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
.
2
,
Feb
r
u
ar
y
20
2
5
,
p
p
.
1
20
1
~
1
20
8
I
SS
N:
2
502
-
4
7
52
,
DOI
:
1
0
.
1
1
5
9
1
/ijee
cs
.v
3
7
.
i
2
.
pp
1
20
1
-
1
20
8
1201
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs
.
ia
esco
r
e.
co
m
Co
mpres
so
r perfo
rma
nce predic
ti
o
n:
g
ra
dient
bo
o
sting
regres
sio
n mo
del
a
nd sensi
tivity
an
a
ly
sis
K
uo
-
Chien L
ia
o
1
,
H
o
m
-
Yu
Wu
2
,
H
un
g
-
T
a
Wen
1
,
J
ui
-
T
a
ng
Su
ng
3
,
M
u
ha
m
a
d H
ida
y
a
t
4
,
Will Wei
-
J
uen Wa
ng
5
1
D
e
p
a
r
t
me
n
t
o
f
A
e
r
o
n
a
u
t
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
,
C
h
a
o
y
a
n
g
U
n
i
v
e
r
s
i
t
y
o
f
T
e
c
h
n
o
l
o
g
y
,
Ta
i
c
h
u
n
g
,
Ta
i
w
a
n
2
D
e
p
a
r
t
me
n
t
o
f
M
e
c
h
a
n
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
,
L
u
n
g
h
w
a
U
n
i
v
e
r
si
t
y
o
f
S
c
i
e
n
c
e
a
n
d
Te
c
h
n
o
l
o
g
y
,
T
a
o
y
u
a
n
,
Ta
i
w
a
n
3
D
e
p
a
r
t
me
n
t
o
f
I
n
d
u
s
t
r
i
a
l
a
n
d
M
a
n
a
g
e
me
n
t
,
M
i
n
g
h
si
n
U
n
i
v
e
r
si
t
y
o
f
S
c
i
e
n
c
e
a
n
d
Te
c
h
n
o
l
o
g
y
,
H
si
n
c
h
u
,
Ta
i
w
a
n
4
D
e
p
a
r
t
me
n
t
o
f
M
e
c
h
a
n
i
c
a
l
E
n
g
i
n
e
e
r
i
n
g
,
S
u
mb
a
w
a
U
n
i
v
e
r
s
i
t
y
o
f
T
e
c
h
n
o
l
o
g
y
,
S
u
mb
a
w
a
,
I
n
d
o
n
e
si
a
5
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
i
c
a
l
a
n
d
C
o
m
p
u
t
e
r
E
n
g
i
n
e
e
r
i
n
g
,
U
n
i
v
e
r
si
t
y
o
f
C
a
l
i
f
o
r
n
i
a
,
D
a
v
i
s
,
U
S
A
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ap
r
7
,
2
0
2
4
R
ev
is
ed
Sep
22
,
2
0
2
4
Acc
ep
ted
Sep
30
,
2
0
2
4
Th
is
stu
d
y
i
n
tro
d
u
c
e
s
th
e
u
se
o
f
g
ra
d
ien
t
b
o
o
stin
g
re
g
re
ss
io
n
(G
BR)
m
o
d
e
ls
to
e
stim
a
te
th
e
c
o
m
p
re
ss
o
r
p
e
rfo
r
m
a
n
c
e
o
f
a
e
ro
-
e
n
g
i
n
e
s.
T
h
e
m
o
d
e
l
e
x
h
i
b
it
s
a
m
e
a
n
a
b
so
lu
te
e
rro
r
(M
AE)
o
f
0
.
0
7
8
,
s
h
o
wc
a
sin
g
su
p
e
rio
r
p
e
rfo
rm
a
n
c
e
c
o
m
p
a
re
d
to
p
re
v
io
u
s
stu
d
ies
.
T
h
ro
u
g
h
se
n
siti
v
it
y
a
n
a
ly
sis,
o
p
ti
m
a
l
v
a
lu
e
s
fo
r
th
re
e
k
e
y
p
a
ra
m
e
ter
s
we
re
d
e
term
in
e
d
:
2
8
0
e
stim
a
to
rs,
a
m
a
x
d
e
p
th
o
f
9
,
a
n
d
a
lea
rn
in
g
ra
te
o
f
0
.
0
8
5
.
F
u
r
th
e
rm
o
re
,
a
c
o
m
p
a
riso
n
with
a
p
rio
r
stu
d
y
re
v
e
a
led
a
n
imp
re
ss
iv
e
M
AE
v
a
l
u
e
lo
we
r
t
h
a
n
0
.
0
0
2
,
h
i
g
h
li
g
h
ti
n
g
th
e
G
BR
m
o
d
e
l
’
s
su
c
c
e
ss
in
a
c
c
u
ra
tely
p
re
d
ictin
g
c
o
m
p
re
ss
o
r
p
e
rf
o
rm
a
n
c
e
.
Th
is
d
e
m
o
n
stra
tes
th
e
m
o
d
e
l
’
s
e
ffe
c
ti
v
e
n
e
ss
a
n
d
p
re
d
ictiv
e
a
c
c
u
ra
c
y
,
m
a
k
in
g
it
a
v
a
lu
a
b
le t
o
o
l
fo
r
a
e
ro
-
e
n
g
i
n
e
c
o
m
p
re
ss
o
r
p
e
rfo
rm
a
n
c
e
e
stim
a
ti
o
n
.
K
ey
w
o
r
d
s
:
C
o
m
p
r
ess
o
r
p
er
f
o
r
m
an
ce
Gr
ad
ien
t b
o
o
s
tin
g
r
eg
r
ess
io
n
Me
an
ab
s
o
lu
te
er
r
o
r
O
p
tim
al
v
alu
es
Sen
s
itiv
ity
an
aly
s
is
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
:
Hu
n
g
-
T
a
W
en
Dep
ar
tm
en
t o
f
Aer
o
n
a
u
tical
E
n
g
in
ee
r
in
g
,
C
h
ao
y
a
n
g
Un
i
v
er
s
ity
o
f
T
ec
h
n
o
lo
g
y
T
aich
u
n
g
4
1
3
,
T
aiwa
n
E
m
ail: w
en
h
u
n
g
ta@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
co
m
p
r
ess
o
r
in
an
ae
r
o
e
n
g
in
e
p
lay
s
a
cr
itical
r
o
le
in
th
e
o
v
er
all
p
er
f
o
r
m
an
ce
an
d
ef
f
icien
cy
o
f
th
e
p
r
o
p
u
ls
io
n
s
y
s
tem
.
T
h
e
co
m
p
r
ess
o
r
in
cr
ea
s
es
th
e
p
r
ess
u
r
e
o
f
th
e
in
co
m
in
g
air
,
all
o
win
g
f
o
r
e
f
f
icien
t
co
m
b
u
s
tio
n
an
d
h
ig
h
er
th
r
u
s
t
g
en
er
atio
n
in
th
e
en
g
in
e
[
1
]
.
An
aly
tical
m
o
d
els
h
av
e
b
ee
n
d
ev
elo
p
ed
f
o
r
p
r
ed
ictin
g
g
as
tu
r
b
in
e
en
g
in
e
co
m
p
r
ess
o
r
p
er
f
o
r
m
an
ce
.
Po
u
r
f
ar
za
n
e
h
et
a
l.
[
2
]
in
tr
o
d
u
c
ed
a
d
im
en
s
io
n
less
m
o
d
elin
g
a
p
p
r
o
ac
h
v
alid
ated
th
r
o
u
g
h
ex
p
e
r
im
en
tal
d
ata.
L
ee
et
a
l.
[
3
]
d
ev
elo
p
ed
a
p
e
r
f
o
r
m
an
ce
p
r
ed
ictio
n
p
r
o
g
r
a
m
v
alid
ated
f
o
r
v
ar
io
u
s
g
as
tu
r
b
in
e
ty
p
es.
R
ec
en
t
ad
v
an
ce
m
e
n
ts
in
co
m
p
u
tatio
n
al
f
lu
id
d
y
n
am
ics
(
C
FD)
h
av
e
s
ig
n
if
i
ca
n
tly
im
p
r
o
v
e
d
th
e
p
r
ed
ictio
n
o
f
co
m
p
r
ess
o
r
p
er
f
o
r
m
an
ce
[
4
]
-
[
6
]
.
Ho
wev
er
,
d
esp
ite
im
p
r
o
v
e
d
p
e
r
f
o
r
m
an
ce
,
u
tili
zin
g
C
FD
r
em
ain
s
p
r
o
h
ib
itiv
el
y
ex
p
e
n
s
iv
e,
r
e
q
u
ir
in
g
s
p
ec
ia
lized
h
ar
d
war
e
an
d
s
o
f
twar
e.
I
t is a
s
o
lu
tio
n
co
n
s
id
er
ed
u
s
in
g
a
r
tific
ial
in
tellig
en
ce
t
ec
h
n
o
l
o
g
ies to
s
av
e
th
e
C
FD c
o
s
t.
Ar
tific
ial
in
tellig
en
ce
tech
n
o
l
o
g
ies
h
av
e
b
ee
n
u
tili
ze
d
wid
ely
in
d
if
f
er
en
t
f
ield
s
in
r
ec
en
t
y
ea
r
s
.
Ma
ch
in
e
lear
n
in
g
is
u
s
ed
f
o
r
r
ea
l
estate
v
alu
atio
n
b
r
ea
s
t
ca
n
ce
r
p
r
e
d
ictio
n
[
7
]
,
[
8
]
,
an
d
d
e
ep
lear
n
in
g
is
u
s
ed
f
o
r
s
to
ck
p
r
ices
an
d
o
il
ac
ci
d
en
t
p
r
ed
ictio
n
s
[9
]
,
[
1
0
]
.
Fu
r
th
er
m
o
r
e,
s
o
m
e
r
esear
c
h
aim
ed
at
g
as
tu
r
b
in
e
en
g
in
e
s
tu
d
ies u
s
in
g
a
r
tific
ial
in
tellig
en
ce
tech
n
o
l
o
g
ies wa
s
s
p
r
ea
d
.
Fei
et
a
l.
[
1
1
]
in
tr
o
d
u
ce
d
a
n
in
n
o
v
ativ
e
f
ee
d
-
f
o
r
war
d
n
eu
r
al
n
etwo
r
k
with
a
Gau
s
s
ian
k
er
n
e
l
f
u
n
ctio
n
,
d
em
o
n
s
tr
atin
g
its
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
t
o
o
th
er
d
ata
-
d
r
iv
en
m
o
d
elin
g
tech
n
iq
u
es.
Yaz
ar
et
a
l.
[
1
2
]
co
m
p
r
eh
e
n
s
iv
ely
c
o
m
p
ar
e
d
v
ar
io
u
s
r
eg
r
ess
io
n
m
o
d
els
to
p
r
ed
ict
co
m
p
r
ess
o
r
a
n
d
tu
r
b
i
n
e
m
ap
p
ar
a
m
eter
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
.
2
,
Feb
r
u
a
r
y
20
2
5
:
1
20
1
-
1
20
8
1202
s
h
o
wca
s
in
g
th
eir
p
o
ten
tial
in
d
ev
el
o
p
in
g
d
y
n
am
ic
m
a
th
em
atica
l
m
o
d
els
f
o
r
g
as
tu
r
b
i
n
e
e
n
g
in
es.
