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.
24
,
No
.
3
,
Dec
em
b
er
2
0
2
1
,
p
p
.
1
6
1
0
~
1
6
1
7
I
SS
N:
2
5
0
2
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/ijeecs.v
24
.i
3
.
pp
1
6
1
0
-
1
6
1
7
1610
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
Para
llel
ex
trem
e
g
ra
dient
bo
o
sting
cla
ss
ifier
for
lun
g
ca
ncer
detec
tion
Ra
na
Dhia
a
Abdu
-
A
lj
a
ba
r
,
O
s
a
m
a
A.
Awa
d
F
a
c
u
lt
y
of
I
n
fo
rm
a
ti
o
n
E
n
g
i
n
e
e
rin
g
,
Al
-
Na
h
ra
in
U
n
iv
e
rsit
y
,
Ba
g
h
d
a
d
,
Ira
q
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
n
2
9
,
2
0
2
1
R
ev
is
ed
Oct
27
,
2
0
2
1
Acc
ep
ted
No
v
1
,
2
0
2
1
M
o
st
lu
n
g
c
a
n
c
e
rs
do
n
o
t
c
a
u
se
sy
m
p
to
m
s
u
n
ti
l
t
h
e
d
ise
a
se
is
in
its
late
r
sta
g
e
.
Th
a
t
led
th
e
lu
n
g
c
a
n
c
e
r
ha
v
in
g
a
h
ig
h
fa
talit
y
ra
te
c
o
m
p
a
r
ed
to
o
th
e
r
c
a
n
c
e
r
ty
p
e
s.
M
a
n
y
sc
ien
ti
sts
try
to
u
se
a
rti
f
icia
l
i
n
telli
g
e
n
c
e
a
lg
o
rit
h
m
s
to
p
ro
d
u
c
e
a
c
c
u
ra
te
lu
n
g
c
a
n
c
e
r
d
e
tec
ti
o
n
.
Th
is
p
a
p
e
r
u
se
d
e
x
trem
e
g
ra
d
ien
t
b
o
o
st
in
g
(XG
Bo
o
st)
m
o
d
e
ls
as
a
b
a
se
m
o
d
e
l
fo
r
its
e
ffe
c
ti
v
e
n
e
ss
.
It
e
n
h
a
n
c
e
d
lu
n
g
c
a
n
c
e
r
d
e
tec
ti
o
n
p
e
rfo
rm
a
n
c
e
by
s
u
g
g
e
stin
g
t
h
re
e
sta
g
e
s
m
o
d
e
l;
fe
a
tu
re
sta
g
e
,
XG
Bo
o
ste
p
a
ra
ll
e
l
sta
g
e
a
n
d
se
lec
ti
o
n
sta
g
e
.
Th
is
stu
d
y
u
se
d
two
ty
p
e
s
of
g
e
n
e
e
x
p
re
ss
io
n
d
a
tas
e
ts;
RNA
-
se
q
u
e
n
c
e
a
n
d
m
icro
a
rra
y
p
r
o
fil
e
s.
Th
e
re
su
lt
s
p
re
se
n
ted
t
h
e
e
ffe
c
ti
v
e
n
e
ss
of
t
h
e
p
ro
p
o
se
d
m
o
d
e
l,
e
s
p
e
c
ially
in
d
e
a
li
n
g
wi
th
imb
a
la
n
c
e
d
d
a
tas
e
ts,
by
h
a
v
i
n
g
1
0
0
%
each
of
s
e
n
siti
v
it
y
,
sp
e
c
ifi
c
it
y
,
p
re
c
isio
n
,
F
1
_
sc
o
re
,
a
re
a
u
n
d
e
r
c
u
rv
e
(AU
C),
a
n
d
a
c
c
u
ra
c
y
m
e
tri
c
s wh
e
n
it
a
p
p
li
e
d
o
n
a
ll
of
t
h
e
d
a
tas
e
ts
u
se
d
i
n
t
h
is stu
d
y
.
K
ey
w
o
r
d
s
:
B
io
in
f
o
r
m
atics
Gen
e
ex
p
r
ess
io
n
L
u
n
g
ca
n
ce
r
d
is
ea
s
e
Ma
ch
in
e
lear
n
in
g
XGBo
o
s
t
T
h
is
is
an
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r
th
e
CC
BY
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
R
an
a
Dh
iaa
Ab
d
u
-
A
ljab
ar
Dep
ar
tm
en
t
of
Sy
s
tem
s
E
n
g
in
ee
r
in
g
,
Facu
lty
of
I
n
f
o
r
m
atio
n
E
n
g
in
ee
r
i
n
g
,
Al
-
Nah
r
ain
Un
i
v
er
s
ity
B
ag
h
d
ad
,
I
r
aq
E
m
ail:
r
an
ad
h
iaa1
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
L
u
n
g
ca
n
ce
r
is
co
m
m
o
n
ca
n
ce
r
th
at
ca
u
s
es
a
h
ig
h
er
f
atality
r
ate
b
etwe
en
ca
n
ce
r
ty
p
es.
T
h
e
f
iv
e
-
y
ea
r
s
u
r
v
iv
al
r
ate
is
ab
o
u
t
5
6
%
f
o
r
p
atien
ts
th
at
ca
n
ce
r
is
s
till
in
th
e
lu
n
g
.
W
h
ile
5
%
f
o
r
t
h
e
ca
s
es,
its
ca
n
ce
r
s
p
r
ea
d
o
u
t
o
f
th
e
lu
n
g
.
On
ly
1
6
%
o
f
lu
n
g
ca
n
ce
r
ca
s
es
ar
e
d
etec
ted
ea
r
ly
[
1
]
.
R
ec
o
g
n
itio
n
an
d
p
r
ed
ictio
n
th
e
lu
n
g
ca
n
ce
r
in
th
e
ea
r
lies
t
s
ta
g
e
ca
n
in
cr
ea
s
e
th
e
s
u
r
v
iv
al
r
ate
o
f
th
e
p
atien
ts
.
L
u
n
g
ca
n
c
er
h
as
n
o
s
y
m
p
to
m
s
in
th
e
ea
r
ly
s
tag
es
[
2
]
,
[
3
]
,
s
o
it
n
ee
d
s
m
o
r
e
t
h
an
tr
a
d
itio
n
al
d
etec
tio
n
to
d
etec
t
it.
C
an
ce
r
ca
n
b
e
d
ef
in
ed
as
a
d
is
ea
s
e
o
f
alter
ed
g
en
e
e
x
p
r
ess
io
n
.
T
h
e
d
e
v
elo
p
m
e
n
t
o
f
g
en
e
e
x
p
r
ess
io
n
tec
h
n
o
lo
g
i
es
h
as
b
ec
o
m
e
t
h
e
s
tan
d
ar
d
tech
n
o
lo
g
y
f
o
r
s
tu
d
y
in
g
th
e
ce
lls
[
4
]
-
[
6
]
.
T
h
e
d
ev
elo
p
m
en
t
o
f
th
is
tech
n
o
lo
g
y
m
ad
e
m
an
y
r
esear
ch
er
s
ap
p
ly
m
a
n
y
s
tu
d
ies
o
n
im
p
r
o
v
in
g
lu
n
g
ca
n
ce
r
p
r
e
d
ictio
n
b
y
an
aly
zi
n
g
t
h
e
ch
an
g
es
in
g
en
e
ex
p
r
ess
io
n
.
So
m
e
r
esear
ch
er
s
s
tu
d
y
g
en
e
ex
p
r
ess
io
n
-
b
ased
p
r
o
g
n
o
s
tic
s
ig
n
atu
r
es
f
o
r
lu
n
g
ca
n
ce
r
[
3
]
.
Oth
er
s
tr
y
to
u
s
e
g
en
e
ex
p
r
ess
io
n
t
ec
h
n
o
lo
g
y
s
u
c
h
as
m
icr
o
ar
r
a
y
an
d
R
NA
-
s
eq
u
en
ce
s
to
d
e
v
e
lo
p
l
u
n
g
ca
n
ce
r
d
etec
tio
n
m
eth
o
d
s
.
Ma
n
y
s
tu
d
ies
u
s
ed
ar
tific
ial
in
tellig
en
ce
to
d
etec
t
lu
n
g
ca
n
ce
r
f
o
r
th
eir
p
o
wer
to
o
ls
.
T
h
ey
u
s
ed
d
if
f
er
en
t
m
eth
o
d
s
,
lik
e
Al
-
An
n
i
et
a
l.
[
7
]
-
[
1
2
]
,
th
ey
p
r
o
p
o
s
ed
d
i
f
f
er
en
t
o
p
tim
izatio
n
m
o
d
els
to
en
h
an
ce
th
e
n
o
n
-
s
m
all
ce
ll
l
u
n
g
ca
n
ce
r
(
NSC
L
C
)
d
etec
tio
n
ac
cu
r
ac
y
u
s
in
g
m
icr
o
ar
r
ay
g
e
n
e
ex
p
r
ess
io
n
d
atasets
.
Als
o
,
Azz
awi
et
a
l.
[
1
3
]
-
[
1
6
]
h
a
v
e
m
u
ltip
le
s
tu
d
ies
o
n
m
u
lticlas
s
im
p
r
o
v
em
e
n
t
u
s
in
g
th
e
GE
P
a
lg
o
r
ith
m
i
n
th
e
lu
n
g
ca
n
ce
r
class
if
icatio
n
s
tag
e
to
d
eter
m
in
e
th
e
s
p
ec
if
ic
th
er
ap
y
an
d
r
ed
u
ce
th
e
f
atality
r
ate.
Hu
et
a
l.
[
1
7
]
p
r
o
p
o
s
ed
d
etec
tin
g
an
d
r
ec
o
g
n
izin
g
d
if
f
er
en
t
life
s
tag
es
o
f
b
lad
d
er
ce
lls
u
s
in
g
two
ca
s
c
ad
ed
co
n
v
o
lu
ti
o
n
al
n
eu
r
al
n
etwo
r
k
s
(
C
NNs).
T
o
d
etec
t
ca
n
ce
r
ce
lls
an
d
th
eir
s
tag
es.
W
h
ile
Sar
ic
et
a
l.
[
1
8
]
p
r
o
p
o
s
ed
a
f
u
l
ly
au
to
m
atic
m
eth
o
d
f
o
r
d
etec
tin
g
lu
n
g
ca
n
ce
r
i
n
lu
n
g
tis
s
u
e.
T
h
ey
u
s
ed
two
co
n
v
o
lu
tio
n
al
n
e
u
r
al
n
etwo
r
k
C
NN
ar
ch
itectu
r
es
(
VGG
an
d
R
esNet)
f
o
r
tr
ain
i
n
g
,
a
n
d
th
eir
p
er
f
o
r
m
an
ce
is
co
m
p
ar
ed
.
T
h
e
r
es
u
lts
o
b
tain
ed
s
h
o
w
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
P
a
r
a
llel e
xtreme
g
r
a
d
ien
t b
o
o
s
tin
g
cla
s
s
ifier
fo
r
lu
n
g
ca
n
ce
r
d
etec
tio
n
(
R
a
n
a
D
h
ia
’
a
A
b
d
u
a
lja
b
a
r
)
1611
th
at
th
e
C
NN
-
b
ased
ap
p
r
o
ac
h
ca
n
h
elp
p
ath
o
lo
g
is
ts
d
iag
n
o
s
e
lu
n
g
ca
n
ce
r
.
Als
o
,
L
i
et
a
l.
[
1
9
]
p
r
o
p
o
s
ed
a
f
u
s
io
n
alg
o
r
ith
m
th
at
c
o
m
b
in
e
s
h
an
d
cr
a
f
ted
f
ea
tu
r
es
i
n
to
th
e
f
ea
tu
r
es
lea
r
n
ed
at
t
h
e
o
u
tp
u
t
lay
er
o
f
a
3
D
d
ee
p
co
n
v
o
l
u
tio
n
al
n
eu
r
al
n
etwo
r
k
(
C
NN)
.
