
o
s
e
o
f
t
h
is
wo
r
k
was
to
im
p
r
o
v
e
th
e
p
r
ed
icti
v
e
p
o
wer
o
f
ML
m
o
d
els
f
o
r
s
tu
d
en
t
p
er
f
o
r
m
an
ce
b
y
in
teg
r
atin
g
w
eig
h
ted
p
ar
am
eter
s
an
d
a
h
y
b
r
id
en
s
em
b
le
with
m
eta
-
lear
n
er
s
.
T
h
e
r
esu
lts
o
f
th
is
s
tu
d
y
clea
r
ly
d
em
o
n
s
tr
at
e
th
e
ef
f
ec
tiv
en
ess
o
f
em
p
lo
y
in
g
weig
h
ted
f
ea
t
u
r
es
in
im
p
r
o
v
in
g
p
r
e
d
ictio
n
ac
cu
r
ac
y
.
B
y
ass
ig
n
in
g
weig
h
ts
to
f
ea
tu
r
es
b
ased
o
n
th
ei
r
im
p
o
r
tan
ce
,
th
e
m
o
d
el
f
o
c
u
s
ed
m
o
r
e
o
n
th
e
p
ar
am
eter
s
th
at
h
a
d
th
e
m
o
s
t
s
ig
n
if
ican
t
im
p
ac
t
o
n
s
tu
d
en
t
p
er
f
o
r
m
an
ce
.
T
h
i
s
ap
p
r
o
ac
h
r
ed
u
ce
d
n
o
is
e
f
r
o
m
less
r
elev
an
t
f
ea
tu
r
es,
r
esu
lti
n
g
in
en
h
a
n
ce
d
ac
c
u
r
ac
y
ac
r
o
s
s
all
m
o
d
els.
Mo
r
e
o
v
er
,
th
e
u
s
e
o
f
weig
h
ted
f
ea
tu
r
es
n
o
t
o
n
ly
im
p
r
o
v
ed
p
r
ed
ictiv
e
p
er
f
o
r
m
a
n
ce
b
u
t
also
allo
wed
f
o
r
m
o
r
e
c
o
m
p
u
tatio
n
al
ef
f
icien
c
y
.
B
y
r
ed
u
cin
g
th
e
f
ea
t
u
r
e
s
et
f
r
o
m
2
3
to
s
m
aller
,
h
ig
h
-
im
p
ac
t
s
u
b
s
ets
(
s
u
ch
as
th
e
to
p
5
o
r
1
0
f
ea
t
u
r
es),
th
e
co
m
p
u
tatio
n
al
tim
e
was
s
ig
n
if
ican
tly
r
ed
u
ce
d
wh
ile
m
a
in
tain
in
g
h
ig
h
ac
cu
r
ac
y
.
T
h
e
h
y
b
r
id
e
n
s
em
b
le
ap
p
r
o
ac
h
,
wh
ich
in
co
r
p
o
r
ates
m
eta
-
lear
n
er
s
p
r
o
v
id
ed
an
a
d
d
itio
n
al
lay
e
r
o
f
im
p
r
o
v
em
e
n
t
o
v
er
tr
ad
itio
n
al
en
s
em
b
le
m
o
d
els.
T
h
is
m
eth
o
d
b
en
ef
ite
d
f
r
o
m
th
e
d
iv
er
s
ity
an
d
c
o
m
p
lem
e
n
tar
y
s
tr
en
g
th
s
o
f
m
u
ltip
le
b
ase
lear
n
er
s
,
im
p
r
o
v
in
g
th
e
m
o
d
el's
g
en
er
aliza
tio
n
an
d
p
r
e
d
i
ctiv
e
p
o
wer
.
H
y
b
r
id
en
s
em
b
les
with
weig
h
ted
f
ea
tu
r
es
co
n
s
is
ten
t
ly
o
u
tp
er
f
o
r
m
ed
b
asic
en
s
em
b
les,
d
em
o
n
s
tr
atin
g
th
eir
s
u
p
er
io
r
ity
in
tack
lin
g
co
m
p
lex
,
non
-
lin
ea
r
r
elatio
n
s
h
ip
s
with
in
th
e
d
ata.
Evaluation Warning : The document was created with Spire.PDF for Python.