L
iu
an
d
Kar
im
i
[
1
3
]
ad
v
an
ce
d
m
ac
h
in
e
lear
n
i
n
g
-
b
ased
m
et
h
o
d
s
to
f
o
r
ec
ast
g
as
tu
r
b
in
e
p
er
f
o
r
m
a
n
ce
th
r
o
u
g
h
s
u
r
r
o
g
ate
m
o
d
els
u
tili
zin
g
h
i
g
h
-
d
im
e
n
s
io
n
al
m
o
d
el
r
e
p
r
es
en
tatio
n
(
HDM
R
)
an
d
ar
tific
ial
n
eu
r
al
n
etwo
r
k
(
ANN
)
.
T
h
ey
p
ar
ticu
lar
ly
em
p
h
asized
th
e
u
tili
ty
o
f
th
e
ANN
m
o
d
el
f
o
r
f
u
ll
-
lo
a
d
p
er
f
o
r
m
an
ce
p
r
e
d
ictio
n
,
lay
in
g
th
e
g
r
o
u
n
d
wo
r
k
f
o
r
c
o
n
tin
u
o
u
s
h
ea
lth
m
o
n
ito
r
i
n
g
a
n
d
f
au
lt
d
iag
n
o
s
is
.
T
h
is
m
eth
o
d
o
lo
g
y
,
ad
ap
ta
b
le
to
an
y
g
as
tu
r
b
in
e
,
q
u
a
n
tifie
s
p
er
f
o
r
m
an
ce
d
eg
r
a
d
atio
n
o
v
er
tim
e.
J
ay
ac
h
an
d
r
a
n
et
a
l.
[
1
4
]
p
r
esen
ted
a
m
ac
h
in
e
lear
n
in
g
-
b
ased
m
o
d
el
to
p
r
e
d
ict
th
e
n
et
ef
f
icien
c
y
o
f
g
a
s
tu
r
b
in
es,
with
t
h
e
r
a
n
d
o
m
f
o
r
est
d
ec
is
io
n
tr
ee
s
h
o
win
g
b
etter
ac
cu
r
ac
y
in
p
r
ed
ictin
g
o
v
er
all
ef
f
icien
c
y
.
Fu
tu
r
e
r
esear
ch
d
ir
ec
tio
n
s
in
c
lu
d
e
ex
p
l
o
r
in
g
k
-
m
ea
n
s
clu
s
ter
in
g
,
h
y
b
r
id
m
o
d
els,
an
d
u
tili
zin
g
ANNs
with
d
ata
f
r
o
m
th
e
ASTR
A
s
im
u
latio
n
to
o
l.
Gh
o
r
b
an
ia
n
an
d
Gh
o
lam
r
ez
a
ei
[
1
5
]
in
v
esti
g
ated
th
e
ap
p
l
icatio
n
o
f
ANNs
f
o
r
c
o
m
p
r
ess
o
r
p
er
f
o
r
m
a
n
ce
p
r
ed
ictio
n
,
c
o
m
p
ar
in
g
d
if
f
e
r
e
n
t
ty
p
es
an
d
co
n
clu
d
in
g
th
at
t
h
e
m
u
ltil
ay
er
p
er
ce
p
tr
o
n
n
etwo
r
k
tech
n
i
q
u
e
is
th
e
m
o
s
t
p
o
wer
f
u
l
ca
n
d
id
ate
f
o
r
in
ter
p
o
latio
n
an
d
ex
tr
ap
o
latio
n
ap
p
licatio
n
s
.
T
h
e
o
th
er
r
esear
ch
o
n
co
m
p
r
ess
o
r
p
er
f
o
r
m
an
ce
b
ased
o
n
th
e
A
NNs
ap
p
r
o
ac
h
,
s
u
ch
as
[
1
6
]
-
[
2
2
]
.
I
n
ad
d
itio
n
,
a
f
ew
s
tu
d
ies
o
n
co
m
p
r
ess
o
r
p
er
f
o
r
m
an
ce
wer
e
p
r
o
p
o
s
ed
u
s
in
g
m
ac
h
in
e
lear
n
in
g
alg
o
r
it
h
m
s
[
2
3
]
,
[
2
4
]
.
T
h
ese
m
ac
h
in
e
lear
n
in
g
m
eth
o
d
s
s
u
cc
ess
f
u
lly
esti
m
ate
co
m
p
r
ess
o
r
p
er
f
o
r
m
an
ce
a
n
d
o
b
tain
s
i
g
n
if
ican
t p
r
ed
ictio
n
r
esu
lts
.
T
h
is
s
tu
d
y
in
v
esti
g
ated
g
as
tu
r
b
in
e
en
g
i
n
e
s
tu
d
ies
b
ased
o
n
ar
tific
ial
in
tellig
en
ce
tech
n
o
lo
g
ies
wid
ely
u
tili
ze
d
in
d
if
f
er
e
n
t
f
i
eld
s
.
T
h
e
g
r
ad
ien
t
b
o
o
s
tin
g
r
e
g
r
ess
io
n
(
GB
R
)
m
o
d
el
was
d
ev
elo
p
ed
to
p
r
ed
ict
th
e
g
as
tu
r
b
in
e
en
g
in
e
’
s
co
m
p
r
ess
o
r
p
er
f
o
r
m
a
n
ce
.
I
n
ad
d
itio
n
,
th
e
m
o
d
el
ac
cu
r
ac
y
was
co
m
p
ar
ed
with
Fei
et
a
l.
[
1
1
]
r
esear
ch
to
v
alid
ate
th
e
m
o
d
el.
2.
M
E
T
H
O
D
2
.
1
.
Da
t
a
A
co
m
p
r
ess
o
r
m
ap
v
is
u
ally
d
is
p
lay
s
a
co
m
p
r
ess
o
r
’
s
p
er
f
o
r
m
an
ce
tr
aits
ac
r
o
s
s
d
if
f
er
en
t
o
p
er
atio
n
a
l
s
ce
n
ar
io
s
.
C
o
m
p
r
ess
o
r
m
ap
s
p
r
o
v
id
e
v
al
u
ab
le
i
n
s
ig
h
ts
in
to
k
ey
p
e
r
f
o
r
m
an
ce
p
ar
am
eter
s
lik
e
p
r
ess
u
r
e
r
atio
(
)
,
co
r
r
ec
ted
m
ass
f
lo
w
(
)
,
ef
f
ici
en
cy
(
)
,
an
d
co
r
r
ec
ted
s
p
ee
d
(
)
.
I
n
th
eo
r
y
,
p
r
ess
u
r
e
r
atio
(
)
an
d
ef
f
icien
cy
(
)
ca
n
b
e
ex
p
r
ess
ed
as
f
u
n
ctio
n
s
o
f
co
r
r
ec
ted
m
ass
f
lo
w
(
)
an
d
co
r
r
ec
ted
s
p
ee
d
(
)
,
as
s
h
o
wn
in
(
1
)
an
d
(
2
)
,
r
esp
ec
tiv
ely
.
A
m
u
ltis
tag
e
ax
ial
f
lo
w
co
m
p
r
ess
o
r
m
ap
was d
r
awn
an
d
s
h
o
wn
i
n
Fig
u
r
e
1
[
1
1
]
.
=
(
,
)
(
1
)
=
(
,
)
(
2
)
Fig
u
r
e
1
.
C
o
m
p
r
ess
o
r
p
er
f
o
r
m
an
ce
m
ap
2
.
2
.
G
ra
dient
bo
o
s
t
ing
re
g
re
s
s
io
n m
o
del
GB
R
i
s
a
m
ac
h
in
e
lear
n
in
g
te
ch
n
iq
u
e
th
at
co
n
s
tr
u
cts
a
p
r
ed
ictiv
e
m
o
d
el
b
y
am
alg
am
atin
g
n
u
m
e
r
o
u
s
w
e
a
k
l
e
a
r
n
e
r
s
,
o
f
t
e
n
d
e
c
i
s
i
o
n
t
r
e
e
s
,
t
o
f
o
r
m
a
r
o
b
u
s
t
p
r
e
d
i
c
t
i
v
e
m
o
d
e
l
.
T
h
i
s
a
l
g
o
r
i
t
h
m
,
p
i
o
n
e
e
r
e
d
b
y
F
r
i
e
d
m
a
n
[
2
5
]
,
en
co
m
p
ass
es
en
s
em
b
le
lear
n
i
n
g
,
wh
ich
a
m
alg
am
ates
th
e
p
r
ed
ictio
n
s
o
f
m
u
ltip
le
in
d
iv
id
u
al
m
o
d
els
(
lear
n
er
s
)
to
en
h
an
ce
o
v
er
all
p
r
ed
ictio
n
ac
cu
r
ac
y
an
d
g
e
n
er
aliza
tio
n
.
W
ea
k
l
ea
r
n
er
s
ar
e
m
o
d
els
th
at
p
er
f
o
r
m
s
lig
h
tly
b
etter
th
an
r
an
d
o
m
g
u
ess
in
g
b
u
t
lack
r
o
b
u
s
tn
ess
.
B
o
o
s
tin
g
is
a
tech
n
iq
u
e
wh
er
e
ea
ch
n
ew
m
o
d
el
in
th
e
en
s
em
b
le
r
ec
tifie
s
th
e
m
is
tak
es
(
r
esid
u
als)
m
ad
e
b
y
t
h
e
p
r
ec
ed
in
g
m
o
d
els.
Gr
ad
ie
n
t
d
es
ce
n
t
o
p
tim
izatio
n
is
an
iter
ativ
e
o
p
tim
izatio
n
alg
o
r
ith
m
th
at
m
in
im
izes
a
lo
s
s
f
u
n
ctio
n
(
e
.
g
.
,
m
ea
n
s
q
u
a
r
ed
e
r
r
o
r
f
o
r
r
eg
r
ess
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
C
o
mp
r
ess
o
r
p
erfo
r
ma
n
ce
p
r
ed
ictio
n
:
g
r
a
d
ien
t b
o
o
s
tin
g
…
(
K
u
o
-
C
h
ien
Lia
o
)
1203
l
og
-
lo
s
s
f
o
r
class
if
icatio
n
)
.
T
h
e
iter
atio
n
p
r
o
ce
s
s
in
v
o
lv
es
tr
ain
in
g
th
e
g
r
ad
ien
t
b
o
o
s
tin
g
m
o
d
el
in
iter
atio
n
s
.
Du
r
in
g
ea
c
h
iter
atio
n
,
a
n
ew
wea
k
lear
n
er
is
in
teg
r
ate
d
in
t
o
th
e
en
s
em
b
le
t
o
r
ec
tif
y
th
e
er
r
o
r
s
m
a
d
e
b
y
t
h
e
p
r
ec
ed
in
g
m
o
d
els.
T
h
e
lear
n
i
n
g
r
ate,
also
r
e
f
er
r
ed
to
as
s
h
r
in
k
ag
e
,
is
a
h
y
p
e
r
p
ar
am
ete
r
th
at
r
e
g
u
lates
th
e
co
n
tr
ib
u
tio
n
o
f
ea
c
h
wea
k
le
ar
n
er
t
o
th
e
f
in
al
p
r
ed
ictio
n
.