Patr
a
[
2
0
]
an
aly
ze
d
v
ar
io
u
s
m
ac
h
in
e
lear
n
in
g
c
lass
if
ier
tech
n
iq
u
es
to
class
if
y
lu
n
g
ca
n
c
er
in
t
o
b
e
n
i
g
n
an
d
m
alig
n
an
t.
L
ai
et
a
l.
[
2
1
]
tr
ai
n
ed
clin
ical
an
d
g
en
e
ex
p
r
ess
io
n
d
ata
with
im
p
r
o
v
e
d
d
ee
p
n
e
u
r
al
n
etwo
r
k
(
DNN)
.
I
t
u
s
ed
p
atien
ts
b
ased
o
n
m
icr
o
ar
r
ay
d
ata
to
p
r
ed
ict
th
e
5
-
y
ea
r
s
u
r
v
iv
al
s
tatu
s
o
f
NS
C
L
C
.
T
h
e
s
tu
d
y
Priy
a
an
d
J
awh
ar
[
2
2
]
p
r
o
p
o
s
ed
an
a
u
to
m
atic
ap
p
r
o
ac
h
to
class
if
y
in
g
th
e
lu
n
g
im
a
g
e
in
to
a
n
o
r
m
a
l
ca
s
e
o
r
ca
n
ce
r
ca
s
e
b
y
p
r
e
p
r
o
ce
s
s
in
g
th
e
co
m
p
u
ted
t
o
m
o
g
r
ap
h
y
(
CT
)
lu
n
g
im
ag
e
to
r
em
o
v
e
n
o
is
e.
T
h
e
n
co
m
b
in
es
th
e
h
is
to
g
r
am
an
aly
s
is
with
m
o
r
p
h
o
lo
g
ical
a
n
d
ex
tr
ac
ts
th
e
lu
n
g
r
eg
io
n
s
b
y
th
r
esh
o
ld
in
g
o
p
er
a
tio
n
s
,
wh
ile
Og
u
n
ley
e
an
d
W
an
g
[
2
3
]
u
s
ed
a
clin
ical
d
atab
ase
to
clas
s
if
y
th
e
p
atien
t
if
h
e
h
as
ch
r
o
n
ic
k
id
n
ey
d
is
ea
s
e
o
r
n
o
t
u
s
in
g
XGB
o
o
s
t.
Desu
k
y
et
a
l.
[
2
4
]
s
u
g
g
ested
a
n
ew
m
eth
o
d
f
o
r
class
if
icatio
n
to
d
ea
l
with
im
b
alan
ce
d
m
ed
ical
d
atasets
.
I
t
u
s
ed
th
e
cr
o
s
s
o
v
er
to
in
c
r
ea
s
e
th
e
m
in
o
r
ity
class
an
d
th
en
u
s
ed
t
h
e
b
o
o
s
tin
g
f
o
r
class
if
icatio
n
.
T
h
is
m
et
h
o
d
e
n
h
an
c
es
th
e
class
if
icatio
n
ac
cu
r
ac
y
r
esu
lts
.
R
u
s
tam
e
t
a
l.
[
2
5
]
u
s
ed
f
ea
tu
r
e
ex
tr
ac
tio
n
f
r
o
m
C
T
im
a
g
es
as
d
ata
to
class
if
y
lu
n
g
ca
n
c
er
.
Fu
zz
y
C
-
Me
an
s
an
d
f
u
zz
y
k
er
n
el
C
-
Me
an
s
wer
e
u
s
ed
to
class
if
y
th
e
lu
n
g
n
o
d
u
le
f
r
o
m
th
e
p
atien
t
in
to
b
en
ig
n
o
r
m
alig
n
an
t.
T
h
e
s
co
r
e
s
h
o
wed
f
u
zz
y
k
e
r
n
el
C
-
Me
an
s
h
ad
h
ig
h
er
ac
c
u
r
ac
y
th
an
f
u
zz
y
C
-
Me
an
s
ac
cu
r
ac
y
.
Pan
d
ian
et
a
l.
[
2
6
]
d
ev
elo
p
e
d
an
alg
o
r
ith
m
t
o
class
if
y
lu
n
g
ca
n
ce
r
m
ed
ical
im
ag
es
as
n
o
r
m
al
an
d
i
n
f
ec
te
d
.
T
h
e
f
ea
tu
r
es
ar
e
ex
tr
ac
ted
f
r
o
m
C
T
im
ag
es
o
f
n
o
r
m
al
lu
n
g
an
d
ca
n
ce
r
af
f
ec
t
ed
lu
n
g
s
wer
e
tak
en
i
n
to
th
e
s
tu
d
y
.
T
h
e
a
r
tific
ial
n
eu
r
al
n
etwo
r
k
is
u
s
ed
in
clas
s
if
icatio
n
.
Hak
im
et
a
l.
[
2
7
]
c
o
m
p
ar
ed
two
p
o
p
u
la
r
f
ea
tu
r
e
s
elec
tio
n
m
o
d
els
to
en
h
an
ce
th
e
s
u
p
p
o
r
t v
ec
to
r
m
a
ch
in
e
(
SVM
)
ca
n
ce
r
class
if
ica
tio
n
.
T
h
ey
s
h
o
wed
th
at
th
e
R
elief
F o
u
tp
er
f
o
r
m
e
d
co
m
p
ar
ed
with
C
FS
as
m
icr
o
ar
r
ay
d
ata
f
ea
t
u
r
e_
s
elec
tio
n
a
p
p
r
o
ac
h
.
Kar
ee
m
et
a
l.
[
2
8
]
d
ev
elo
p
e
d
th
e
C
T
s
ca
n
n
in
g
d
ata
s
et
u
s
in
g
im
ag
i
n
g
/co
m
p
u
ter
v
is
io
n
al
g
o
r
ith
m
s
f
o
r
d
iar
y
o
f
h
ea
lth
y
a
n
d
tu
m
o
r
o
u
s
c
h
est
s
ca
n
s
;
T
h
is
co
m
p
r
is
es
th
r
ee
p
r
e
p
r
o
ce
s
s
in
g
s
tep
s
:
i)
im
p
r
o
v
em
e
n
t
o
f
im
ag
es,
ii)
s
eg
m
en
tatio
n
o
f
im
ag
es,
an
d
iii)
s
tr
ateg
ies
f
o
r
f
ea
tu
r
e
ex
tr
a
ctio
n
.
I
n
th
e
last
s
tag
e,
a
s
u
p
p
o
r
t v
ec
to
r
m
ac
h
in
e
(
SVM)
is
u
tili
ze
d
to
ca
teg
o
r
ize
s
lid
e
in
s
tan
ce
s
as
o
n
e
o
f
3
ty
p
es
(
n
o
r
m
al,
b
en
i
g
n
,
o
r
m
alig
n
an
t)
b
y
u
s
in
g
class
if
icatio
n
tech
n
o
lo
g
y
.
T
h
e
b
est
ac
cu
r
ac
y
,
8
9
,
8
8
7
6
p
er
c
e
n
t,
was
o
b
tain
ed
wh
en
ap
p
l
y
in
g
th
is
tech
n
iq
u
e
t
o
th
e
n
ew
d
ataset.
Selwal
an
d
R
ao
o
f
[
2
9
]
d
ev
el
o
p
ed
a
MA
T
L
AB
-
b
ased
C
NN
f
o
r
au
to
m
at
ed
d
etec
tio
n
o
f
ca
n
ce
r
o
u
s
ce
r
v
ix
ce
lls
wh
er
e
th
e
tem
p
lates
s
eg
m
en
ted
th
e
n
u
cleu
s
o
f
th
e
ce
lls
.
T
h
e
s
im
u
latio
n
r
esu
lts
s
h
o
w
th
at
th
e
p
r
o
p
o
s
ed
C
NN
alg
o
r
ith
m
ca
n
au
to
m
atica
lly
d
etec
t
th
e
ce
r
v
ix
ca
n
ce
r
ce
lls
with
m
o
r
e
th
an
8
8
%
ac
cu
r
ac
y
.
R
aju
e
t
a
l.
[
3
0
]
u
s
ed
h
ig
h
r
eso
lu
tio
n
co
m
p
u
ter
to
m
o
g
r
a
p
h
y
(
HR
C
T
)
im
ag
es
with
m
u
lti
-
class
if
icatio
n
to
class
if
y
1
7
in
ter
s
titi
al
lu
n
g
d
is
ea
s
es
with
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
(
C
NN)
ar
ch
it
ec
tu
r
e
ca
lled
Sm
aller
VGGN
et.
I
t
o
b
tain
ed
9
5
%
av
er
ag
ed
ac
c
u
r
ac
y
.
Ali
et
a
l.
[3
1
]
d
ee
p
n
e
u
r
al
n
etwo
r
k
s
wer
e
u
s
ed
,
th
at
is
,
th
e
en
h
an
ce
r
D
ee
p
B
elief
Netwo
r
k
(
DB
N)
,
wh
ich
is
co
n
s
tr
u
cted
f
r
o
m
two
R
estricte
d
B
o
ltzm
an
n
Ma
ch
in
es
(
R
B
M)
.
T
h
e
e
n
h
an
ce
r
DB
N
was
tr
ain
ed
b
y
b
ac
k
p
r
o
p
a
g
atio
n
n
eu
r
al
n
etwo
r
k
(
B
PNN)
.
I
t
f
o
u
n
d
th
at
L
ASSO
with
L
R
g
iv
es
th
e
b
est
ac
cu
r
ac
y
in
th
eir
s
tu
d
y
d
ataset.
Ab
d
u
ll
ah
et
a
l.
[
3
2
]
d
ev
elo
p
ed
a
M
AT
L
AB
-
b
ased
C
NN
f
o
r
au
t
o
m
ated
d
etec
tio
n
o
f
ca
n
ce
r
o
u
s
c
er
v
ix
ce
lls
wh
er
e
t
h
e
tem
p
lates
s
eg
m
e
n
ted
th
e
n
u
cleu
s
o
f
th
e
ce
lls
.
T
h
e
p
r
o
p
o
s
ed
C
NN
alg
o
r
ith
m
d
etec
t
ed
th
e
ce
r
v
ix
ca
n
ce
r
ce
lls
au
to
m
atica
lly
with
m
o
r
e
th
a
n
8
8
% a
cc
u
r
ac
y
.
A
p
r
ev
io
u
s
s
tu
d
y
[
3
3
]
co
m
p
a
r
es
m
u
ltip
le
cu
r
r
en
t
m
ac
h
in
e
lear
n
in
g
an
d
f
o
u
n
d
th
at
th
e
XGBo
o
s
t
is
th
e
m
o
s
t
ac
cu
r
ate
s
y
s
tem
in
b
alan
ce
an
d
im
b
alan
ce
d
atasets
.
T
h
is
s
tu
d
y
tr
ied
to
im
p
r
o
v
e
th
e
XGBo
o
s
t
b
y
u
s
in
g
a
f
ea
tu
r
e
s
elec
tio
n
to
u
s
e
o
n
ly
th
e
g
e
n
es
r
esp
o
n
s
ib
le
f
o
r
lu
n
g
ca
n
ce
r
d
is
ea
s
e
in
th
e
lear
n
in
g
s
tag
e
an
d
ap
p
lied
a
p
ar
allel
XGBo
o
s
t
(
PXGB
)
wi
th
d
if
f
er
en
t
h
y
p
e
r
p
ar
am
eter
s
to
in
cr
ea
s
e
th
e
s
y
s
tem
v
ar
iety
an
d
d
ec
r
ea
s
e
th
e
o
v
er
f
itti
n
g
a
n
d
u
n
d
er
f
itti
n
g
.
T
h
e
PXGB
s
h
o
w
ed
m
o
r
e
ac
cu
r
ate
p
r
ed
ictio
n
v
alu
es
f
o
r
d
etec
tin
g
ca
n
ce
r
an
d
n
o
r
m
al
lu
n
g
s
tat
e,
esp
ec
ially
f
o
r
im
b
ala
n
ce
d
d
ata
s
ets.