R
eg
u
lar
izatio
n
:
g
r
ad
ien
t
b
o
o
s
tin
g
m
o
d
els
o
f
ten
in
clu
d
e
r
eg
u
lar
izatio
n
tech
n
i
q
u
es to
p
r
ev
en
t o
v
er
f
itti
n
g
,
s
u
c
h
as lim
itin
g
th
e
m
ax
im
u
m
d
e
p
th
o
f
d
ec
is
io
n
tr
ee
s
(
m
ax
_
d
e
p
th
)
,
a
d
ju
s
tin
g
th
e
m
i
n
im
u
m
s
am
p
les
p
er
leaf
(
m
in
_
s
am
p
les_
leaf
)
,
o
r
u
s
in
g
ea
r
ly
s
to
p
p
in
g
b
ased
o
n
v
alid
atio
n
er
r
o
r
.
Pre
d
ictio
n
:
o
n
ce
tr
ain
in
g
is
co
m
p
lete,
th
e
g
r
ad
ien
t
b
o
o
s
tin
g
m
o
d
el
ca
n
m
ak
e
p
r
e
d
ictio
n
s
f
o
r
n
ew
d
ata
b
y
ag
g
r
e
g
atin
g
th
e
p
r
ed
ictio
n
s
o
f
all
wea
k
lea
r
n
er
s
ac
co
r
d
in
g
to
th
eir
weig
h
ts
.
E
n
s
em
b
les
ar
e
b
u
ilt
u
s
in
g
d
e
cisi
o
n
tr
ee
m
o
d
els,
with
tr
ee
s
ad
d
ed
in
cr
e
m
en
tally
to
th
e
e
n
s
em
b
le
to
r
ec
tify
p
r
ed
ictio
n
e
r
r
o
r
s
f
r
o
m
p
r
ev
io
u
s
m
o
d
els.
Fig
u
r
e
2
illu
s
tr
ates a
n
ex
am
p
l
e
o
f
th
e
g
r
ad
i
en
t b
o
o
s
tin
g
m
o
d
el
iter
atio
n
p
r
o
ce
d
u
r
e.
T
h
is
iter
a
tiv
e
p
r
o
ce
s
s
co
n
tin
u
es
u
n
til
s
atis
f
ac
to
r
y
r
esu
lts
ar
e
ac
h
ie
v
e
d
,
m
ak
i
n
g
g
r
ad
ien
t
b
o
o
s
tin
g
a
p
o
wer
f
u
l m
ac
h
in
e
l
ea
r
n
in
g
alg
o
r
ith
m
t
h
at
m
in
im
i
ze
s
th
e
lo
s
s
g
r
ad
ien
t e
f
f
ec
tiv
el
y
.
Fig
u
r
e
2
.
Gr
a
d
ien
t b
o
o
s
tin
g
m
o
d
el
iter
atio
n
p
r
o
ce
d
u
r
e
Fig
u
r
e
3
s
h
o
ws
th
e
f
lo
wch
ar
t
o
f
th
e
p
r
o
p
o
s
ed
GB
R
alg
o
r
ith
m
m
o
d
el
f
o
r
co
m
p
r
ess
o
r
p
er
f
o
r
m
an
ce
p
r
ed
ictio
n
.
T
h
is
s
tu
d
y
d
iv
id
es th
e
d
ataset
in
to
two
s
u
b
s
ets:
4
8
s
am
p
les
ar
e
allo
ca
ted
f
o
r
m
o
d
el
tr
ain
in
g
,
wh
ile
1
8
s
am
p
les
ar
e
d
esig
n
ate
d
f
o
r
m
o
d
el
test
in
g
.
T
h
is
d
ata
-
s
p
litt
in
g
s
ch
em
e
ali
g
n
s
with
th
e
a
p
p
r
o
ac
h
u
s
ed
in
th
e
s
tu
d
y
b
y
Fei
et
a
l
.
[
1
1
]
.
T
h
e
m
o
s
t
cr
itical
s
te
p
d
u
r
i
n
g
m
o
d
el
b
u
ild
in
g
is
h
y
p
er
p
ar
a
m
et
er
tu
n
in
g
.
Dif
f
er
e
n
t
m
o
d
el
p
ar
am
eter
s
a
f
f
ec
t
th
e
m
o
d
el
p
er
f
o
r
m
an
ce
an
d
ac
cu
r
ac
y
.
T
h
e
cr
itical
p
ar
am
eter
s
o
f
th
e
GB
R
m
o
d
el
ar
e
as
f
o
llo
ws:
a
m
a
x
_
d
e
p
th
o
f
9
,
N_
esti
m
ato
r
s
s
et
at
2
8
0
,
an
d
a
lea
r
n
in
g
_
r
ate
o
f
0
.
0
8
5
.
T
h
e
ac
cu
r
ac
y
o
f
th
e
m
o
d
el
is
ass
ess
ed
th
r
o
u
g
h
p
er
f
o
r
m
an
ce
m
etr
ic
s
u
ch
as
m
ea
n
ab
s
o
lu
te
er
r
o
r
(
MA
E
)
.
I
n
(
3
)
d
ef
in
es
th
e
m
etr
ics
wh
er
e
P
i
r
ep
r
esen
ts
th
e
p
r
ed
ict
ed
v
alu
e
f
r
o
m
t
h
e
m
o
d
el
an
d
T
i
r
ep
r
esen
ts
th
e
tar
g
et
v
alu
e
f
r
o
m
th
e
d
ataset.
Fig
u
r
e
3
.
Flo
wch
ar
t
o
f
th
e
p
r
o
p
o
s
ed
g
r
a
d
ien
t b
o
o
s
tin
g
r
e
g
r
ess
io
n
m
o
d
el
=
1
∑
|
−
|
=
1
(
3
)
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
3
.
1
.
M
o
del per
f
o
rma
nce
T
h
e
GB
R
m
o
d
el
’
s
p
er
f
o
r
m
a
n
c
e,
as
d
em
o
n
s
tr
ated
in
T
a
b
le
1
,
was
v
alid
ated
b
y
co
m
p
ar
in
g
it
with
th
e
o
u
tco
m
es
r
ep
o
r
te
d
b
y
Fei
et
a
l.
[
1
1
]
.
No
tab
ly
,
th
e
p
r
esen
t
s
tu
d
y
’
s
MA
E
v
alu
e
was
lo
wer
t
h
an
Fei
’
s
m
o
d
el
’
s
.
T
h
is
co
m
p
ar
is
o
n
u
n
d
er
s
co
r
e
s
th
e
ef
f
icac
y
o
f
o
u
r
GB
R
m
o
d
el
in
ac
c
u
r
a
tely
p
r
e
d
ictin
g
co
m
p
r
ess
o
r
p
er
f
o
r
m
an
ce
,
in
d
icatin
g
its
p
o
t
en
tial f
o
r
p
r
ac
tical
ap
p
licatio
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
7
,
No
.
2
,
Feb
r
u
a
r
y
20
2
5
:
1
20
1
-
1
20
8
1204
T
ab
le
1
.
Per
f
o
r
m
an
ce
MA
E
s
tatis
tics
r
esu
lt o
f
th
e
GB
R
m
o
d
el
M
A
E
0
.
0
7
8
0
.
0
8
0
[
1
1
]
T
h
e
co
m
p
ar
is
o
n
o
f
co
m
p
r
ess
o
r
p
e
r
f
o
r
m
an
ce
p
r
e
d
ictio
n
s
at
N=
1
b
etwe
en
th
e
GB
R
m
o
d
e
l
an
d
Fei
’
s
s
tu
d
y
,
as
d
ep
icted
in
Fig
u
r
e
4
,
r
ev
ea
ls
n
o
tab
le
im
p
r
o
v
e
m
en
t
s
in
th
e
GB
R
m
o
d
el
’
s
p
er
f
o
r
m
an
ce
.
Sp
ec
if
ically
,
th
e
GB
R
m
o
d
el
d
em
o
n
s
tr
ates
s
u
p
er
io
r
ac
cu
r
ac
y
co
m
p
ar
ed
t
o
Fei
’
s
s
tu
d
y
,
p
ar
ticu
lar
ly
e
v
id
en
t
in
th
e
r
a
n
g
e
o
f
co
r
r
ec
ted
f
lo
w
ar
o
u
n
d
0
.
0
7
6
t
o
0
.
0
7
8
,
wh
er
e
t
h
e
p
r
ed
ictio
n
d
ev
iatio
n
is
s
ig
n
if
ican
tly
r
e
d
u
ce
d
.
Fu
r
t
h
er
m
o
r
e,
th
e
GB
R
m
o
d
el
ex
h
ib
its
in
cr
ea
s
in
g
ac
cu
r
ac
y
with
h
ig
h
er
co
r
r
ec
ted
f
lo
w
v
alu
es,
with
p
r
ed
ictio
n
s
alig
n
in
g
clo
s
ely
with
ac
tu
al
v
alu
es
f
r
o
m
th
e
test
in
g
d
ataset,
s
ig
n
if
ic
an
tly
ab
o
v
e
0
.
0
8
0
co
r
r
ec
ted
f
l
o
w.
T
h
ese
f
in
d
i
n
g
s
h
ig
h
lig
h
t
t
h
e
GB
R
m
o
d
el
’
s
r
o
b
u
s
tn
ess
an
d
ca
p
ab
ilit
y
t
o
a
cc
u
r
ately
p
r
ed
ict
c
o
m
p
r
ess
o
r
p
er
f
o
r
m
an
ce
ac
r
o
s
s
v
ar
y
in
g
f
lo
w
co
n
d
itio
n
s
,
s
h
o
wca
s
in
g
its
p
o
ten
tial
f
o
r
e
n
h
an
cin
g
p
r
e
d
ictiv
e
ac
cu
r
ac
y
an
d
r
eliab
ilit
y
in
p
r
ac
t
ical
ap
p
licatio
n
s
with
in
t
h
e
co
m
p
r
ess
o
r
in
d
u
s
tr
y
.
Fig
u
r
e
4
.
Pre
d
ictio
n
o
f
co
m
p
r
ess
o
r
p
er
f
o
r
m
a
n
ce
3
.
2
.
Sens
it
iv
it
y
a
na
l
y
s
is
3
.
2
.
1
.
E
f
f
ec
t
o
f
t
he
nu
m
ber
o
f
estim
a
t
o
rs
T
h
e
n
u
m
b
e
r
o
f
esti
m
ato
r
s
is
a
c
r
itical
h
y
p
er
p
ar
am
eter
i
n
GB
R
m
o
d
els,
d
ir
ec
tly
in
f
lu
en
cin
g
th
e
m
o
d
el
’
s
p
er
f
o
r
m
an
ce
an
d
co
m
p
lex
ity
.
I
n
ess
en
ce
,
ea
ch
esti
m
ato
r
r
ep
r
esen
ts
an
in
d
iv
id
u
al
d
ec
is
io
n
tr
ee
in
th
e
en
s
em
b
le,
an
d
in
cr
ea
s
in
g
th
e
n
u
m
b
e
r
o
f
esti
m
ato
r
s
ca
n
h
av
e
b
o
th
p
o
s
itiv
e
a
n
d
n
eg
ati
v
e
ef
f
ec
ts
.
Fig
u
r
e
5
s
h
o
ws
th
e
o
b
s
er
v
ed
v
ar
iatio
n
in
MA
E
v
alu
es
ac
r
o
s
s
d
if
f
er
en
t
n
u
m
b
er
s
o
f
esti
m
ato
r
s
,
r
a
n
g
in
g
f
r
o
m
1
8
0
to
3
3
0
,
w
h
ich
p
r
o
v
i
d
es
v
alu
ab
le
in
s
ig
h
ts
in
to
m
o
d
el
b
e
h
av
io
r
.
I
n
ter
esti
n
g
l
y
,
a
s
ig
n
if
ica
n
t
r
ed
u
ctio
n
in
MA
E
o
cc
u
r
s
wh
en
th
e
n
u
m
b
e
r
o
f
esti
m
ato
r
s
r
ea
ch
es
2
8
0
,
in
d
ic
atin
g
im
p
r
o
v
ed
p
r
e
d
ictiv
e
ac
cu
r
ac
y
a
n
d
m
o
d
el
p
er
f
o
r
m
an
ce
.
T
h
is
f
in
d
in
g
s
u
g
g
ests
th
at
ad
d
in
g
m
o
r
e
b
o
o
s
tin
g
s
tag
es
b
ey
o
n
d
th
is
p
o
in
t
m
ay
n
o
t
n
ec
ess
ar
ily
lead
to
f
u
r
th
er
en
h
an
ce
m
en
ts
in
p
r
ed
ictio
n
q
u
ality
,
as
e
v
id
e
n
ce
d
b
y
th
e
s
tab
le
lo
w
MA
E
v
alu
es
ev
en
at
3
3
0
esti
m
ato
r
s
.
T
h
e
s
tab
ilit
y
o
f
th
e
lo
west
MA
E
v
alu
e
as
t
h
e
n
u
m
b
er
o
f
esti
m
ato
r
s
in
cr
ea
s
es
t
o
3
3
0
h
i
g
h
lig
h
ts
a
n
in
tr
ig
u
in
g
asp
ec
t
o
f
m
o
d
el
c
o
n
v
er
g
e
n
ce
an
d
o
p
tim
izatio
n
.
I
t
s
u
g
g
ests
th
at
th
e
m
o
d
el
’
s
ca
p
ac
ity
to
g
en
er
alize
an
d
ca
p
tu
r
e
u
n
d
er
ly
in
g
p
atter
n
s
m
ay
h
av
e
r
ea
ch
e
d
a
p
latea
u
b
ey
o
n
d
wh
ic
h
ad
d
itio
n
al
co
m
p
lex
ity
d
o
es
n
o
t
s
ig
n
if
ican
tly
im
p
r
o
v
e
p
er
f
o
r
m
an
ce
.
T
h
ese
r
esu
lts
u
n
d
er
s
c
o
r
e
th
e
im
p
o
r
tan
ce
o
f
h
y
p
er
p
ar
am
eter
tu
n
in
g
in
GB
R
m
o
d
els,
s
p
ec
if
ically
th
e
n
u
m
b
er
o
f
esti
m
ato
r
s
,
to
ac
h
i
ev
e
o
p
tim
al
p
r
e
d
ictiv
e
ac
cu
r
a
cy
wh
ile
ef
f
ec
tiv
el
y
m
an
ag
in
g
co
m
p
u
tatio
n
al
r
eso
u
r
ce
s
.
Fig
u
r
e
5
.
Sen
s
itiv
ity
o
f
th
e
G
B
R
m
o
d
el
to
th
e
n
u
m
b
er
o
f
esti
m
ato
r
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
C
o
mp
r
ess
o
r
p
erfo
r
ma
n
ce
p
r
ed
ictio
n
:
g
r
a
d
ien
t b
o
o
s
tin
g
…
(
K
u
o
-
C
h
ien
Lia
o
)
1205
3
.
2
.
2
.
E
f
f
ec
t
o
f
t
he
m
ax
-
dept
h
T
h
e
m
ax
_
d
ep
th
h
y
p
e
r
p
ar
am
e
ter
in
GB
R
m
o
d
els
is
p
iv
o
tal
in
m
an
ag
in
g
m
o
d
el
co
m
p
lex
ity
an
d
en
s
u
r
in
g
o
p
tim
al
p
er
f
o
r
m
an
ce
.
I
t
d
ictates
th
e
m
a
x
im
u
m
d
ep
th
o
f
ea
ch
d
ec
is
io
n
tr
ee
in
th
e
en
s
em
b
le,
d
ir
ec
tl
y
in
f
lu
en
cin
g
th
e
tr
ee
s
’
co
m
p
le
x
ity
an
d
th
e
o
v
e
r
all
m
o
d
el
’
s
ab
ilit
y
to
ca
p
tu
r
e
in
tr
icate
p
a
tter
n
s
in
th
e
d
ata.
Fig
u
r
e
6
s
h
o
ws
th
e
o
b
s
er
v
e
d
r
an
g
e
o
f
m
ax
_
d
ep
th
v
alu
es
f
r
o
m
4
to
1
0
,
w
h
ich
p
r
o
v
id
es
a
c
o
m
p
r
eh
e
n
s
iv
e
v
iew
o
f
h
o
w
v
a
r
y
in
g
th
is
h
y
p
er
p
a
r
am
eter
im
p
ac
ts
m
o
d
el
p
e
r
f
o
r
m
an
ce
.
Du
r
i
n
g
th
e
m
o
d
el
tr
ain
in
g
an
d
test
in
g
p
h
ases
,
th
e
MA
E
v
alu
es
wer
e
clo
s
ely
m
o
n
ito
r
ed
ac
r
o
s
s
d
if
f
e
r
en
t
m
ax
_
d
ep
t
h
v
al
u
es.
A
n
o
t
ab
le
tr
en
d
em
er
g
ed
wh
er
e
th
e
m
o
d
el
’
s
MA
E
r
ea
c
h
ed
its
lo
west
v
alu
e
wh
en
th
e
m
ax
_
d
ep
t
h
p
ar
am
eter
was
s
et
to
9
.
T
h
is
f
in
d
in
g
s
u
g
g
ests
th
at
a
m
ax
_
d
ep
th
o
f
9
s
tr
ik
es
a
b
alan
ce
b
etwe
en
ca
p
tu
r
in
g
ess
en
tial
f
ea
tu
r
es
in
th
e
d
ata
wh
ile
av
o
id
in
g
o
v
er
f
itti
n
g
,
r
esu
ltin
g
in
im
p
r
o
v
e
d
p
r
e
d
ictiv
e
ac
cu
r
a
cy
an
d
g
e
n
er
aliza
tio
n
ab
ilit
y
.
T
h
e
s
ig
n
if
ican
ce
o
f
th
is
o
p
tim
al
m
ax
_
d
ep
th
v
alu
e
u
n
d
er
s
co
r
es
th
e
cr
itical
r
o
le
o
f
h
y
p
e
r
p
ar
am
eter
tu
n
in
g
in
GB
R
m
o
d
els.
Fin
e
-
t
u
n
in
g
m
ax
_
d
ep
th
with
in
a
r
ea
s
o
n
ab
le
r
an
g
e
ca
n
s
u
b
s
tan
tially
im
p
r
o
v
e
m
o
d
el
p
e
r
f
o
r
m
a
n
ce
with
o
u
t
s
ac
r
if
icin
g
co
m
p
u
tatio
n
al
ef
f
ic
ien
cy
.
Fig
u
r
e
6
.
Sen
s
itiv
ity
o
f
th
e
G
B
R
m
o
d
el
to
th
e
m
ax
-
d
ep
th
3
.
2
.
3
.
E
f
f
ec
t
o
f
t
he
l
ea
rning
r
a
t
e
T
h
e
lear
n
in
g
r
ate
is
a
cr
itical
h
y
p
er
p
ar
am
eter
i
n
th
e
GB
R
m
o
d
el,
d
ictatin
g
th
e
im
p
ac
t o
f
e
ac
h
tr
ee
o
n
th
e
en
s
em
b
le
d
u
r
in
g
tr
ain
in
g
.
Fig
u
r
e
7
r
ev
ea
ls
th
e
r
an
g
e
o
f
lear
n
in
g
r
ates
ex
p
l
o
r
ed
,
f
r
o
m
0
.
0
6
5
to
0
.
1
,
an
d
s
h
ed
s
lig
h
t
o
n
h
o
w
th
is
p
ar
a
m
eter
in
f
lu
e
n
ce
s
m
o
d
el
p
er
f
o
r
m
an
ce
.
O
b
s
er
v
in
g
MA
E
v
al
u
es
ac
r
o
s
s
d
if
f
er
e
n
t
lear
n
in
g
r
ates
d
u
r
in
g
m
o
d
el
tr
ain
in
g
an
d
test
in
g
p
r
o
v
i
d
es
es
s
en
tial
in
s
ig
h
ts
in
to
o
p
tim
al
lear
n
in
g
r
ate
s
elec
tio
n
.
T
h
e
tr
en
d
o
b
s
er
v
e
d
in
d
icate
s
th
at
in
cr
ea
s
in
g
th
e
lear
n
in
g
r
ate
in
itially
r
e
d
u
ce
s
MA
E
,
with
th
e
lo
west M
AE
ac
h
iev
ed
at
a
lear
n
in
g
r
ate
o
f
0
.
0
8
5
.
T
h
is
f
in
d
i
n
g
u
n
d
er
s
co
r
es th
e
im
p
o
r
tan
ce
o
f
f
in
d
in
g
th
e
r
ig
h
t
b
alan
ce
;
a
m
o
d
er
ate
lear
n
in
g
r
ate
o
p
tim
izes
m
o
d
el
co
n
v
e
r
g
e
n
ce
an
d
ac
cu
r
ac
y
with
o
u
t
co
m
p
r
o
m
is
in
g
s
tab
ilit
y
o
r
r
is
k
in
g
o
v
er
f
itti
n
g
.
Ho
we
v
er
,
a
r
ap
id
r
is
e
in
MA
E
is
n
o
ted
b
ey
o
n
d
th
is
o
p
tim
al
lear
n
in
g
r
ate,
in
d
icatin
g
d
im
in
is
h
in
g
r
etu
r
n
s
o
r
ev
en
a
d
eter
io
r
atio
n
in
m
o
d
el
p
er
f
o
r
m
an
ce
as
th
e
lear
n
in
g
r
ate
b
ec
o
m
es
to
o
h
ig
h
.
T
h
is
o
b
s
er
v
ed
r
elatio
n
s
h
ip
b
e
twee
n
lear
n
in
g
r
ate
an
d
MA
E
h
ig
h
lig
h
ts
th
e
d
elica
te
tr
ad
e
-
o
f
f
b
et
wee
n
m
o
d
el
o
p
tim
izatio
n
an
d
av
o
i
d
in
g
o
v
e
r
f
itti
n
g
.
Pro
p
er
h
y
p
er
p
a
r
am
eter
tu
n
i
n
g
is
cr
u
cial
f
o
r
m
o
d
el
s
u
cc
ess
.
L
ea
r
n
in
g
r
ate
is
a
cr
itical
h
y
p
e
r
p
ar
am
eter
th
at
s
ig
n
if
ican
tly
im
p
ac
ts
m
o
d
el
p
er
f
o
r
m
an
ce
.
Settin
g
it
to
o
h
ig
h
ca
n
lead
to
d
iv
e
r
g
en
ce
wh
ile
s
ettin
g
i
t
to
o
lo
w
ca
n
r
esu
lt in
s
lo
w
co
n
v
er
g
en
ce
.
C
ar
ef
u
l o
p
tim
izatio
n
is
ess
en
tial to
ac
h
iev
e
o
p
tim
al
r
es
u
lts
.
Fig
u
r
e
7
.
Sen
s
itiv
ity
o
f
th
e
G
B
R
m
o
d
el
to
th
e
l
ea
r
n
in
g
r
ate
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
.