2.
XG
B
O
O
ST
A
L
G
O
RI
T
H
M
XGBo
o
s
t
is
a
d
ec
is
io
n
-
tr
ee
-
b
ased
en
s
em
b
le
m
ac
h
i
n
e
lear
n
i
n
g
alg
o
r
ith
m
d
ev
elo
p
ed
b
y
T
i
an
q
i
C
h
en
an
d
C
ar
lo
s
Gu
estrin
.
T
h
ey
em
p
lo
y
m
ac
h
in
e
lea
r
n
in
g
alg
o
r
it
h
m
s
u
n
d
er
th
e
g
r
ad
ien
t
b
o
o
s
tin
g
f
r
am
ewo
r
k
.
T
h
ey
in
tr
o
d
u
ce
d
th
eir
wo
r
k
at
th
e
SI
GKDD
co
n
f
er
en
ce
in
2
0
1
6
[
34
]
.
First,
let
u
s
clar
if
y
th
e
co
n
ce
p
t
o
f
b
o
o
s
tin
g
.
I
t
is
a
n
e
n
s
em
b
le
m
eth
o
d
t
h
at
s
ee
k
s
to
cr
ea
te
a
r
o
b
u
s
t
class
if
ier
(
m
o
d
el)
b
ased
o
n
"
wea
k
"
class
if
ier
s
.
I
n
th
is
co
n
t
ex
t,
wea
k
an
d
r
o
b
u
s
t
r
ef
er
to
h
o
w
co
r
r
elate
d
th
e
lear
n
er
s
ar
e
to
th
e
ac
tu
al
tar
g
e
t
v
ar
iab
le.
B
y
ad
d
in
g
m
o
d
els
o
n
to
p
o
f
ea
c
h
o
th
er
iter
ativ
el
y
,
th
e
er
r
o
r
s
o
f
th
e
p
r
ev
io
u
s
m
o
d
el
ar
e
co
r
r
ec
ted
b
y
th
e
n
ex
t p
r
ed
icto
r
,
s
eq
u
en
ti
ally
in
th
e
tr
ain
in
g
s
tag
e,
as it
ap
p
ea
r
ed
in
f
ig
u
r
e1
u
n
til
t
h
e
tr
ain
in
g
d
ata
is
ac
cu
r
ately
p
r
ed
icted
o
r
r
ep
r
o
d
u
ce
d
b
y
th
e
m
o
d
el.
Fin
ally
,
as
s
ee
n
in
F
ig
u
r
e
1
,
it
p
r
o
v
id
es
a
p
ar
allel
tr
ee
b
o
o
s
tin
g
f
o
r
th
e
test
in
g
s
tag
e
th
at
q
u
ick
ly
an
d
ac
cu
r
ately
s
o
lv
es
m
an
y
d
ata
s
cien
ce
p
r
o
b
lem
s
.
I
t
o
f
f
er
s
a
r
an
g
e
o
f
h
y
p
er
p
ar
am
eter
s
th
at
g
iv
e
f
in
e
-
g
r
ain
ed
co
n
tr
o
l
o
v
er
th
e
m
o
d
el
tr
ain
in
g
p
r
o
ce
d
u
r
e.
I
t
wo
r
k
ed
v
e
r
y
well
with
th
e
im
b
alan
ce
d
atab
ase.
I
t
h
ad
m
an
y
f
ea
tu
r
es
s
u
itab
le
f
o
r
lar
g
e
d
ata
b
ases
lik
e;
p
ar
alleliza
tio
n
,
d
is
tr
ib
u
te
d
co
m
p
u
tin
g
,
o
u
t
-
of
-
co
r
e
c
o
m
p
u
tin
g
(
f
o
r
m
a
n
ag
in
g
th
e
la
r
g
e
d
ataset
with
th
e
m
em
o
r
y
)
,
c
ac
h
e
o
p
tim
izatio
n
(
to
th
e
b
est u
s
e
o
f
h
ar
d
war
e)
[
3
5
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
24
,
No
.
3
,
Dec
em
b
er
2
0
2
1
:
1
6
1
0
-
1
6
1
7
1612
3.
L
UNG
CANC
E
R
DA
T
AS
E
T
S
T
h
e
d
atasets
u
s
ed
i
n
th
is
s
t
u
d
y
ar
e
m
ic
r
o
ar
r
a
y
a
n
d
R
N
A
-
s
eq
u
en
ce
d
atasets
.
T
h
e
d
a
ta
g
ath
er
ed
th
r
o
u
g
h
m
ic
r
o
ar
r
a
y
s
r
ep
r
esen
ts
th
e
g
e
n
e
e
x
p
r
ess
io
n
p
r
o
f
iles
,
wh
ich
s
h
o
w
s
im
u
ltan
e
o
u
s
ch
an
g
es
in
th
e
ex
p
r
ess
io
n
o
f
m
an
y
g
en
es
in
r
esp
o
n
s
e
to
a
p
ar
ticu
lar
co
n
d
i
tio
n
o
r
tr
ea
tm
en
t.
T
h
e
y
r
ep
r
e
s
en
t
th
e
m
o
lecu
lar
lev
el
s
tates
o
f
t
h
e
ce
ll
[
6
]
.
R
N
A
-
s
eq
u
en
ce
d
atasets
u
s
ed
a
s
eq
u
en
cin
g
tech
n
iq
u
e
(
n
e
x
t
-
g
e
n
er
atio
n
s
eq
u
en
cin
g
)
to
d
is
clo
s
e
th
e
p
r
esen
ce
an
d
q
u
an
tity
o
f
R
NA
in
a
b
io
lo
g
ical
s
am
p
le
at
a
g
iv
en
m
o
m
en
t,
an
aly
zi
n
g
th
e
co
n
tin
u
o
u
s
ly
ch
a
n
g
in
g
ce
llu
la
r
tr
an
s
cr
ip
to
m
e
[
36
]
.
T
h
is
s
tu
d
y
ap
p
lied
th
e
p
r
o
p
o
s
ed
m
o
d
el
o
n
two
m
icr
o
ar
r
a
y
d
atasets
an
d
o
n
e
R
NA
-
s
eq
u
e
n
ce
d
ataset
as
s
h
o
wn
in
T
ab
le
1
.
All
d
ataset
s
wer
e
d
o
wn
lo
ad
ed
f
r
o
m
th
e
g
e
n
e
ex
p
r
ess
io
n
o
m
n
ib
u
s
s
ite
(
GE
O
).
Fig
u
r
e
1.
XGBo
o
s
t
alg
o
r
ith
m
[
3
3
]
3
.
1
.
Da
t
a
s
et
info
r
m
a
t
io
n
E
ac
h
d
ataset
u
s
ed
h
as
a
d
if
f
er
en
t
way
o
f
ex
tr
ac
tin
g
th
e
g
en
e
ex
p
r
ess
io
n
,
th
e
n
u
m
b
e
r
o
f
f
ea
tu
r
es
an
d
th
e
n
u
m
b
er
o
f
ca
s
es.
T
h
e
f
ir
s
t
is
(
G
SE3
0
2
1
9
)
d
ataset
r
ep
r
esen
tin
g
th
e
g
en
e
ex
p
r
ess
io
n
b
y
m
icr
o
a
r
r
ay
tech
n
o
lo
g
y
.
I
t
h
as
1
4
n
o
r
m
al
lu
n
g
s
am
p
les
an
d
2
9
3
lu
n
g
ca
n
ce
r
s
am
p
les
[
37
]
.
T
h
e
s
ec
o
n
d
(
GSE7
4
7
0
6
)
d
ataset
is
also
r
ep
r
esen
ted
b
y
m
icr
o
ar
r
ay
tech
n
o
lo
g
y
.
I
t
e
x
p
r
ess
es
d
ata
o
f
ea
r
ly
-
s
tag
e
NSC
L
C
.
I
t
h
as
1
8
l
u
n
g
ca
n
ce
r
s
am
p
les
a
n
d
1
8
n
o
r
m
a
l
lu
n
g
s
am
p
les.
T
h
e
last
d
atas
et
(
GSE8
1
0
8
9
)
[
3
8
]
h
as
2
1
8
ca
s
es
ex
p
r
ess
ed
b
y
R
NA
-
s
eq
u
en
cin
g
,
wh
ich
is
c
alled
n
ex
t
-
g
e
n
er
atio
n
s
eq
u
en
c
in
g
[
3
9
]
;
R
NA
-
Seq
allo
ws
r
e
s
ea
r
ch
er
s
to
d
etec
t
g
en
e
f
u
s
io
n
s
v
ar
ian
ts
,
b
o
t
h
k
n
o
wn
a
n
d
n
o
v
el
f
ea
t
u
r
es
a
n
d
o
th
e
r
f
ea
tu
r
es
with
o
u
t
th
e
lim
itatio
n
o
f
p
r
io
r
k
n
o
wled
g
e
[
40
]
.
I
t
h
as 1
9
9
l
u
n
g
ca
n
ce
r
s
am
p
les with
lu
n
g
ca
n
ce
r
ty
p
e
NSC
L
C
an
d
1
9
h
ea
l
th
y
lu
n
g
s
am
p
les.
T
ab
le
1
.
Data
s
et'
s
in
f
o
r
m
atio
n
D
a
t
a
s
e
t
s
Ty
p
e
p
a
t
i
e
n
t
s
F
e
a
t
u
r
e
s
Th
e
C
l
a
ss
S
a
mp
l
e
d
i
st
r
i
b
u
t
i
o
n
C
a
n
c
e
r
c
a
s
e
N
o
r
mal
c
a
se
G
S
E3
0
2
1
9
M
i
c
r
o
a
r
r
a
y
3
0
7
5
4
6
7
5
C
a
n
c
e
r
/
N
o
r
m
a
l
2
9
3
14
G
S
E7
4
7
0
6
M
i
c
r
o
a
r
r
a
y
36
3
4
1
8
2
C
a
n
c
e
r
/
N
o
r
m
a
l
18
18
G
S
E8
1
0
8
9
N
e
w
G
e
n
e
r
a
t
i
o
n
S
e
q
u
e
n
c
i
n
g
(NGS)
2
1
8
6
3
1
2
9
C
a
n
c
e
r
/
N
o
r
mal
1
9
9
19
3
.
2
.
Da
t
a
prepro
ce
s
s
ing
Data
p
r
ep
r
o
ce
s
s
in
g
in
m
ac
h
i
n
e
lear
n
in
g
is
a
n
ess
en
tial
s
tep
in
e
n
h
an
ci
n
g
d
ata
q
u
ality
to
r
aise
m
ea
n
in
g
f
u
l
p
er
ce
p
tiv
e
n
ess
.
I
t
r
ef
er
s
to
clea
n
in
g
a
n
d
o
r
g
an
iz
in
g
th
e
r
aw
d
ata
to
m
ak
e
it
s
u
itab
le
f
o
r
b
u
ild
i
n
g
an
d
tr
ain
in
g
m
ac
h
in
e
lear
n
in
g
m
o
d
els.
I
n
b
io
lo
g
ical
d
at
a,
it
is
cr
u
cial
to
clea
n
th
e
d
ata
to
i
m
p
r
o
v
e
th
e
q
u
ality
o
f
th
e
d
ata
f
o
r
s
ea
r
ch
in
g
an
d
an
aly
zin
g
.
T
o
d
o
th
at,
it
r
u
n
s
a
p
r
o
ce
s
s
to
d
etec
t
an
d
r
em
o
v
e
co
r
r
u
p
t
o
r
in
ac
cu
r
ate
r
ec
o
r
d
s
f
r
o
m
th
e
d
a
tab
ase.
E
ac
h
r
ec
o
r
d
with
m
is
s
in
g
d
ata
m
u
s
t
b
e
d
elete
d
b
ec
a
u
s
e
it
is
r
eg
ar
d
ed
as
ir
r
elev
an
t
an
d
ca
u
s
e
in
ap
p
r
o
p
r
iate
lear
n
in
g
r
esu
lts
.