2
,
Feb
r
u
a
r
y
20
2
5
:
1
20
1
-
1
20
8
1206
T
h
is
s
tu
d
y
th
o
r
o
u
g
h
l
y
in
v
esti
g
ated
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
G
B
R
m
o
d
el
th
r
o
u
g
h
a
d
etailed
a
n
aly
s
is
o
f
k
ey
m
o
d
el
p
ar
am
ete
r
s
,
in
clu
d
in
g
th
e
n
u
m
b
er
o
f
esti
m
ato
r
s
,
m
ax
d
e
p
th
,
an
d
lear
n
in
g
r
ate
.
T
h
ese
p
ar
am
eter
s
p
lay
a
p
iv
o
tal
r
o
le
in
s
h
a
p
in
g
th
e
m
o
d
el
’
s
tr
ain
in
g
an
d
t
esti
n
g
o
u
tco
m
es,
d
ir
ec
tly
im
p
ac
tin
g
its
o
v
er
all
p
er
f
o
r
m
an
ce
a
n
d
p
r
ed
ictiv
e
a
cc
u
r
ac
y.
T
h
e
in
v
esti
g
atio
n
d
el
v
ed
in
to
u
n
d
er
s
tan
d
in
g
th
e
m
o
d
el
’
s
s
en
s
itiv
ity
to
th
ese
p
ar
am
eter
s
,
aim
in
g
t
o
id
en
tify
th
e
o
p
tim
al
co
n
f
ig
u
r
atio
n
s
th
at
wo
u
ld
y
iel
d
th
e
b
est
r
esu
lts
.
T
h
is
in
v
o
lv
ed
co
n
d
u
ctin
g
a
co
m
p
r
eh
e
n
s
iv
e
s
en
s
itiv
ity
an
aly
s
is
b
y
ad
ju
s
tin
g
th
e
p
ar
a
m
eter
s
’
r
an
g
es
an
d
o
b
s
er
v
in
g
th
eir
ef
f
ec
ts
o
n
th
e
m
o
d
el
’
s
p
er
f
o
r
m
an
ce
m
etr
ics.
Su
ch
an
an
aly
s
is
is
cr
u
cial
i
n
g
ain
in
g
v
alu
a
b
le
in
s
ig
h
ts
in
to
h
o
w
c
h
an
g
es
i
n
t
h
ese
p
ar
am
eter
s
in
f
lu
en
ce
th
e
m
o
d
el
’
s
b
e
h
av
io
r
d
u
r
in
g
th
e
t
r
ain
in
g
an
d
test
in
g
p
h
ases
.
T
h
e
s
en
s
iti
v
ity
an
aly
s
is
p
r
o
v
id
ed
r
esear
ch
er
s
with
cr
itical
in
f
o
r
m
atio
n
f
o
r
f
in
e
-
tu
n
in
g
th
e
m
o
d
el
p
ar
am
eter
s
.
B
y
ca
r
ef
u
lly
ad
ju
s
tin
g
th
e
r
an
g
es
an
d
o
b
s
er
v
in
g
th
e
co
r
r
esp
o
n
d
i
n
g
v
ar
iatio
n
s
in
m
o
d
el
p
er
f
o
r
m
an
ce
,
r
esear
ch
er
s
co
u
ld
m
ak
e
i
n
f
o
r
m
ed
d
ec
is
io
n
s
ab
o
u
t
th
e
o
p
tim
a
l
s
ettin
g
s
f
o
r
ea
ch
p
ar
am
eter
.
T
h
is
iter
ativ
e
p
r
o
ce
s
s
o
f
an
aly
s
is
an
d
ad
ju
s
tm
en
t
m
ax
im
izes
th
e
m
o
d
el
’
s
p
r
ed
ictiv
e
ac
cu
r
a
cy
an
d
en
s
u
r
es
th
at
it p
er
f
o
r
m
s
o
p
tim
ally
ac
r
o
s
s
d
i
f
f
er
en
t scen
ar
io
s
a
n
d
d
atasets
.
Fo
llo
win
g
th
e
s
en
s
itiv
ity
an
al
y
s
is
,
th
e
s
tu
d
y
s
u
cc
es
s
f
u
lly
i
d
en
tifie
d
th
e
o
p
tim
al
p
ar
am
et
er
s
f
o
r
t
h
e
GB
R
m
o
d
el.
T
h
ese
o
p
tim
al
co
n
f
ig
u
r
atio
n
s
wer
e
s
u
m
m
ar
i
ze
d
an
d
p
r
esen
ted
i
n
T
ab
le
2
,
p
r
o
v
id
i
n
g
a
clea
r
r
ef
er
en
ce
f
o
r
f
u
t
u
r
e
r
esear
ch
an
d
p
r
ac
tical
im
p
lem
en
tatio
n
s
.
B
y
estab
lis
h
in
g
th
e
id
ea
l
p
ar
am
eter
v
alu
es,
re
s
ea
r
ch
er
s
ca
n
s
tr
ea
m
lin
e
th
e
m
o
d
el
d
ev
elo
p
m
en
t
p
r
o
ce
s
s
an
d
en
h
an
ce
its
p
er
f
o
r
m
an
ce
w
ith
o
u
t
u
n
n
ec
ess
ar
y
tr
ial
an
d
er
r
o
r
.
T
ab
le
2
.
Op
ti
m
a
l
p
a
ra
m
e
ters
o
f
t
h
e
G
BR
m
o
d
e
l
n
u
m
b
e
r
o
f
e
st
i
ma
t
o
r
s
max
d
e
p
t
h
l
e
a
r
n
i
n
g
r
a
t
e
2
8
0
9
0
.
0
8
5
4.
CO
NCLU
SI
O
N
T
h
e
co
m
p
r
ess
o
r
p
er
f
o
r
m
an
c
e
m
o
d
el
co
n
s
tr
u
cted
u
s
in
g
th
e
GB
R
ap
p
r
o
ac
h
was
d
ev
elo
p
e
d
s
u
cc
ess
f
u
lly
,
an
d
it
o
b
tain
ed
i
m
p
r
o
v
e
d
m
o
d
el
p
er
f
o
r
m
an
ce
co
m
p
ar
ed
t
o
p
r
e
v
io
u
s
s
tu
d
ies.
T
h
is
s
tar
k
co
n
tr
ast
u
n
d
er
s
co
r
es
t
h
e
ad
v
an
tag
es
o
f
em
p
lo
y
i
n
g
th
e
GB
R
m
o
d
el
f
o
r
s
u
ch
task
s
.
F
o
llo
win
g
a
m
e
ticu
lo
u
s
s
en
s
itiv
ity
an
aly
s
is
,
th
e
s
tu
d
y
id
en
tifie
d
o
p
tim
al
v
alu
es
f
o
r
t
h
r
ee
k
ey
p
a
r
am
eter
s
:
th
e
n
u
m
b
e
r
o
f
esti
m
ato
r
s
,
lear
n
in
g
r
ate,
an
d
m
ax
d
ep
t
h
.
Me
ticu
lo
u
s
ly
ex
p
lo
r
ed
a
n
d
r
ef
in
e
d
with
in
s
p
ec
if
ic
r
an
g
es,
th
ese
v
alu
es
wer
e
in
s
tr
u
m
en
tal
in
fi
ne
-
tu
n
in
g
th
e
m
o
d
el
f
o
r
s
u
p
e
r
io
r
p
r
e
d
ictiv
e
ac
cu
r
ac
y
an
d
p
er
f
o
r
m
a
n
ce
.
Fu
tu
r
e
wo
r
k
r
elate
d
to
th
is
s
tu
d
y
c
o
u
ld
c
o
n
ce
n
t
r
ate
o
n
s
ev
er
al
k
ey
ar
ea
s
to
f
u
r
th
er
e
n
h
an
ce
th
e
m
o
d
el
’
s
ca
p
ab
ilit
ies.
On
e
av
en
u
e
o
f
ex
p
lo
r
atio
n
in
v
o
lv
es
f
in
e
-
tu
n
in
g
h
y
p
e
r
p
ar
am
et
er
s
b
ey
o
n
d
t
h
o
s
e
s
cr
u
tin
ized
in
th
e
cu
r
r
en
t
s
tu
d
y
.
R
esear
ch
er
s
ca
n
u
n
l
o
ck
f
u
r
th
er
m
o
d
el
ac
cu
r
ac
y
an
d
r
o
b
u
s
tn
ess
im
p
r
o
v
em
e
n
ts
b
y
d
elv
in
g
in
to
o
th
er
p
o
ten
tial
p
ar
am
eter
s
an
d
o
p
tim
izin
g
th
eir
s
etti
n
g
s
.
Ad
d
itio
n
ally
,
co
n
d
u
ctin
g
b
r
o
a
d
er
s
en
s
itiv
ity
an
aly
s
es
with
an
ex
p
a
n
d
e
d
r
an
g
e
o
f
p
ar
am
eter
v
alu
es
p
r
esen
ts
an
o
th
er
p
r
o
m
is
in
g
d
ir
ec
tio
n
f
o
r
f
u
t
u
r
e
r
esear
ch
.
T
h
is
ap
p
r
o
ac
h
allo
ws
f
o
r
a
m
o
r
e
co
m
p
r
eh
e
n
s
iv
e
u
n
d
er
s
tan
d
in
g
o
f
h
o
w
h
y
p
er
p
ar
a
m
eter
v
ar
iatio
n
s
im
p
ac
t
th
e
m
o
d
el
’
s
p
er
f
o
r
m
an
ce
ac
r
o
s
s
d
if
f
er
en
t
s
ce
n
ar
io
s
an
d
d
atasets
.
Em
p
lo
y
in
g
ad
v
an
ce
d
o
p
tim
izatio
n
tech
n
i
q
u
es,
s
u
ch
as
B
a
y
esian
o
p
tim
izatio
n
o
r
g
e
n
etic
alg
o
r
ith
m
s
,
co
u
l
d
also
p
r
o
v
e
b
en
ef
icial
in
o
p
tim
i
zin
g
th
e
m
o
d
el
’
s
p
ar
am
ete
r
s
m
o
r
e
ef
f
icien
tly
an
d
ef
f
ec
tiv
el
y
.
RE
F
E
R
E
NC
E
S
[
1
]
R
.
O
.
B
u
l
l
o
c
k
,
A
e
r
o
d
y
n
a
m
i
c
d
e
si
g
n
o
f
a
x
i
a
l
f
l
o
w
c
o
m
p
r
e
sso
r
s
.
W
a
s
h
i
n
g
t
o
n
,
D
C
,
U
S
A
:
N
a
t
i
o
n
a
l
A
e
r
o
n
a
u
t
i
c
s
a
n
d
S
p
a
c
e
A
d
mi
n
i
s
t
r
a
t
i
o
n
,
1
9
6
5
.
[
2
]
H
.
P
o
u
r
f
a
r
z
a
n
e
h
,
A
.
H
a
j
i
l
o
u
y
-
B
e
n
i
si
,
a
n
d
M
.
F
a
r
sh
c
h
i
,
“
A
n
a
n
a
l
y
t
i
c
a
l
m
o
d
e
l
o
f
a
g
a
s
t
u
r
b
i
n
e
c
o
m
p
o
n
e
n
t
s
p
e
r
f
o
r
m
a
n
c
e
a
n
d
i
t
s
e
x
p
e
r
i
m
e
n
t
a
l
v
a
l
i
d
a
t
i
o
n
,
”
i
n
T
u
r
b
o
E
x
p
o
:
Po
w
e
r
f
o
r
L
a
n
d
,
S
e
a
,
a
n
d
Ai
r
,
O
c
t
.
2
0
1
0
,
p
p
.
3
3
5
–
3
4
0
,
d
o
i
:
1
0
.
1
1
1
5
/
G
T2
0
1
0
-
2
3
3
6
9
.
[
3
]
J.
J.
L
e
e
,
D
.
W
.
K
a
n
g
,
a
n
d
T
.
S
.
K
i
m,
“
D
e
v
e
l
o
p
me
n
t
o
f
a
g
a
s
t
u
r
b
i
n
e
p
e
r
f
o
r
man
c
e
a
n
a
l
y
si
s
p
r
o
g
r
a
m
a
n
d
i
t
s
a
p
p
l
i
c
a
t
i
o
n
,
”
E
n
e
r
g
y
,
v
o
l
.
3
6
,
n
o
.
8
,
p
p
.
5
2
7
4
–
5
2
8
5
,
A
u
g
.
2
0
1
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
n
e
r
g
y
.
2
0
1
1
.
0
6
.
0
3
2
.
[
4
]
M
.
K
e
t
h
i
d
i
,
A
.
K
o
v
a
c
e
v
i
c
,
N
.
S
t
o
s
i
c
,
a
n
d
I
.
K
.
S
mi
t
h
,
“
E
v
a
l
u
a
t
i
o
n
o
f
v
a
r
i
o
u
s
t
u
r
b
u
l
e
n
c
e
m
o
d
e
l
s
i
n
p
r
e
d
i
c
t
i
n
g
screw
c
o
m
p
r
e
ss
o
r
f
l
o
w
p
r
o
c
e
sses
b
y
C
F
D
,
”
i
n
7
t
h
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
C
o
m
p
r
e
ss
o
rs
a
n
d
t
h
e
i
r
S
y
st
e
m
s
2
0
1
1
,
El
s
e
v
i
e
r
,
2
0
1
1
,
p
p
.
3
4
7
–
3
5
7
.
[
5
]
J.
J
.
M
o
o
r
e
,
D
.
L.
R
a
n
s
o
m,
a
n
d
F
.
V
i
a
n
a
,
“
R
o
t
o
r
d
y
n
a
m
i
c
f
o
r
c
e
p
r
e
d
i
c
t
i
o
n
o
f
c
e
n
t
r
i
f
u
g
a
l
c
o
mp
r
e
sso
r
i
m
p
e
l
l
e
r
s
u
s
i
n
g
c
o
m
p
u
t
a
t
i
o
n
a
l
f
l
u
i
d
d
y
n
a
m
i
c
s
,
”
J
o
u
rn
a
l
o
f
E
n
g
i
n
e
e
r
i
n
g
f
o
r
G
a
s
T
u
rb
i
n
e
s
a
n
d
Po
w
e
r
,
v
o
l
.
1
3
3
,
n
o
.
4
,
A
p
r
.
2
0
1
1
,
d
o
i
:
1
0
.
1
1
1
5
/
1
.
2
9
0
0
9
5
8
.
[
6
]
E.
L
.
L.
P
e
r
e
i
r
a
,
C
.
J.
D
e
s
c
h
a
m
p
s,
a
n
d
F
.
A
.
R
i
b
a
s,
“
P
e
r
f
o
r
m
a
n
c
e
a
n
a
l
y
s
i
s
o
f
r
e
c
i
p
r
o
c
a
t
i
n
g
c
o
m
p
r
e
ss
o
r
s
t
h
r
o
u
g
h
c
o
m
p
u
t
a
t
i
o
n
a
l
f
l
u
i
d
d
y
n
a
mi
c
s,
”
i
n
Pr
o
c
e
e
d
i
n
g
s
o
f
t
h
e
I
n
s
t
i
t
u
t
i
o
n
o
f
Me
c
h
a
n
i
c
a
l
En
g
i
n
e
e
rs
,
P
a
rt
E:
J
o
u
rn
a
l
o
f
Pr
o
c
e
ss
M
e
c
h
a
n
i
c
a
l
E
n
g
i
n
e
e
ri
n
g
,
N
o
v
.
2
0
0
8
,
v
o
l
.
2
2
2
,
n
o
.
4
,
p
p
.
1
8
3
–
1
9
2
,
d
o
i
:
1
0
.
1
2
4
3
/
0
9
5
4
4
0
8
9
JP
M
E1
9
4
.
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
C
o
mp
r
ess
o
r
p
erfo
r
ma
n
ce
p
r
ed
ictio
n
:
g
r
a
d
ien
t b
o
o
s
tin
g
…
(
K
u
o
-
C
h
ien
Lia
o
)
1207
[
7
]
A
.
B
a
r
l
y
b
a
y
e
v
,
A
.
S
a
n
k
i
b
a
y
e
v
,
R
.
N
i
y
a
z
o
v
a
,
a
n
d
G
.
A
k
i
m
b
e
k
o
v
a
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
f
o
r
r
e
a
l
e
st
a
t
e
v
a
l
u
a
t
i
o
n
:
A
st
a
n
a
,
K
a
z
a
k
h
s
t
a
n
c
a
se
,
”
I
n
d
o
n
e
si
a
n
J
o
u
rn
a
l
o
f
El
e
c
t
ri
c
a
l
E
n
g
i
n
e
e
r
i
n
g
a
n
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
(
I
J
E
EC
S
)
,
v
o
l
.
3
5
,
n
o
.
2
,
p
p
.
1
1
1
0
–
1
1
2
1
,
A
u
g
.
2
0
2
4
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
e
e
c
s
.
v
3
5
.
i
2
.
p
p
1
1
1
0
-
1
1
2
1
.
[
8
]
S
.
S
.
R
e
d
d
y
,
N
.
P
i
l
l
i
,
P
.
V
o
o
s
a
l
a
,
a
n
d
S
.
R
.
C
h
i
g
u
r
u
p
a
t
i
,
“
A
c
o
m
p
a
r
a
t
i
v
e
st
u
d
y
t
o
p
r
e
d
i
c
t
b
r
e
a
s
t
c
a
n
c
e
r
u
si
n
g
mac
h
i
n
e
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s,
”
I
n
d
o
n
e
s
i
a
n
J
o
u
r
n
a
l
o
f
El
e
c
t
r
i
c
a
l
E
n
g
i
n
e
e
ri
n
g
a
n
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
(
I
J
EE
C
S
)
,
v
o
l
.
2
7
,
n
o
.
1
,
p
p
.
1
7
1
–
1
8
0
,
Ju
l
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
e
e
c
s
.
v
2
7
.
i
1
.
p
p
1
7
1
-
1
8
0
.
[
9
]
Z.
S
i
d
e
k
,
S
.
S
.
S
.
A
h
ma
d
,
a
n
d
N
.
H
.
I
.
Te
o
,
“
A
sso
c
i
a
t
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
a
n
d
t
h
e
n
e
w
s
h
e
a
d
l
i
n
e
s
se
n
t
i
me
n
t
f
o
r
B
u
r
sa
s
t
o
c
k
p
r
i
c
e
p
r
e
d
i
c
t
i
o
n
,
”
I
n
d
o
n
e
si
a
n
J
o
u
rn
a
l
o
f
El
e
c
t
r
i
c
a
l
En
g
i
n
e
e
r
i
n
g
a
n
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
(
I
J
EE
C
S
)
,
v
o
l
.
3
1
,
n
o
.
2
,
p
p
.
1
0
4
1
–
1
0
4
9
,
A
u
g
.
2
0
2
3
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
e
e
c
s
.
v
3
1
.
i
2
.
p
p
1
0
4
1
-
1
0
4
9
.
[
1
0
]
Y
.
S
.
A
.
El
W
a
h
a
b
,
M
.
M
.
N
a
sr
,
a
n
d
F
.
K
.
A
l
S
h
e
r
e
f
,
“
A
n
i
n
t
e
l
l
i
g
e
n
t
o
i
l
a
c
c
i
d
e
n
t
p
r
e
d
i
c
t
i
n
g
a
n
d
c
l
a
ss
i
f
y
i
n
g
s
y
st
e
m
u
s
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
s
,
”
I
n
d
o
n
e
si
a
n
J
o
u
r
n
a
l
o
f
El
e
c
t
r
i
c
a
l
En
g
i
n
e
e
ri
n
g
a
n
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
(
I
J
E
EC
S
)
,
v
o
l
.
2
9
,
n
o
.
1
,
p
p
.
4
6
0
–
4
7
1
,
Jan
.
2
0
2
3
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
e
e
c
s
.
v
2
9
.
i
1
.
p
p
4
6
0
-
4
7
1
.
[
1
1
]
J.
F
e
i
,
N
.
Z
h
a
o
,
Y
.
S
h
i
,
Y
.
F
e
n
g
,
a
n
d
Z.
W
a
n
g
,
“
C
o
mp
r
e
ss
o
r
p
e
r
f
o
r
ma
n
c
e
p
r
e
d
i
c
t
i
o
n
u
s
i
n
g
a
n
o
v
e
l
f
e
e
d
-
f
o
r
w
a
r
d
n
e
u
r
a
l
n
e
t
w
o
r
k
b
a
s
e
d
o
n
G
a
u
ssi
a
n
k
e
r
n
e
l
f
u
n
c
t
i
o
n
,
”
A
d
v
a
n
c
e
s
i
n
M
e
c
h
a
n
i
c
a
l
En
g
i
n
e
e
ri
n
g
,
v
o
l
.
8
,
n
o
.
1
,
p
.
1
6
8
7
8
1
4
0
1
6
6
2
8
3
9
6
,
2
0
1
6
,
d
o
i
:
1
0
.
1
1
7
7
/
1
6
8
7
8
1
4
0
1
6
6
2
8
3
9
6
.
[
1
2
]
I
.
Y
a
z
a
r
,
H
.
S
.
Y
a
v
u
z
,
a
n
d
A
.
A
.
Y
a
v
u
z
,
“
C
o
mp
a
r
i
s
o
n
o
f
v
a
r
i
o
u
s
r
e
g
r
e
ssi
o
n
m
o
d
e
l
s
f
o
r
p
r
e
d
i
c
t
i
n
g
c
o
m
p
r
e
sso
r
a
n
d
t
u
r
b
i
n
e
p
e
r
f
o
r
m
a
n
c
e
p
a
r
a
me
t
e
r
s,
”
E
n
e
rg
y
,
v
o
l
.
1
4
0
,
p
p
.
1
3
9
8
–
1
4
0
6
,
2
0
1
7
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
n
e
r
g
y
.
2
0
1
7
.
0
5
.
0
6
1
.
[
1
3
]
Z.
L
i
u
a
n
d
I
.
A
.
K
a
r
i
mi
,
“
G
a
s
t
u
r
b
i
n
e
p
e
r
f
o
r
m
a
n
c
e
p
r
e
d
i
c
t
i
o
n
v
i
a
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
,
”
E
n
e
r
g
y
,
v
o
l
.
1
9
2
,
p
.
1
1
6
6
2
7
,
F
e
b
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
n
e
r
g
y
.
2
0
1
9
.
1
1
6
6
2
7
.
[
1
4
]
A
.
V
.
T
.
J
a
y
a
c
h
a
n
d
r
a
n
,
H
.
O
m
a
r
,
A
.
T
k
a
c
h
e
n
k
o
,
a
n
d
A
.