T
h
e
XGBo
o
s
t
clas
s
if
icatio
n
d
ea
ls
with
th
e
n
u
m
er
ic
r
ep
r
esen
tatio
n
in
th
e
d
ec
is
io
n
class
.
I
n
co
n
tr
ast,
th
e
class
es
in
th
e
lu
n
g
ca
n
ce
r
d
atas
ets
ar
e
in
n
o
m
i
n
al
r
ep
r
esen
tatio
n
,
lik
e
n
o
r
m
al
/can
c
er
.
T
h
e
r
ef
o
r
e
,
it m
u
s
t c
h
an
g
e
th
em
to
n
u
m
er
ic
r
e
p
r
esen
tatio
n
(
0
/
1
)
.
4.
T
H
E
P
ARA
L
L
E
L
_
XG
B
O
O
ST
(
P
XG
B
)
T
h
er
e
is
n
o
way
to
teac
h
o
n
e
m
ac
h
in
e
lear
n
in
g
to
f
it
all
k
in
d
s
o
f
in
f
o
r
m
atio
n
.
I
n
o
u
r
ca
s
e,
th
e
XGBo
o
s
t
s
u
cc
ee
d
ed
in
lear
n
in
g
o
n
s
o
m
e
d
atasets
with
h
ig
h
ac
cu
r
ac
y
b
u
t
lo
wer
in
o
th
er
s
.
T
h
at
is
b
ec
au
s
e
o
f
its
f
ir
m
r
elian
ce
o
n
its
h
y
p
er
p
ar
am
eter
s
ettin
g
.
T
h
is
s
tu
d
y
d
ev
elo
p
ed
an
XGBo
o
s
ts
s
tr
u
ctu
r
e
to
ac
c
o
m
m
o
d
ate
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
P
a
r
a
llel e
xtreme
g
r
a
d
ien
t b
o
o
s
tin
g
cla
s
s
ifier
fo
r
lu
n
g
ca
n
ce
r
d
etec
tio
n
(
R
a
n
a
D
h
ia
’
a
A
b
d
u
a
lja
b
a
r
)
1613
d
if
f
er
en
t
ty
p
es
o
f
d
atasets
b
y
co
n
n
ec
tin
g
m
u
ltip
le
n
u
m
b
er
s
o
f
XGBo
o
s
ts
o
n
p
ar
allel
with
v
ar
io
u
s
v
alu
es
o
f
h
y
p
er
p
ar
am
eter
s
.
T
h
en
it
tak
es
th
e
m
ax
im
u
m
p
r
o
b
a
b
ilit
y
f
o
r
its
p
r
ed
ictio
n
,
as
s
h
o
wn
in
Fig
u
r
e
2
.
All
th
e
XGb
o
o
s
ts
ar
e
wo
r
k
in
g
in
p
ar
allel
n
o
t
to
ca
u
s
e
a
d
elay
in
l
ea
r
n
in
g
tim
e.
As
s
ee
n
i
n
Fig
u
r
e
2
,
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
l
o
g
y
h
as th
r
ee
s
tag
es:
i)
Featu
r
e
s
elec
tio
n
s
tag
e
:
T
h
e
b
en
ef
it
o
f
u
s
in
g
XGBo
o
s
t
in
f
ea
tu
r
e
s
elec
tio
n
is
th
at
af
ter
th
e
b
o
o
s
ted
tr
ee
s
ar
e
c
o
n
s
tr
u
cted
,
th
ey
will
r
etr
i
ev
e
th
e
im
p
o
r
tan
ce
s
co
r
es
f
o
r
ea
ch
f
ea
tu
r
e
.
T
h
e
im
p
o
r
tan
ce
s
co
r
e
r
ef
er
s
to
h
o
w
u
s
ef
u
l
o
r
v
alu
a
b
le
ea
ch
f
ea
tu
r
e
was
in
co
n
s
tr
u
ctin
g
th
e
m
o
d
el
b
o
o
s
ted
d
ec
is
io
n
t
r
ee
s
.
T
h
e
m
o
r
e
f
ea
tu
r
e
is
u
s
ed
,
th
e
h
ig
h
er
its
im
p
o
r
tan
ce
s
co
r
e.
T
h
is
im
p
o
r
tan
ce
is
co
m
p
u
ted
f
o
r
ev
er
y
f
ea
tu
r
e
in
th
e
d
ataset,
allo
win
g
th
e
r
an
k
in
g
a
n
d
co
m
p
ar
is
o
n
b
etwe
en
th
em
.
T
h
e
im
p
o
r
tan
ce
o
f
e
v
er
y
d
ec
is
io
n
tr
ee
is
esti
m
ated
b
y
ca
lcu
latin
g
th
e
n
u
m
b
er
o
f
o
b
s
er
v
atio
n
s
r
esp
o
n
s
ib
le
f
o
r
ea
ch
f
ea
tu
r
e
s
p
lit
an
d
in
cr
ea
s
in
g
t
h
e
m
e
asu
r
em
en
t
o
f
p
er
f
o
r
m
an
ce
.
I
n
e
v
e
r
y
d
ec
is
io
n
tr
ee
in
th
e
m
o
d
el,
th
e
attr
ib
u
te
im
p
o
r
ts
ar
e
th
en
av
er
a
g
ed
[
2
3
]
.
I
n
th
is
p
ap
er
,
th
e
im
p
o
r
tan
ce
s
co
r
e
th
r
esh
o
ld
s
ettin
g
was
(
1
0
-
6
)
.
E
ac
h
attr
ib
u
te
less
th
an
th
is
th
r
esh
o
ld
w
ill
b
e
n
eg
lecte
d
.
T
h
e
f
ea
tu
r
es
o
f
GSE3
0
2
1
9
,
GSE7
4
7
0
6
,
an
d
GSE8
1
0
8
9
d
a
tasets
wer
e
(
5
4
6
7
5
)
,
(
3
4
1
8
2
)
a
n
d
(
6
3
1
2
9
)
,
r
esp
ec
tiv
ely
,
b
u
t
af
ter
th
e
f
ea
tu
r
e
s
elec
tio
n
s
tag
e
,
it b
ec
o
m
es (
2
0
)
,
(
1
)
an
d
(
8
)
f
ea
tu
r
es.
Fig
u
r
e
2.
T
h
e
p
r
o
p
o
s
al
lear
n
i
n
g
m
o
d
el
(
PXGB
)
ii)
Par
allel
XGBo
o
s
t
s
tag
e
:
Af
ter
th
e
f
ea
tu
r
e
s
elec
tio
n
s
tag
e,
th
e
d
ata
will
b
e
s
u
b
s
et
to
7
0
%
f
o
r
tr
ai
n
in
g
an
d
3
0
%
f
o
r
test
in
g
,
th
e
n
en
ter
ed
in
to
ea
ch
XGBo
o
s
t
s
im
u
ltan
eo
u
s
ly
.
I
n
o
u
r
ca
s
e,
it
n
ee
d
s
to
u
s
e
d
if
f
er
en
t
ty
p
es
o
f
b
io
_
d
ataset.
T
h
is
d
at
aset
is
u
s
u
ally
n
o
is
y
,
s
o
it
n
ee
d
s
th
e
m
o
d
el
to
tu
n
e
its
h
y
p
er
p
ar
am
eter
s
ea
ch
tim
e
to
av
o
id
o
v
er
f
itti
n
g
o
r
u
n
d
er
f
itti
n
g
t
o
h
an
d
le
a
wid
e
r
an
g
e
o
f
d
atasets
.
Fo
r
th
at
r
ea
s
o
n
,
I
t
u
s
ed
m
u
lti
XGBo
o
s
t
m
o
d
els
co
n
n
ec
ted
i
n
p
ar
allel.
E
ac
h
XGBo
o
s
t
h
a
s
its
h
y
p
er
p
ar
am
eter
s
s
ettin
g
d
if
f
er
e
n
t
f
r
o
m
ea
ch
o
th
er
.
T
h
is
s
tu
d
y
will
ta
k
e
s
ix
s
ets
o
f
XGBo
o
s
t
h
y
p
er
p
ar
am
eter
s
f
r
o
m
th
e
m
o
s
t
co
m
m
o
n
r
an
g
e
t
h
at
co
n
s
id
er
th
e
X
GB
o
o
s
t
m
o
d
el
o
f
ten
wo
r
k
s
well
in
th
em
.
T
h
e
h
y
p
er
p
ar
a
m
eter
s
r
an
g
es
ar
e;
th
e
s
u
b
s
am
p
le
[
0
.
5
-
1
]
,
t
h
e
Ma
x
_
d
ep
t
h
[
2
-
7
]
,
th
e
lea
r
n
in
g
r
ate
[
0
.
0
5
-
0
.
3
]
,
t
h
e
n
_
esti
m
ato
r
s
(
n
o
.
o
f
tr
ee
s
)
[
5
-
5
0
]
,
an
d
th
e
last
is
th
e
m
in
_
ch
ild
weig
h
t
f
r
o
m
[
1
-
6
]
.
T
h
eir
a
r
r
an
g
em
en
t
d
ep
en
d
s
o
n
t
h
e
m
o
s
t
v
alu
es
th
at
d
o
n
o
t
ca
u
s
e
o
v
er
f
itti
n
g
b
u
t
m
ay
s
o
m
etim
es
ca
u
s
e
an
u
n
d
er
f
itti
n
g
(
lev
el
o
n
e)
,
to
th
e
m
o
r
e
v
alu
es
th
at
m
ay
ca
u
s
e
o
v
er
f
itti
n
g
b
u
t
less
lik
ely
ca
u
s
in
g
u
n
d
er
f
itti
n
g
(
lev
el
s
ix
)
,
s
ee
T
ab
le
2
.
T
h
e
en
d
o
f
th
is
s
t
ag
e
will
h
a
v
e
a
p
r
o
b
a
b
ilit
y
p
r
ed
ictio
n
f
o
r
b
o
t
h
class
es
in
ea
ch
lev
el
.
iii)
Selectio
n
s
tag
e:
At
th
i
s
s
tag
e
,
it
will
tak
e
th
e
m
ax
im
u
m
p
r
o
b
ab
ilit
y
v
alu
e
o
f
all
XGBo
o
s
t
lev
els
.
T
h
e
r
esu
lt is
th
at
th
e
class
with
m
a
x
im
u
m
p
r
o
b
a
b
ilit
y
is
ch
o
s
en
a
s
th
e
f
in
al
class
p
r
ed
ictio
n
.
T
ab
le
2.
T
h
e
s
ettin
g
of
ea
c
h
X
GB
o
o
s
t
h
y
p
er
p
ar
am
ete
r
s
in
th
e
PXGB
X
G
B
o
o
st
s
e
q
u
e
n
c
e
in
t
h
e
p
a
r
a
l
l
e
l
st
a
g
e
X
G
B
o
o
st
h
y
p
e
r
p
a
r
a
me
t
e
r
s
su
b
s
a
m
b
l
e
M
a
x
_
d
e
p
t
h
Le
a
r
n
i
n
g
r
a
t
e
n
_
e
st
i
ma
t
o
r
s
mi
n
_
c
h
i
l
d
_
w
e
i
g
h
t
F
i
r
st
l
e
v
e
l
0
.
5
2
0
.
3
5
6
S
e
c
o
n
d
l
e
v
e
l
0
.
6
3
0
.
2
5
10
5
Th
i
r
d
l
e
v
e
l
0
.
7
4
0
.
2
20
4
F
o
u
r
t
h
l
e
v
e
l
0
.
8
5
0
.
1
5
30
3
F
i
f
t
h
l
e
v
e
l
0
.
9
6
0
.
1
40
2
S
i
x
t
h
l
e
v
e
l
1
7
0
.
0
5
50
1
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
24
,
No
.