K
r
i
s
h
n
a
k
u
m
a
r
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
p
r
e
d
i
c
t
o
r
f
o
r
mi
c
r
o
g
a
s
t
u
r
b
i
n
e
p
e
r
f
o
r
m
a
n
c
e
e
v
a
l
u
a
t
i
o
n
,
”
A
e
ro
n
a
u
t
i
c
s
a
n
d
A
e
r
o
sp
a
c
e
O
p
e
n
A
c
c
e
ss
J
o
u
r
n
a
l
,
v
o
l
.
4
,
n
o
.
4
,
p
p
.
1
7
2
–
1
8
0
,
D
e
c
.
2
0
2
0
,
d
o
i
:
1
0
.
1
5
4
0
6
/
a
a
o
a
j
.
2
0
2
0
.
0
4
.
0
0
1
2
0
.
[
1
5
]
K
.
G
h
o
r
b
a
n
i
a
n
a
n
d
M
.
G
h
o
l
a
mr
e
z
a
e
i
,
“
A
n
a
r
t
i
f
i
c
i
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
a
p
p
r
o
a
c
h
t
o
c
o
m
p
r
e
ss
o
r
p
e
r
f
o
r
m
a
n
c
e
p
r
e
d
i
c
t
i
o
n
,
”
Ap
p
l
i
e
d
E
n
e
r
g
y
,
v
o
l
.
8
6
,
n
o
.
7
–
8
,
p
p
.
1
2
1
0
–
1
2
2
1
,
Ju
l
.
2
0
0
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
a
p
e
n
e
r
g
y
.
2
0
0
8
.
0
6
.
0
0
6
.
[
1
6
]
M
.
G
h
o
l
a
mr
e
z
a
e
i
a
n
d
K
.
G
h
o
r
b
a
n
i
a
n
,
“
A
p
p
l
i
c
a
t
i
o
n
o
f
i
n
t
e
g
r
a
t
e
d
f
u
z
z
y
l
o
g
i
c
a
n
d
n
e
u
r
a
l
n
e
t
w
o
r
k
s t
o
t
h
e
p
e
r
f
o
r
m
a
n
c
e
p
r
e
d
i
c
t
i
o
n
o
f
a
x
i
a
l
c
o
m
p
r
e
sso
r
s
,
”
Pro
c
e
e
d
i
n
g
s
o
f
t
h
e
I
n
st
i
t
u
t
i
o
n
o
f
M
e
c
h
a
n
i
c
a
l
E
n
g
i
n
e
e
rs,
P
a
r
t
A:
J
o
u
rn
a
l
o
f
P
o
w
e
r
a
n
d
En
e
r
g
y
,
v
o
l
.
2
2
9
,
n
o
.
8
,
p
p
.
9
2
8
–
9
4
7
,
D
e
c
.
2
0
1
5
,
d
o
i
:
1
0
.
1
1
7
7
/
0
9
5
7
6
5
0
9
1
5
5
9
6
8
7
7
.
[
1
7
]
W
.
W
a
n
g
,
Y
.
Y
a
o
,
a
n
d
Z
.
M
a
,
“
M
o
d
e
l
o
f
c
o
mp
r
e
s
so
r
p
e
r
f
o
r
m
a
n
c
e
p
r
e
d
i
c
t
i
o
n
b
a
s
e
d
o
n
e
r
r
o
r
b
a
c
k
-
p
r
o
p
a
g
a
t
i
o
n
a
r
t
i
f
i
c
i
a
l
n
e
u
r
a
l
w
o
r
k
,
”
F
l
u
i
d
M
a
c
h
,
v
o
l
.
3
3
,
n
o
.
4
,
2
0
0
5
.
[
1
8
]
P
.
C
.
V
i
l
a
l
t
a
,
H
.
W
a
n
,
a
n
d
S
.
S
.
P
a
t
n
a
i
k
,
“
C
e
n
t
r
i
f
u
g
a
l
c
o
m
p
r
e
ss
o
r
p
e
r
f
o
r
m
a
n
c
e
p
r
e
d
i
c
t
i
o
n
u
si
n
g
g
a
u
ssi
a
n
p
r
o
c
e
s
s
r
e
g
r
e
s
si
o
n
a
n
d
a
r
t
i
f
i
c
i
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s,”
AS
ME
I
n
t
e
rn
a
t
i
o
n
a
l
Me
c
h
a
n
i
c
a
l
E
n
g
i
n
e
e
ri
n
g
C
o
n
g
r
e
ss
a
n
d
Ex
p
o
s
i
t
i
o
n
,
Pr
o
c
e
e
d
i
n
g
s
(
I
ME
C
E)
,
v
o
l
.
5
9
4
5
2
,
p
.
V
0
0
8
T
0
9
A
0
4
5
,
2
0
1
9
,
d
o
i
:
1
0
.
1
1
1
5
/
I
M
EC
E
2
0
1
9
-
1
1
9
3
6
.
[
1
9
]
J.
M
a
,
X
.
D
i
n
g
,
W
.
T
.
H
o
r
t
o
n
,
a
n
d
D
.
Zi
v
i
a
n
i
,
“
D
e
v
e
l
o
p
me
n
t
o
f
a
n
a
u
t
o
m
a
t
e
d
c
o
m
p
r
e
ss
o
r
p
e
r
f
o
r
m
a
n
c
e
m
a
p
p
i
n
g
u
si
n
g
a
r
t
i
f
i
c
i
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
a
n
d
mu
l
t
i
p
l
e
c
o
m
p
r
e
sso
r
t
e
c
h
n
o
l
o
g
i
e
s,”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
Re
f
ri
g
e
r
a
t
i
o
n
,
v
o
l
.
1
2
0
,
p
p
.
6
6
–
8
0
,
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
i
j
r
e
f
r
i
g
.
2
0
2
0
.
0
8
.
0
0
1
.
[
2
0
]
O
.
C
o
r
t
é
s,
G
.
U
r
q
u
i
z
a
,
a
n
d
J.
A
.
H
e
r
n
á
n
d
e
z
,
“
O
p
t
i
m
i
z
a
t
i
o
n
o
f
o
p
e
r
a
t
i
n
g
c
o
n
d
i
t
i
o
n
s
f
o
r
c
o
m
p
r
e
ss
o
r
p
e
r
f
o
r
m
a
n
c
e
b
y
me
a
n
s
o
f
n
e
u
r
a
l
n
e
t
w
o
r
k
i
n
v
e
r
se,
”
A
p
p
l
i
e
d
En
e
rg
y
,
v
o
l
.
8
6
,
n
o
.
1
1
,
p
p
.
2
4
8
7
–
2
4
9
3
,
2
0
0
9
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
a
p
e
n
e
r
g
y
.
2
0
0
9
.
0
3
.
0
0
1
.
[
2
1
]
Y
.
Y
u
,
L.
C
h
e
n
,
F
.
S
u
n
,
a
n
d
C
.
W
u
,
“
N
e
u
r
a
l
-
n
e
t
w
o
r
k
b
a
s
e
d
a
n
a
l
y
si
s
a
n
d
p
r
e
d
i
c
t
i
o
n
o
f
a
c
o
m
p
r
e
ss
o
r
’
s
c
h
a
r
a
c
t
e
r
i
st
i
c
p
e
r
f
o
r
ma
n
c
e
map
,
”
Ap
p
l
i
e
d
En
e
rg
y
,
v
o
l
.
8
4
,
n
o
.
1
,
p
p
.
4
8
–
5
5
,
2
0
0
7
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
a
p
e
n
e
r
g
y
.
2
0
0
6
.
0
4
.
0
0
5
.
[
2
2
]
T.
P
a
l
m
é
,
P
.
W
a
n
i
c
z
e
k
,
H
.
H
ö
n
e
n
,
M
.
A
ssa
d
i
,
a
n
d
P
.
J
e
sc
h
k
e
,
“
C
o
m
p
r
e
ss
o
r
ma
p
p
r
e
d
i
c
t
i
o
n
b
y
n
e
u
r
a
l
n
e
t
w
o
r
k
s,
”
J
o
u
rn
a
l
o
f
e
n
e
r
g
y
a
n
d
p
o
w
e
r
e
n
g
i
n
e
e
ri
n
g
,
v
o
l
.
6
,
n
o
.
1
0
,
p
p
.
1
6
5
1
–
1
6
6
2
,
2
0
1
2
.
[
2
3
]
L.
Z
h
o
n
g
,
R
.
Li
u
,
X
.
M
i
a
o
,
Y
.
C
h
e
n
,
S
.
Li
,
a
n
d
H
.
J
i
,
“
C
o
m
p
r
e
ss
o
r
p
e
r
f
o
r
man
c
e
p
r
e
d
i
c
t
i
o
n
b
a
s
e
d
o
n
t
h
e
i
n
t
e
r
p
o
l
a
t
i
o
n
met
h
o
d
a
n
d
su
p
p
o
r
t
v
e
c
t
o
r
m
a
c
h
i
n
e
,
”
Ae
r
o
sp
a
c
e
,
v
o
l
.
1
0
,
n
o
.
6
,
p
.
5
5
8
,
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
a
e
r
o
s
p
a
c
e
1
0
0
6
0
5
5
8
.
[
2
4
]
J.
M
a
r
x
,
S
.
G
a
n
t
n
e
r
,
J.
S
t
ä
d
i
n
g
,
a
n
d
J.
F
r
i
e
d
r
i
c
h
s,
“
A
mac
h
i
n
e
l
e
a
r
n
i
n
g
b
a
s
e
d
a
p
p
r
o
a
c
h
o
f
p
e
r
f
o
r
m
a
n
c
e
e
st
i
mat
i
o
n
f
o
r
h
i
g
h
-
p
r
e
ss
u
r
e
c
o
mp
r
e
sso
r
a
i
r
f
o
i
l
s,”
T
u
r
b
o
E
x
p
o
:
Po
w
e
r
f
o
r
L
a
n
d
,
S
e
a
,
a
n
d
A
i
r
,
v
o
l
.
5
1
0
2
9
,
p
.
V
0
2
D
T4
6
A
0
0
4
,
2
0
1
8
,
d
o
i
:
1
0
.
1
1
1
5
/
g
t
2
0
1
8
-
7
5
3
7
2
.
[
2
5
]
J.
H
.
F
r
i
e
d
m
a
n
,
“
G
r
e
e
d
y
f
u
n
c
t
i
o
n
a
p
p
r
o
x
i
mat
i
o
n
:
a
g
r
a
d
i
e
n
t
b
o
o
st
i
n
g
ma
c
h
i
n
e
,
”
A
n
n
a
l
s
o
f
S
t
a
t
i
st
i
c
s
,
v
o
l
.
2
9
,
n
o
.
5
,
p
p
.
1
1
8
9
–
1
2
3
2
,
2
0
0
1
,
d
o
i
:
1
0
.
1
2
1
4
/
a
o
s/
1
0
1
3
2
0
3
4
5
1
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
K
u
o
-
Chi
e
n
Li
a
o
c
u
rre
n
t
ly
se
rv
in
g
a
s
a
n
As
sista
n
t
P
ro
fe
ss
o
r
in
th
e
De
p
a
rtme
n
t
o
f
Ae
ro
n
a
u
ti
c
a
l
En
g
i
n
e
e
rin
g
a
t
C
h
a
o
y
a
n
g
U
n
iv
e
rsit
y
o
f
Tec
h
n
o
lo
g
y
,
Taiwa
n
.
M
y
a
c
a
d
e
m
ic
jo
u
r
n
e
y
b
e
g
a
n
a
t
t
h
e
Na
ti
o
n
a
l
Ch
in
-
Yi
U
n
iv
e
rsit
y
o
f
Tec
h
n
o
l
o
g
y
,
wh
e
re
he
p
u
rs
u
e
d
h
is
u
n
d
e
r
g
ra
d
u
a
te
a
n
d
M
a
ste
r
’
s
d
e
g
re
e
s
in
In
d
u
strial
E
n
g
i
n
e
e
rin
g
a
n
d
M
a
n
a
g
e
m
e
n
t.
F
u
e
ll
e
d
b
y
a
p
a
ss
io
n
f
o
r
e
n
g
in
e
e
rin
g
a
n
d
a
d
riv
e
f
o
r
c
o
n
ti
n
u
o
u
s
lea
rn
i
n
g
,
he
fu
rth
e
re
d
h
is
stu
d
ies
a
n
d
o
b
tai
n
e
d
a
P
h
.
D.
i
n
M
e
c
h
a
n
ica
l
En
g
i
n
e
e
rin
g
fr
o
m
Th
e
Na
ti
o
n
a
l
Ch
u
n
g
Hs
in
g
Un
i
v
e
rsity
,
Taiwa
n
.
He
is
c
o
m
m
it
ted
to
im
p
a
rti
n
g
k
n
o
wle
d
g
e
,
fo
ste
ri
n
g
c
rit
ica
l
th
in
k
in
g
,
a
n
d
in
s
p
iri
n
g
in
n
o
v
a
ti
o
n
a
m
o
n
g
h
is
stu
d
e
n
ts.
His
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
a
ir
c
ra
ft
m
a
in
ten
a
n
c
e
,
tu
rb
o
fa
n
e
n
g
in
e
re
p
a
ir,
m
u
lt
i
-
a
x
is
u
n
m
a
n
n
e
d
a
e
rial
v
e
h
icle
u
ti
l
iza
ti
o
n
,
a
n
d
o
c
c
u
p
a
ti
o
n
a
l
sa
fe
ty
a
n
d
h
e
a
lt
h
m
a
n
a
g
e
m
e
n
t.
Th
ro
u
g
h
h
i
s
re
se
a
rc
h
,
he
a
im
t
o
a
d
d
re
ss
c
o
n
tem
p
o
ra
ry
c
h
a
l
len
g
e
s
i
n
a
e
ro
n
a
u
ti
c
a
l
e
n
g
i
n
e
e
rin
g
a
n
d
c
o
n
tri
b
u
te
t
o
d
e
v
e
l
o
p
i
n
g
su
sta
i
n
a
b
le
so
lu
ti
o
n
s
fo
r
t
h
e
fu
t
u
re
.
He
is
d
e
e
p
l
y
i
n
v
e
ste
d
in
t
h
e
a
c
a
d
e
m
ic
c
o
m
m
u
n
it
y
a
n
d
is
a
lwa
y
s
e
a
g
e
r
to
c
o
ll
a
b
o
ra
te
a
n
d
e
x
c
h
a
n
g
e
id
e
a
s
with
fe
ll
o
w
re
se
a
rc
h
e
rs
a
n
d
p
ro
fe
ss
io
n
a
ls.
F
o
r
a
n
y
in
q
u
iri
e
s
o
r
c
o
l
lab
o
ra
ti
o
n
o
p
p
o
rt
u
n
it
ies
,
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
jam
e
s1
9
8
3
1
1
1
1
@
g
m
a
il
.
c
o
m
.
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
.
2
,
Feb
r
u
a
r
y
20
2
5
:
1
20
1
-
1
20
8
1208
H
o
m
-
Y
u
W
u
re
c
e
iv
e
d
t
h
e
B.
S
c
.
d
e
g
re
e
in
De
p
a
rtme
n
t
o
f
Ae
ro
n
a
u
ti
c
a
l
En
g
i
n
e
e
rin
g
fro
m
t
h
e
Ch
u
n
g
C
h
e
n
g
I
n
stit
u
te
o
f
Tec
h
n
o
l
o
g
y
(CCIT
),
Taiwa
n
,
th
e
M
.
S
c
.
d
e
g
re
e
in
Wea
p
o
n
sy
ste
m
e
n
g
i
n
e
e
rin
g
fr
o
m
t
h
e
Na
ti
o
n
a
l
De
fe
n
se
U
n
iv
e
rs
it
y
(ND
U),
Taiwa
n
,
a
n
d
th
e
P
h
.
D.
d
e
g
re
e
i
n
S
c
h
o
o
l
o
f
De
fe
n
se
S
c
ien
c
e
fro
m
t
h
e
Na
ti
o
n
a
l
De
fe
n
se
Un
iv
e
rsity
(ND
U),
Taiwa
n
.
He
is
a
n
a
ss
istan
t
p
ro
fe
ss
o
r
in
t
h
e
De
p
a
rtme
n
t
o
f
M
e
c
h
a
n
ica
l
En
g
i
n
e
e
rin
g
,
L
u
n
g
h
wa
Un
iv
e
rsity
o
f
S
c
ien
c
e
a
n
d
Tec
h
n
o
l
o
g
y
,
Taiwa
n
.
His
re
se
a
rc
h
i
n
tere
sts
in
c
l
u
d
e
a
e
ro
e
n
g
i
n
e
p
e
rfo
rm
a
n
c
e
a
n
a
ly
sis,
jet
p
ro
p
u
lsi
o
n
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
wu
h
o
m
y
u
@g
m
a
il
.
c
o
m
.
H
u
n
g
-
Ta
We
n
re
c
e
iv
e
d
t
h
e
B.
S
c
.
d
e
g
re
e
in
De
p
a
rtme
n
t
o
f
Ae
ro
n
a
u
ti
c
a
l
En
g
i
n
e
e
rin
g
fro
m
t
h
e
Ch
u
n
g
C
h
e
n
g
I
n
stit
u
te
o
f
Tec
h
n
o
l
o
g
y
(CCIT
),
Taiwa
n
,
th
e
M
.
S
c
.
d
e
g
re
e
in
Wea
p
o
n
S
y
ste
m
En
g
in
e
e
rin
g
fr
o
m
th
e
Na
ti
o
n
a
l
De
fe
n
se
Un
iv
e
rsity
(ND
U),
Taiwa
n
,
a
n
d
t
h
e
P
h
.
D.
d
e
g
re
e
in
M
e
c
h
a
n
ica
l
En
g
i
n
e
e
rin
g
fro
m
Th
e
Na
ti
o
n
a
l
C
h
u
n
g
Hs
in
g
Un
i
v
e
rsity
,
Taiwa
n
.
He
is
a
n
a
ss
istan
t
p
ro
fe
ss
o
r
in
th
e
De
p
a
rtme
n
t
o
f
Ae
ro
n
a
u
t
ica
l
En
g
i
n
e
e
rin
g
,
C
h
a
o
y
a
n
g
Un
iv
e
rsity
o
f
S
c
ien
c
e
a
n
d
Tec
h
n
o
l
o
g
y
,
Taiwa
n
.
His
re
se
a
rc
h
i
n
tere
sts
in
c
l
u
d
e
a
e
ro
e
n
g
i
n
e
p
e
rfo
rm
a
n
c
e
a
n
a
ly
sis
,
m
a
c
h
in
e
lea
rn
in
g
,
d
e
e
p
lea
rn
in
g
,
a
n
d
d
ies
e
l
e
n
g
in
e
e
m
issio
n
p
re
d
ictio
n
.
H
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
we
n
h
u
n
g
ta
@
g
m
a
il
.
c
o
m
.
J
ui
-
Ta
ng
Su
n
g
Ob
tain
e
d
a
b
a
c
h
e
lo
r
’
s
d
e
g
re
e
fro
m
t
h
e
De
p
a
rtme
n
t
o
f
M
e
c
h
a
n
ica
l
En
g
i
n
e
e
rin
g
o
f
Ch
u
n
g
Ch
e
n
g
I
n
stit
u
te
o
f
Tec
h
n
o
lo
g
y
(CCIT)
i
n
T
a
iwa
n
,
a
m
a
ste
r
’
s
d
e
g
re
e
in
in
d
u
strial
e
n
g
in
e
e
rin
g
fro
m
C
h
u
n
g
Yu
a
n
C
h
risti
a
n
U
n
iv
e
rsit
y
(
CY
C
U),
a
n
d
a
d
o
c
to
ra
l
st
u
d
e
n
t
fro
m
th
e
S
c
h
o
o
l
o
f
M
a
n
a
g
e
m
e
n
t
o
f
Ch
u
n
g
H
u
a
Un
iv
e
rsit
y
(
CH
U)
in
Taiwa
n
.
His
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
h
u
m
a
n
fa
c
to
rs
e
n
g
i
n
e
e
rin
g
a
n
d
v
a
rio
u
s
AI
-
re
late
d
d
o
m
a
i
n
s.
F
o
r
fu
r
th
e
r
in
q
u
iry
,
h
e
m
a
y
b
e
re
a
c
h
e
d
a
t
rts
o
u
n
g
2
@
g
m
a
il
.
c
o
m
.
Mu
h
a
m
a
d
H
id
a
y
a
t
re
c
e
iv
e
d
t
h
e
Ph
.
D
.
i
n
Co
n
stru
c
ti
o
n
En
g
i
n
e
e
rin
g
fro
m
Taiwa
n
Ch
a
o
y
a
n
g
Un
i
v
e
rsity
o
f
Tec
h
n
o
lo
g
y
i
n
Ju
n
e
2
0
2
4
.
Cu
rre
n
tl
y
tea
c
h
i
n
g
a
t
th
e
De
p
a
rtme
n
t
o
f
M
e
c
h
a
n
ica
l
E
n
g
i
n
e
e
rin
g
,
S
u
m
b
a
wa
Un
i
v
e
rsity
o
f
Tec
h
n
o
lo
g
y
.
His
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
stru
c
tu
re
a
n
d
d
e
sig
n
i
n
e
n
g
in
e
e
rin
g
.
H
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
m
u
h
a
m
a
d
.
h
i
d
a
y
a
t@u
ts.ac
.
i
d
.
Wil
l
We
i
-
J
u
e
n
W
a
n
g
C
u
rre
n
tl
y
p
u
rs
u
in
g
a
M
a
ste
r
’
s
d
e
g
re
e
in
El
e
c
tri
c
a
l
a
n
d
Co
m
p
u
ter
E
n
g
i
n
e
e
rin
g
a
t
t
h
e
Un
iv
e
rsit
y
o
f
Ca
li
fo
r
n
ia,
Da
v
is.
He
re
c
e
iv
e
d
h
is
Ba
c
h
e
lo
r
’
s
d
e
g
re
e
in
In
f
o
rm
a
ti
o
n
E
n
g
i
n
e
e
rin
g
fr
o
m
S
h
a
n
g
h
a
i
Jia
o
To
n
g
Un
i
v
e
rsity
,
S
h
a
n
g
h
a
i,
Ch
i
n
a
.
His
re
se
a
rc
h
in
tere
st
li
e
s
in
th
e
field
s
o
f
ro
b
o
ti
c
s
a
n
d
c
o
m
p
u
ter
v
isi
o
n
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
ww
wa
@u
c
d
a
v
is.ed
u
.
Evaluation Warning : The document was created with Spire.PDF for Python.