3
,
Dec
em
b
er
2
0
2
1
:
1
6
1
0
-
1
6
1
7
1614
5.
T
H
E
R
E
SU
L
T
S
AND
D
I
SC
USSI
O
N
T
o
s
t
at
e
t
h
e
f
i
n
d
i
n
g
s
o
f
t
h
e
r
e
s
e
a
r
c
h
,
i
t
a
r
r
a
n
g
e
d
t
h
e
r
e
s
u
lt
s
i
n
a
l
o
g
i
ca
l
s
e
q
u
e
n
c
e
.
A
t
f
i
r
s
t,
it
c
o
m
p
a
r
e
d
t
h
e
r
e
s
u
l
t
o
f
t
h
e
P
X
GB
m
o
d
e
l
w
i
t
h
t
h
e
o
r
i
g
i
n
a
l
X
GB
o
o
s
t
[
3
5
]
t
o
r
e
p
r
e
s
e
n
t
i
t
s
i
m
p
r
o
v
e
m
e
n
t
.
I
t
t
h
e
n
c
o
m
p
a
r
e
d
i
t
s
r
e
s
u
l
t
w
i
t
h
r
e
p
r
es
e
n
t
a
ti
v
e
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
s
t
o
g
i
v
e
t
h
e
w
h
o
l
e
p
e
r
f
o
r
m
a
n
c
e
s
t
a
t
e
o
f
t
h
e
p
r
o
p
o
s
a
l
.
5
.
1
.
XG
B
o
o
s
t
hy
perpa
ra
m
et
er
s
s
et
t
ing
T
h
e
PXGB
s
et
s
s
ev
en
co
m
m
o
n
h
y
p
e
r
p
ar
am
eter
s
f
o
r
ea
ch
X
GB
o
o
s
t
clas
s
if
icatio
n
m
o
d
el,
as
illu
s
tr
ated
in
T
ab
le
2
.
T
h
ese
h
y
p
er
p
ar
a
m
eter
s
ar
e
ch
o
s
en
d
ep
en
d
in
g
o
n
th
eir
ar
r
a
n
g
em
en
t
f
r
o
m
m
o
s
t
h
y
p
er
p
ar
a
m
eter
s
th
at
m
ay
ca
u
s
e
th
e
o
v
er
f
itti
n
g
s
itu
atio
n
to
th
e
m
o
s
t
p
ar
am
eter
s
th
at
m
ay
ca
u
s
e
th
e
u
n
d
er
f
i
ttin
g
s
itu
atio
n
.
T
h
e
s
ettin
g
o
f
th
e
o
r
ig
in
al
XGBo
o
s
t
an
d
SVM
[4
1
]
,
d
ee
p
f
o
r
e
s
t
(
g
cf
o
r
est
)
[
4
2
]
,
k
-
n
ea
r
est
n
eig
h
b
o
r
s
alg
o
r
ith
m
(
KNN)
,
an
d
n
aiv
e
B
ay
es,
wh
i
ch
ar
e
th
e
m
ac
h
in
e
lea
r
n
in
g
s
th
at
u
s
ed
in
th
is
s
tu
d
y
,
h
a
v
e
a
p
ar
ticu
lar
s
ettin
g
illu
s
tr
ated
in
T
ab
le
3
.
T
ab
le
3.
Par
am
eter
s
s
ettin
g
of
r
ep
r
esen
tativ
e
m
o
d
els
X
G
B
o
o
st
S
V
M
g
c
F
o
r
e
st
K
N
N
N
a
i
v
e
B
a
y
e
s
P
a
r
a
me
t
e
r
v
a
l
u
e
P
a
r
a
me
t
e
r
s
v
a
l
u
e
P
a
r
a
me
t
e
r
v
a
l
u
e
P
a
r
a
me
t
e
r
v
a
l
u
e
P
a
r
a
me
t
e
r
v
a
l
u
e
max
_
d
e
p
t
h
6
k
e
r
n
e
l
RBF
max
_
d
e
p
t
h
6
n
_
n
e
i
g
h
b
o
r
2
v
a
r
_
smo
o
t
h
i
n
g
1e
-
9
n
_
e
st
i
ma
t
o
r
s
(
Tr
e
e
s)
2
g
a
mm
a
1
n
o
.
o
f
t
r
e
e
s
i
n
e
a
c
h
f
o
r
e
st
st
a
g
e
s=
5
0
0
5
0
0
w
e
i
g
h
t
s
u
n
i
f
o
r
m
samp
l
e
_
w
e
i
g
h
t
N
o
n
e
Le
a
r
n
i
n
g
r
a
t
e
0
.
3
t
o
l
e
r
a
n
c
e
0
.
0
0
1
W
i
n
d
.
si
z
e
5
0
0
a
l
g
o
r
i
t
h
m
a
u
t
o
mi
n
_
c
h
i
l
d
_
w
e
i
g
h
t
1
C
1
S
t
e
p
1
0
0
l
e
a
f
_
si
z
e
1
S
u
b
s
a
m
p
l
e
0
.
7
M
i
n
_
sam
p
l
e
s_
s
p
l
i
t
0
.
7
5
.
2
.
Co
m
pa
riso
n o
f
diff
er
en
t
cla
s
s
if
iers
Dif
f
er
en
t
r
esu
lts
wer
e
o
b
tain
e
d
af
ter
ap
p
ly
i
n
g
th
e
PXGB
m
o
d
el
an
d
o
t
h
er
m
ac
h
in
e
lear
n
i
n
g
m
o
d
els
to
th
e
lu
n
g
ca
n
ce
r
d
atasets
.
T
ab
les
4
illu
s
tr
ate
ea
ch
m
o
d
el'
s
s
en
s
itiv
ity
,
s
p
ec
if
icity
,
p
r
ec
is
io
n
,
F1
_
s
co
r
e,
ar
ea
u
n
d
er
cu
r
v
e
(
AUC
)
,
ac
cu
r
ac
y
,
an
d
lear
n
i
n
g
tim
e
m
etr
ics.
Fu
r
th
er
m
o
r
e
,
F
ig
u
r
es
3
,
4
,
an
d
5
s
h
o
w
th
e
r
ec
ei
v
er
o
p
er
atin
g
c
h
ar
ac
ter
is
tic
(
R
OC
)
d
r
awin
g
s
an
d
th
e
AUC
v
alu
e
s
o
f
ea
ch
m
ac
h
in
e
lear
n
in
g
m
o
d
el.
T
ab
le
4
.
C
o
m
p
a
r
is
o
n
r
esu
lts
of
lu
n
g
ca
n
ce
r
d
etec
tio
n
f
o
r
all
d
ataset
G
S
E8
1
0
8
9
d
a
t
a
se
t
C
l
a
s
si
f
i
e
r
N
a
m
e
S
e
n
s
i
t
i
v
i
t
y
S
p
e
c
i
f
i
c
i
t
y
P
r
e
c
i
s
i
o
n
F
1
_
s
c
o
r
e
AUC
A
c
c
u
r
a
c
y
Ti
me
(
mi
n
.
)
P
X
G
B
S
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
0
0
:
0
3
X
G
B
o
o
st
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
0
0
:
0
4
S
V
M
0
.
2
0
.
8
3
0
.
5
0
.
2
9
0
.
5
2
0
.
5
5
0
0
:
0
1
g
c
F
o
r
e
st
1
.
0
0
0
.
4
5
0
.
6
3
0
.
5
0
0
.
4
5
0
0
:
3
6
K
N
N
0
.
8
1
.
0
1
.
0
0
.
8
9
0
.
9
0
0
.
9
1
0
0
:
0
1
N
a
i
v
e
B
a
y
e
s
0
.
6
0
.
6
7
0
.
6
0
.
6
0
.
6
3
0
.
6
4
0
0
:
0
1
C
l
a
s
si
f
i
e
r
N
a
m
e
S
e
n
s
i
t
i
v
i
t
y
S
p
e
c
i
f
i
c
i
t
y
P
r
e
c
i
s
i
o
n
F
1
_
s
c
o
r
e
AUC
A
c
c
u
r
a
c
y
Ti
me
(
mi
n
.
)
P
X
G
B
S
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
0
0
:
1
3
X
G
B
o
o
st
1
.
0
0
.
9
5
1
.
0
0
.
9
9
0
.
9
9
0
.
9
8
0
0
:
2
4
S
V
M
1
.
0
0
.
5
0
.
9
5
0
.
9
8
0
.
7
5
0
.
9
5
0
0
:
0
5
g
c
F
o
r
e
st
0
.
9
8
0
.
8
3
0
.
9
8
0
.
9
8
0
.
9
1
0
.
9
7
0
3
:
3
7
K
N
N
0
.
9
5
0
.
5
0
.
9
5
0
.
9
5
0
.
7
2
0
.
9
1
0
0
:
2
9
N
a
i
v
e
B
a
y
e
s
1
.
0
0
.
1
7
0
.
9
2
0
.
9
6
0
.
5
8
0
.
9
2
0
0
:
0
2
G
S
E7
4
7
0
6
d
a
t
a
se
t
C
l
a
s
si
f
i
e
r
N
a
m
e
S
e
n
s
i
t
i
v
i
t
y
S
p
e
c
i
f
i
c
i
t
y
P
r
e
c
i
s
i
o
n
F
1
_
s
c
o
r
e
AUC
A
c
c
u
r
a
c
y
Ti
me
(
mi
n
.
)
P
X
G
B
S
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
0
0
:
1
3
X
G
B
o
o
st
0
.
9
9
1
.
0
0
.
9
9
1
.
0
0
.
9
9
0
.
9
9
0
0
:
1
7
S
V
M
1
.
0
0
0
.
9
6
0
.
9
8
0
.
5
0
.
9
6
0
0
:
0
7
g
c
F
o
r
e
st
0
.
9
8
0
.
7
5
0
.
9
8
0
.
9
8
0
.
8
7
0
.
9
7
0
3
:
2
6
K
N
N
0
.
9
8
1
.
0
1
.
0
0
.
9
9
0
.
9
9
0
.
9
9
0
0
:
1
2
N
a
i
v
e
B
a
y
e
s
0
.
9
9
1
.
0
1
.
0
0
.
9
9
0
.
9
9
0
.
9
9
0
0
:
0
2
5
.
3
.
Ana
ly
zing
m
et
rics
Fro
m
T
ab
le
4
,
it
is
s
ee
n
th
at
a
ll
PXGB
m
etr
ics
h
av
e
ex
ce
lle
n
t
v
alu
es
wh
e
n
a
p
p
lied
to
all
d
atasets
.
I
t
s
u
cc
ee
d
ed
in
d
etec
tin
g
all
ca
s
es
(
ca
n
ce
r
an
d
n
o
r
m
al
ca
s
es)
in
all
d
atasets
.
I
n
co
n
tr
ast,
XGBo
o
s
t
s
u
cc
es
s
f
u
lly
p
r
ed
icts
all
ca
s
es
o
n
ly
in
GSE
8
1
0
8
9
d
ataset
b
ec
au
s
e
it
h
as
o
n
ly
o
n
e
s
et
o
f
h
y
p
er
p
ar
am
et
er
s
,
wh
ile
XGBo
o
s
t
h
as
a
r
an
g
e
o
f
h
y
p
er
p
ar
am
et
er
s
th
at
let
it
b
u
ild
m
u
ltip
le
XGBo
o
s
t
s
tr
u
ctu
r
es
in
th
e
tr
ain
in
g
s
tag
e.
PXGB
g
iv
es
th
e
f
lex
ib
ilit
y
to
d
ea
l
wi
th
d
if
f
er
e
n
t
d
atasets
an
d
allo
w
s
all
th
e
XGBo
o
s
t
s
tr
u
ctu
r
es
t
o
co
n
tr
ib
u
te
to
th
e
class
d
etec
tio
n
in
t
h
e
test
s
tag
e
an
d
th
en
ch
o
o
s
e
th
e
b
est
p
r
e
d
ictio
n
b
y
s
elec
tin
g
th
e
class
with
th
e
m
a
x
im
u
m
p
r
ed
ictio
n
v
alu
e.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
P
a
r
a
llel e
xtreme
g
r
a
d
ien
t b
o
o
s
tin
g
cla
s
s
ifier
fo
r
lu
n
g
ca
n
ce
r
d
etec
tio
n
(
R
a
n
a
D
h
ia
’
a
A
b
d
u
a
lja
b
a
r
)
1615
T
h
e
PXGB
im
p
r
o
v
ed
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
XGBo
o
s
t.
I
t
h
as
b
ec
o
m
e
m
o
r
e
p
o
wer
f
u
l
a
n
d
r
eliab
le
f
o
r
a
v
ar
ian
t
ty
p
e
o
f
d
ataset
with
o
u
t
ch
an
g
in
g
its
h
y
p
er
p
ar
a
m
eter
s
.
Alth
o
u
g
h
th
e
n
ai
v
e
B
ay
es
h
as
th
e
s
h
o
r
test
lear
n
in
g
tim
e
in
m
o
s
t
d
atasets
,
th
e
PXGB
h
as
an
ac
ce
p
ted
l
ea
r
n
in
g
tim
e
r
an
g
in
g
f
r
o
m
3
t
o
1
3
s
ec
o
n
d
s
.
I
t
is
ev
en
s
h
o
r
ter
th
an
th
e
o
r
ig
i
n
al
XGBo
o
s
t
r
an
g
in
g
f
r
o
m
4
to
2
3
s
ec
o
n
d
s
b
ec
au
s
e
o
f
t
h
e
s
elec
tio
n
f
ea
tu
r
e
p
r
o
ce
s
s
,
an
d
th
e
m
u
ltip
le
XG
B
o
o
s
t a
r
e
wo
r
k
ed
in
p
ar
allel,
d
ec
r
ea
s
in
g
th
e
s
y
s
tem
o
v
er
h
ea
d
.
Fig
u
r
e
3
.
T
h
e
R
OC
cu
r
v
es a
n
d
AUC v
alu
es f
o
r
all
co
m
p
ar
ativ
e
m
o
d
els o
n
GSE8
1
0
8
9
d
ataset
Fig
u
r
e
4
.
T
h
e
R
OC
cu
r
v
es a
n
d
AUC v
alu
es f
o
r
all
co
m
p
ar
ativ
e
m
o
d
els o
n
GSE3
0
2
1
9
d
ataset
Fig
u
r
e
5
.
T
h
e
R
OC
cu
r
v
es a
n
d
AUC v
alu
es f
o
r
all
co
m
p
a
r
ativ
e
m
o
d
els o
n
GSE7
4
7
0
6
d
atas
et
6.
CO
NCLU
SI
O
N
T
h
is
s
tu
d
y
p
r
o
p
o
s
ed
a
lu
n
g
c
an
ce
r
d
etec
tio
n
s
y
s
tem
with
m
u
lti
-
s
tag
es
to
r
ea
ch
o
p
tim
al
r
esu
lts
.
I
t
u
s
es
th
e
XGBo
o
s
t
as
a
f
ea
tu
r
e
s
elec
tio
n
to
ch
o
o
s
e
o
n
l
y
ac
tiv
e
g
en
es
t
h
at
h
a
v
e
an
ess
en
tial
r
o
le
in
lu
n
g
ca
n
ce
r
d
is
ea
s
e
an
d
s
u
g
g
ested
a
f
lex
ib
le
m
ac
h
in
e
lear
n
in
g
b
y
u
s
in
g
m
u
ltip
le
XGBo
o
s
t
class
if
icatio
n
s
to
r
u
n
in
p
ar
allel
.
E
ac
h
XGBo
o
s
t
in
p
ar
allel
s
ta
g
e
h
as
d
if
f
er
en
t
s
ets
o
f
h
y
p
er
p
ar
am
eter
s
,
r
an
g
in
g
f
r
o
m
th
e
m
o
s
t
v
alu
es
th
at
m
ay
lead
to
o
v
er
f
itti
n
g
to
th
e
p
ar
am
eter
s
'
v
alu
es
th
at
m
ig
h
t
ca
u
s
e
th
e
u
n
d
er
f
itti
n
g
.
T
h
at
led
to
o
b
tain
in
g
v
ar
io
u
s
tr
ee
b
u
il
d
in
g
s
,
wh
ic
h
g
iv
es
th
e
PXGB
f
lex
ib
ilit
y
a
n
d
r
eliab
ilit
y
;
wh
e
n
a
p
p
lied
t
o
d
if
f
er
en
t
d
atasets
.
Mo
r
eo
v
er
,
u
s
in
g
f
ea
tu
r
e
s
elec
tio
n
im
p
r
o
v
e
d
th
e
d
etec
tio
n
a
cc
u
r
ac
y
;
it
also
s
p
ed
u
p
th
e
lear
n
in
g
tim
e.
T
h
e
r
esu
lts
s
h
o
wed
th
at
th
e
PXGB
m
o
d
el
,
th
e
p
r
o
p
o
s
ed
m
o
d
el,
i
m
p
r
o
v
e
d
lu
n
g
ca
n
ce
r
d
etec
tio
n
p
er
f
o
r
m
an
ce
.
T
h
is
im
p
r
o
v
em
e
n
t
is
b
etter
th
an
th
e
o
r
ig
in
al
XGBo
o
s
t
an
d
o
th
er
co
m
p
ar
ativ
e
m
ac
h
in
e
lear
n
in
g
,
esp
ec
ially
f
o
r
im
b
alan
ce
d
d
atasets
an
d
with
i
n
an
ac
ce
p
ta
b
le
tim
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
24
,
No
.
3
,
Dec
em
b
er
2
0
2
1
:
1
6
1
0
-
1
6
1
7
1616
RE
F
E
R
E
NC
E
S
[1
]
Wo
rld
He
a
lt
h
Or
g
a
n
iza
ti
o
n
,
Ca
n
c
e
r,
2
0
2
1
.
[O
n
li
n
e
].
Av
a
il
a
b
le:
h
tt
p
s:/
/www
.
wh
o
.
in
t
/n
e
ws
-
ro
o
m
/fac
t
-
sh
e
e
ts/d
e
tail/
c
a
n
c
e
r
.
[
2
]
R
.
P
a
r
k
,
J
.
W
.
S
h
a
w
,
A
.
K
o
r
n
,
a
n
d
J
.
M
c
A
u
l
i
f
f
e
,
"
T
h
e
v
a
l
u
e
of
im
m
u
n
o
t
h
e
r
a
p
y
f
o
r
s
u
r
v
i
v
o
r
s
of
s
t
a
g
e
IV
non
-
s
m
a
l
l
c
e
l
l
l
u
n
g
c
a
n
c
e
r
:
p
a
t
i
e
n
t
p
e
r
s
p
e
c
t
i
v
e
s
on
q
u
a
l
i
t
y
of
l
i
f
e
,
"
J
.
C
a
n
c
e
r
S
u
r
v
i
v
,
v
o
l
.
1
4
,
n
o
.
3,
pp.
3
6
3
-
3
7
6
,
2
0
2
0
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
1
7
6
4
-
020
-
0
0
8
5
3
-
3.
[3
]
Y.
Wan
g
e
t
a
l
.
,
"A
n
o
v
e
l
4
-
g
e
n
e
sig
n
a
tu
re
f
o
r
o
v
e
ra
ll
su
r
v
iv
a
l
p
re
d
ictio
n
in
l
u
n
g
a
d
e
n
o
c
a
rc
in
o
m
a
p
a
ti
e
n
ts
with
ly
m
p
h
n
o
d
e
m
e
tas
tas
is
Ca
n
c
e
r,
"
Ce
ll
In
t.
,
v
o
l
19,
n
o
.
1
0
0
,
2
0
1
9
,
d
o
i:
1
0
.
1
1
8
6
/s
1
2
9
3
5
-
0
1
9
-
0
8
2
2
-
1
.
[4
]
E.
F
.
Nu
wa
y
sir,
M
.
Bit
t
n
e
r,
J.
Tr
e
n
t,
J.
C.
Ba
rre
tt
,
a
n
d
C.
A
.
Afs
h
a
ri
,
“
M
icro
a
rra
y
s
a
n
d
to
x
ico
l
o
g
y
:
th
e
a
d
v
e
n
t
of
to
x
ic
o
g
e
n
o
m
ics
,
”
M
o
lec
u
l
a
r
C
a
rc
in
o
g
e
n
e
sis
,
v
o
l.
24
,
n
o
.
3,
pp.
1
5
3
-
1
5
9
,
1
9
9
9
,
d
o
i:
1
0
.
1
0
0
2
/(sici)
1
0
9
8
-
2
7
4
4
(
1
9
9
9
0
3
)
2
4
:
3
%
3
C1
5
3
::
a
i
d
-
m
c
1
%
3
E
3
.
0
.
c
o
;2
-
p.
[5
]
Y
.
Ya
n
g
,
E
.
A
.
G
.
Blo
m
m
e
,
a
n
d
J
.
F
.
Warin
g
,
"
To
x
ico
g
e
n
o
m
ic
s
in
d
ru
g
d
isc
o
v
e
r:
F
r
o
m
p
re
c
li
n
ica
l
stu
d
ies
to
c
li
n
ica
l
tri
a
ls
,
"
C
h
e
m.
Bi
o
l.
I
n
ter
a
c
t
,
v
o
l.
1
5
0
,
no.
1,
p
p
.
71
-
8
5
,
2
0
0
4
,
d
o
i:
1
0
.
1
0
1
6
/j
.
c
b
i.
2
0
0
4
.
0
9
.
0
1
3
.
[6
]
H
.
A
.
Ru
e
d
a
-
Zára
te,
I
.
Im
a
z
-
Ro
s
sh
a
n
d
ler,
R
.
A
.
Cá
rd
e
n
a
s
-
Ov
a
n
d
o
,
J
.
E
.
Ca
stil
l
o
-
F
e
rn
á
n
d
e
z
,
J
.
N
o
g
u
e
z
-
M
o
n
r
o
y
,
a
n
d
C
.
Ra
n
g
e
l
-
Esc
a
re
ñ
o
,
"A
c
o
m
p
u
tati
o
n
a
l
t
o
x
ic
o
g
e
n
o
m
ics
a
p
p
r
o
a
c
h
id
e
n
t
ifi
e
s
a
li
st
of
h
i
g
h
ly
h
e
p
a
to
t
o
x
ic
c
o
m
p
o
u
n
d
s
fr
o
m
a
larg
e
m
icro
a
rr
a
y
d
a
tab
a
se
,
"
Pl
o
s
On
e
,
v
o
l.
12,
n
o
.
4,
2
0
1
7
,
d
o
i:
1
0
.
1
3
7
1
/
jo
u
rn
a
l.
p
o
n
e
.
0
1
7
6
2
8
4
.
[7
]
R
.
Al
-
A
n
n
i,
J
.
Ho
u
,
R
.
D
.
Ab
d
u
-
Aljab
a
r
,
a
n
d
Y
.
Xia
n
g
,
"
P
re
d
ict
io
n
of
NSCL
C
re
c
u
rre
n
c
e
fro
m
m
icro
a
rra
y
d
a
ta
with
G
EP
,
"
IE
T
sy
ste
ms
b
io
l
o
g
y
,
v
o
l.
1
1
,
no.
3,
p
p
.
77
-
78.
2
0
1
7
,
d
o
i:
1
0
.
1
0
4
9
/i
e
t
-
sy
b
.
2
0
1
6
.
0
0
3
3
.
[8
]
R.
Al
-
A
n
n
i,
J.
H
o
u
,
H.
Az
z
a
wi
a
n
d
Y.
Xia
n
g
,
"
Ca
n
c
e
r
a
d
j
u
v
a
n
t
c
h
e
m
o
th
e
ra
p
y
p
re
d
icti
o
n
m
o
d
e
l
fo
r
non
-
sm
a
ll
c
e
ll
lu
n
g
c
a
n
c
e
r"
,
IE
T
sy
ste
ms
b
io
l
o
g
y
,
v
o
l.
13,
n
o
.
3,
p
p
.
1
2
9
-
1
3
5
,
2
0
1
9
,
d
o
i:
1
0
.
1
0
4
9
/
iet
-
sy
b
.
2
0
1
8
.
5
0
6
0
.
[9
]
R.
Al
-
A
n
n
i
,
J.
Ho
u
,
H.
Az
z
a
wi
,
a
n
d
Y.
Xia
n
g
,
“
Risk
c
las
sifica
t
io
n
f
o
r
NSCL
C
s
u
rv
i
v
a
l
u
si
n
g
m
icro
a
rra
y
a
n
d
c
li
n
ica
l
d
a
ta,”
In
ter
n
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
A
d
v
a
n
c
e
s in
El
e
c
tro
n
ics
a
n
d
Co
m
p
u
ter
S
c
ie
n
c
e
,
v
o
l.
6
,
n
o
.
3
,
2
0
1
9
.
[1
0
]
R.
Al
-
An
n
i
,
J.
Ho
u
,
H.
Az
z
a
wi
a
n
d
Y.
Xia
n
g
,
"A
n
o
v
e
l
g
e
n
e
se
lec
ti
o
n
a
lg
o
rit
h
m
fo
r
c
a
n
c
e
r
c
las
sifica
ti
o
n
u
sin
g
m
icro
a
rra
y
d
a
tas
e
ts,
”
BM
C
M
e
d
.
Ge
n
o
mic
s
,
v
o
l.
12,
n
o
.
10,
2
0
1
8
,
d
o
i:
1
0
.
1
1
8
6
/s1
2
9
2
0
-
0
1
8
-
0
4
4
7
-
6.
[1
1
]
R.
Al
-
An
n
i,
J.
Ho
u
,
H.
Az
z
a
wi
a
n
d
Y.
Xia
n
g
,
"
D
eep
g
e
n
e
se
l
e
c
ti
o
n
m
e
th
o
d
to
se
lec
t
g
e
n
e
s
f
ro
m
m
icro
a
rra
y
d
a
tas
e
ts
fo
r
c
a
n
c
e
r
c
las
sifica
ti
o
n
,
"
BM
C
-
i
n
fo
rm
a
ti
c
s
,
v
o
l.
2
0
,
n
o
.
6
0
8
,
2
0
1
8
,
d
o
i:
1
0
.
1
1
8
6
/s
1
2
8
5
9
-
0
1
9
-
3
1
6
1
-
2.
[1
2
]
R.
Al
-
An
n
i
,
J.
H
o
u
,
H.
Az
z
a
wi
a
n
d
Y.
Xia
n
g
,
“
Ne
w
G
e
n
e
S
e
lec
ti
o
n
M
e
th
o
d
Us
i
n
g
G
e
n
e
Ex
p
re
ss
i
o
n
P
ro
g
ra
m
in
g
Ap
p
ro
a
c
h
on
M
icro
a
rra
y
Da
ta
S
e
ts
,
”
In
t.
Co
n
f.
on
Co
mp
.
a
n
d
I
n
fo
.
S
c
ien
c
e
4
th
,
v
o
l.
7
9
1
,
p
p
.
1
7
-
3
1
,
2
0
1
9
,
d
o
i:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
3
1
9
-
9
8
6
9
3
-
7
_
2
.
[1
3
]
H.
Az
z
a
wi,
J.
Ho
u
,
Y.
Xia
n
g
a
n
d
R.
Ala
n
n
i
,
"
Lu
n
g
c
a
n
c
e
r
p
re
d
i
c
ti
o
n
fr
o
m
m
icro
a
rra
y
d
a
ta
by
g
e
n
e
e
x
p
re
ss
io
n
p
ro
g
ra
m
m
in
g
,
"
IET
S
y
st.
Bi
o
l
.,
v
o
l.
1
0
,
no.
5,
p
p
.
1
6
8
-
1
7
8
,
2
0
1
6
,
d
o
i:
1
0
.
1
0
4
9
/i
e
t
-
sy
b
.
2
0
1
5
.
0
0
8
2
.
[1
4
]
H.
Az
z
a
wi,
J.
Ho
u
,
Y.
Xia
n
g
,
R.
Ala
n
n
i,
R.
Ab
d
u
-
a
lj
a
b
a
r
a
n
d
A.
Az
z
a
wi,
“
M
u
lt
icla
ss
lu
n
g
c
a
n
c
e
r
d
iag
n
o
sis
by
g
e
n
e
e
x
p
re
ss
io
n
p
r
o
g
ra
m
m
in
g
a
n
d
m
icro
a
rra
y
d
a
tas
e
ts,”
1
3
t
h
In
t.
Co
n
f.
on
A
d
v
a
n
c
e
d
Da
t
a
M
i
n
in
g
a
n
d
Ap
p
li
c
a
ti
o
n
s
,
v
o
l
.
3
8
,
2
0
1
7
,
pp
5
4
1
-
5
5
3
,
d
o
i:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
3
1
9
-
6
9
1
7
9
-
4
_
3
8
.
[1
5
]
H.
Az
z
a
wi,
J.
H
o
u
,
R.
Al
n
n
n
i,
a
n
d
Y.
Xia
n
g
,
“
S
BC:
A
Ne
w
S
tr
a
teg
y
f
o
r
M
u
l
ti
c
las
s
Lu
n
g
Ca
n
c
e
r
Clas
sifica
ti
o
n
Ba
se
d
o
n
Tu
m
o
u
r
S
tru
c
t
u
ra
l
In
fo
r
m
a
ti
o
n
a
n
d
M
icr
o
a
rra
y
Da
ta,”
2
0
1
8
IEE
E/
A
CIS
1
7
t
h
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Co
mp
u
ter
a
n
d
In
fo
rm
a
t
io
n
S
c
ien
c
e
(ICIS
)
,
2
0
1
8
,
p
p
.
6
8
-
7
3
,
d
o
i:
1
0
.
1
1
0
9
/ICIS
.
2
0
1
8
.
8
4
6
6
4
4
8
.
[1
6
]
H.
Az
z
a
wi,
J.
Ho
u
,
R.
Ala
n
n
i,
a
n
d
Y.
Xia
n
,
“
A
h
y
b
rid
n
e
u
ra
l
n
e
two
rk
a
p
p
ro
a
c
h
f
o
r
l
u
n
g
c
a
n
c
e
r
c
las
sifica
ti
o
n
wit
h
g
e
n
e
e
x
p
re
ss
io
n
d
a
tas
e
t
a
n
d
p
ri
o
r
b
io
l
o
g
ica
l
k
n
o
wle
d
g
e
,”
In
t.
Co
n
f
.
on
M
a
c
h
in
e
L
e
a
rn
i
n
g
fo
r
Ne
two
rk
in
g
,
pp.
2
7
9
-
293
,
2
0
1
9
,
d
o
i:
1
0
.
1
0
0
7
/
9
7
8
-
3
-
0
3
0
-
1
9
9
4
5
-
6
_
2
0
.
[1
7
]
H.
Hu
,
Q.
G
u
a
n
,
S
.
Ch
e
n
,
Z.
Ji
a
n
d
Y.
Li
n
,
“
De
tec
ti
o
n
a
n
d
Re
c
o
g
n
it
io
n
fo
r
Li
fe
S
tate
o
f
Ce
ll
Ca
n
c
e
r
Us
in
g
Two
-
S
tag
e
Ca
sc
a
d
e
CNN
s,”
in
IEE
E/
ACM
T
r
a
n
s
a
c
ti
o
n
s
o
n
C
o
mp
u
ta
t
i
o
n
a
l
Bi
o
l
o
g
y
a
n
d
B
io
i
n
fo
rm
a
ti
c
s
,
v
o
l.
1
7
,
n
o
.
3
,
p
p
.
8
8
7
-
8
9
8
,
1
M
a
y
-
J
u
n
e
2
0
2
0
,
d
o
i:
1
0
.
1
1
0
9
/
TCBB.
2
0
1
7
.
2
7
8
0
8
4
2
.
[1
8
]
M
.
Š
a
rić,
M
.
Ru
ss
o
,
M
.
S
tella,
a
n
d
M
.
S
i
k
o
ra
,
“
CNN
-
b
a
se
d
M
e
t
h
o
d
fo
r
Lu
n
g
Ca
n
c
e
r
De
tec
ti
o
n
i
n
Wh
o
le
S
l
id
e
Histo
p
a
th
o
lo
g
y
Im
a
g
e
s,”
2
0
1
9
4
th
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
S
ma
rt
a
n
d
S
u
st
a
i
n
a
b
le
T
e
c
h
n
o
lo
g
ies
(S
p
l
iT
e
c
h
)
,
2
0
1
9
,
p
p
.
1
-
4
,
d
o
i:
1
0
.
2
3
9
1
9
/S
p
li
Tec
h
.
2
0
1
9
.
8
7
8
3
0
4
1
.
[1
9
]
S
.
L
i
e
t
a
l
.
,
"
P
re
d
ictin
g
l
u
n
g
n
o
d
u
le
m
a
li
g
n
a
n
c
ies
by
c
o
m
b
in
i
n
g
d
e
e
p
c
o
n
v
o
lu
t
io
n
a
l
n
e
u
ra
l
n
e
two
r
k
a
n
d
h
a
n
d
c
ra
fted
fe
a
tu
re
s
P
h
y
sic
s
in
M
e
d
ici
n
e
&
Bio
lo
g
y
,
"
P
h
y
sic
s
in
M
e
d
icin
e
a
n
d
B
io
l
o
g
y
,
v
o
l
.
64
n
o
.
17,
2
0
1
9
,
d
o
i
:
1
0
.
1
0
8
8
/
1
3
6
1
-
6
5
6
0
/ab
3
2
6
a
.
[2
0
]
R.
P
a
tra,
“
P
re
d
icti
o
n
of
L
u
n
g
Ca
n
c
e
r
Us
in
g
M
a
c
h
i
n
e
Lea
rn
i
n
g
Clas
sifier,
”
I
n
t.
C
o
n
f.
on
Co
mp
u
ti
n
g
S
c
ien
c
e
,
Co
mm
u
n
ica
ti
o
n
a
n
d
S
e
c
u
rity
Co
mp
u
t
in
g
S
c
ien
c
e
,
Co
mm
u
n
ic
a
ti
o
n
a
n
d
S
e
c
u
rity
,
p
p
.
1
3
2
-
1
4
2
,
2
0
2
0
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
9
8
1
-
15
-
6
6
4
8
-
6
_
1
1
.
[2
1
]
Y
.
-
H
.
Lai,
W
.
-
N
.
Ch
e
n
,
T
.
-
C
.
Hs
u
,
C
.
Li
n
,
Y
.
Tsa
o
,
a
n
d
S
.
Wu
,
"
Ov
e
ra
ll
su
rv
iv
a
l
p
re
d
ictio
n
of
n
o
n
-
sm
a
ll
c
e
ll
lu
n
g
c
a
n
c
e
r
by
in
te
g
ra
ti
n
g
m
icro
a
rra
y
a
n
d
c
li
n
ica
l
d
a
ta
with
d
e
e
p
lea
rn
i
n
g
,
"
S
c
ien
ti
fi
c
Rep
o
rts
,
v
o
l.
1
0
,
n
o
.
4
6
7
9
,
2
0
2
0
,
d
o
i:
1
0
.
1
0
3
8
/s4
1
5
9
8
-
0
2
0
-
6
1
5
8
8
-
w.
[2
2
]
M.
M.
A.
P
ri
y
a
a
n
d
S.
J.
Ja
wh
a
r,
"
Ad
v
a
n
c
e
d
lu
n
g
c
a
n
c
e
r
c
las
sifica
ti
o
n
a
p
p
r
o
a
c
h
a
d
o
p
ti
n
g
m
o
d
ifi
e
d
g
ra
p
h
c
lu
ste
ri
n
g
a
n
d
wh
a
le
o
p
ti
m
isa
ti
o
n
-
b
a
se
d
fe
a
tu
re
se
lec
ti
o
n
tec
h
n
iq
u
e
a
c
c
o
m
p
a
n
ied
by
a
h
y
b
rid
e
n
se
m
b
le
c
las
sifier
,
"
IET
Im
a
g
e
Pro
c
e
ss
in
g
,
v
o
l
.
1
4
,
n
o
.
1
0
,
p
p
.
2
2
0
4
-
2
2
1
5
,
2
0
2
0
,
d
o
i:
1
0
.
1
0
4
9
/
iet
-
ip
r.
2
0
1
9
.
0
1
7
8
.
[2
3
]
A
.
Og
u
n
le
y
e
a
n
d
Q.
-
G
.
Wan
g
,
"
XG
Bo
o
st
M
o
d
e
l
fo
r
Ch
ro
n
i
c
Kid
n
e
y
Dise
a
se
Dia
g
n
o
sis,"
in
IEE
E/
AC
M
T
ra
n
sa
c
ti
o
n
s
o
n
Co
m
p
u
t
a
ti
o
n
a
l
Bi
o
lo
g
y
a
n
d
Bi
o
in
f
o
rm
a
ti
c
s
,
v
o
l
.
1
7
,
n
o
.
6
,
p
p
.
2
1
3
1
-
2
1
4
0
,
1
No
v
.
-
De
c
.
2
0
2
0
,
d
o
i
:
1
0
.
1
1
0
9
/
TCBB.
2
0
1
9
.
2
9
1
1
0
7
1
.
[2
4
]
A
.
S
.
De
su
k
y
,
A
.
H
.
Om
a
r
,
a
n
d
N
.
M
.
M
o
st
a
fa
,
"
B
o
o
sti
n
g
wi
th
c
ro
ss
o
v
e
r
fo
r
imp
r
o
v
in
g
imb
a
lan
c
e
d
m
e
d
ica
l
d
a
tas
e
ts
c
las
sifica
ti
o
n
,
"
B
u
ll
e
ti
n
o
f
El
e
c
trica
l
E
n
g
i
n
e
e
rin
g
a
n
d
I
n
fo
rm
a
ti
c
s
,
v
o
l.
10,
n
o
.
5,
pp.
2
7
3
3
-
2
7
4
1
,
2
0
2
1
,
d
o
i:
1
0
.
1
1
5
9
1
/ee
i.
v
1
0
i
5
.
3
1
2
1
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
P
a
r
a
llel e
xtreme
g
r
a
d
ien
t b
o
o
s
tin
g
cla
s
s
ifier
fo
r
lu
n
g
ca
n
ce
r
d
etec
tio
n
(
R
a
n
a
D
h
ia
’
a
A
b
d
u
a
lja
b
a
r
)
1617
[2
5
]
Z
.
Ru
sta
m
,
A
.
P
u
rwa
n
to
,
S
.
Ha
rti
n
i
,
a
n
d
G
.
S
.
S
a
ra
g
ih
,
"
Lu
n
g
c
a
n
c
e
r
c
las
sifica
ti
o
n
u
sin
g
fu
z
z
y
c
-
m
e
a
n
s
a
n
d
fu
z
z
y
k
e
rn
e
l
C
-
M
e
a
n
s
b
a
se
d
on
CT
sc
a
n
ima
g
e
,
"
I
n
d
o
n
e
sia
n
J
o
u
rn
a
l
o
f
Ap
p
li
e
d
In
fo
rm
a
t
ics
,
v
o
l.
10,
n
o
.
2,
p
p
.
2
9
1
-
2
9
7
,
2
0
2
1
,
d
o
i:
1
0
.
1
1
5
9
1
/i
jai.
v
1
0
.
i
2
.
p
p
2
9
1
-
2
9
7
.
[2
6
]
R.
P
a
n
d
ian
,
D.
N.
S
.
R
.
K
u
m
a
r
,
a
n
d
R.
R
.
K
u
m
a
r
,
"
De
v
e
l
o
p
m
e
n
t
of
a
lg
o
rit
h
m
fo
r
id
e
n
ti
fica
ti
o
n
of
m
a
li
g
a
n
t
g
ro
wt
h
in
c
a
n
c
e
r
u
sin
g
a
rti
ficia
l
n
e
u
ra
l
n
e
two
rk
"
,
I
n
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
ter
En
g
i
n
e
e
rin
g
,
v
o
l.
10,
n
o.
6,
p
p
.
5
7
0
9
-
5
7
1
3
,
2
0
2
0
,
d
o
i:
1
0
.
1
1
5
9
1
/
ij
e
c
e
.
v
1
0
i6
.
[2
7
]
M
.
A
.
N
.
Ha
k
im,
Ad
iwij
a
y
a
,
a
n
d
W
.
As
tu
ti
,
"
C
o
m
p
a
ra
ti
v
e
a
n
a
ly
sis
of
Re
li
e
fF
-
S
VM
a
n
d
CF
S
-
S
V
M
fo
r
m
icro
a
rra
y
d
a
ta
c
las
sifica
ti
o
n
,
"
I
n
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
Co
m
p
u
ter
En
g
in
e
e
rin
g
,
v
o
l.
1
1
,
n
o
.
4,
pp.
3
3
9
3
-
3
4
0
2
,
2
0
2
1
,
d
o
i:
1
0
.
1
1
5
9
1
/i
jec
e
.
v
1
1
i
4
.
p
p
3
3
9
3
-
3
4
0
2
.
[2
8
]
H
.
F
.
Ka
re
e
m
,
M
.
S
.
AL
-
Hu
se
i
n
y
,
F
.
Y.
M
o
h
se
n
,
E
.
A.
Kh
a
li
l
,
a
n
d
Z
.
S
.
Ha
ss
a
n
,
"
E
v
a
lu
a
ti
o
n
of
S
V
M
p
e
rf
o
rm
a
n
c
e
in
th
e
d
e
tec
ti
o
n
of
l
u
n
g
c
a
n
c
e
r
in
m
a
rk
e
d
CT
sc
a
n
d
a
tas
e
t
,
"
I
n
d
o
n
e
sia
n
J
o
u
r
n
a
l
o
f
E
lec
trica
l
E
n
g
i
n
e
e
rin
g
a
n
d
Co
mp
u
ter
S
c
ien
c
e
(
IJ
EE
CS
)
,
v
o
l.
21
n
o
.
3,
p
p
.
1
7
3
1
-
1
7
3
8
,
2
0
2
1
,
d
o
i:
1
0
.
1
1
5
9
1
/i
jee
c
s.v
2
1
.
i
3
.
p
p
1
7
3
1
-
1
7
3
8
.
[2
9
]
A.
S
e
lwa
l
a
n
d
I.
Ra
o
o
f
,
“
A
M
u
lt
i
-
lay
e
r
p
e
rc
e
p
tr
o
n
b
a
se
d
in
telli
g
e
n
t
t
h
y
r
o
id
d
ise
a
se
p
re
d
ictio
n
sy
ste
m
,”
In
d
o
n
e
si
a
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
E
n
g
in
e
e
rin
g
a
n
d
Co
mp
u
ter
S
c
ien
c
e
,
v
o
l.
1
7
,
n
o
.
1,
p
p
.
5
2
4
-
5
3
2
.
2
0
2
0
,
d
o
i:
1
0
.
1
1
5
9
1
/i
jee
c
s.v
1
7
.
i
1
.
p
p
5
2
4
-
5
3
2
.
[3
0
]
N.
Ra
ju
,
H.
B.
An
i
ta,
a
n
d
P
.
Au
g
u
sti
n
e
,
"
Id
e
n
ti
fica
ti
o
n
of
i
n
ters
ti
ti
a
l
lu
n
g
d
ise
a
se
s
u
si
n
g
d
e
e
p
lea
rn
i
n
g
"
,
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
tr
ica
l
a
n
d
Co
mp
u
ter
En
g
in
e
e
rin
g
,
v
o
l.
1
0
,
n
o
.
6,
p
p
.
6
2
8
3
-
6
2
9
1
,
2
0
2
0
,
d
o
i:
1
0
.
1
1
5
9
1
/
ij
e
c
e
.
v
1
0
i
6
.
p
p
6
2
8
3
-
6
2
9
1
.
[3
1
]
N.
M
.
Ali,
N.
A.
A.
Az
iz,
a
n
d
R.
Be
sa
r,
“
Co
m
p
a
riso
n
of
m
icro
a
rra
y
b
re
a
st
c
a
n
c
e
r
c
las
sifica
ti
o
n
u
si
n
g
su
p
p
o
r
t
v
e
c
to
r
m
a
c
h
in
e
a
n
d
lo
g
isti
c
re
g
re
ss
io
n
wit
h
LAS
S
O
a
n
d
b
o
r
u
ta
fe
a
tu
re
se
lec
ti
o
n
,”
In
d
o
n
e
sia
n
J
o
u
rn
a
l
o
f
E
lec
trica
l
En
g
i
n
e
e
rin
g
a
n
d
C
o
mp
u
ter
S
c
ien
c
e
,
v
o
l.
2
0
,
no.
2,
p
p
.
7
1
2
-
7
1
9
,
2
0
2
0
,
d
o
i:
1
0
.
1
1
5
9
1
/i
jee
c
s.v
2
0
.
i
2
.
p
p
7
1
2
-
7
1
9
.
[3
2
]
A.
A.
Ab
d
u
ll
a
h
,
A.
F
.
D.
G
io
n
g
,
a
n
d
N.
A.
H.
Zah
ri
,
"
Ce
r
v
ica
l
c
a
n
c
e
r
d
e
tec
ti
o
n
m
e
th
o
d
u
sin
g
an
imp
ro
v
e
d
c
e
ll
u
la
r
n
e
u
ra
l
n
e
two
rk
(CNN
)
a
lg
o
r
it
h
m
,
"
I
n
d
o
n
e
sia
n
J
o
u
r
n
a
l
o
f
El
e
c
trica
l
En
g
in
e
e
rin
g
a
n
d
C
o
mp
u
ter
S
c
ien
c
e
,
v
o
l.
1
4
,
n
o
.
1
,
pp.
2
1
0
-
2
1
8
,
2
0
1
9
,
d
o
i:
1
0
.
1
1
5
9
1
/i
jee
c
s.v
1
4
.
i1
.
p
p
2
1
0
-
2
1
8
.
[3
3
]
R.
D
.
Ab
d
u
-
a
lj
a
b
a
r
a
n
d
O.
A
.
Aw
a
d
,
“
A
Co
m
p
a
ra
ti
v
e
a
n
a
ly
sis
stu
d
y
of
lu
n
g
c
a
n
c
e
r
d
e
tec
ti
o
n
a
n
d
re
l
a
p
se
p
re
d
ictio
n
u
sin
g
XG
Bo
o
st
c
las
sifier
,”
IO
P
Co
n
f.
S
e
r.
M
a
ter
.
S
c
i
.
E
n
g
.,
v
o
l.
1
0
7
6
,
no.
1,
2
0
2
1
,
d
o
i:
1
0
.
1
0
8
8
/
1
7
5
7
-
8
9
9
X/
1
0
7
6
/1
/
0
1
2
0
4
8
.
[3
4
]
T.
Ch
e
n
a
n
d
C.
G
u
e
strin
,
“
XG
Bo
o
st:
A
S
c
a
lab
le
Tree
Bo
o
stin
g
S
y
ste
m
,”
In
Pro
c
.
of
th
e
2
2
n
d
ACM
S
IGKD
D
In
t.
Co
n
f.
on
Kn
o
wled
g
e
Disc
o
v
e
ry
and
D
a
ta
M
i
n
in
g
,
2
0
1
6
,
p
p
.
7
8
5
-
7
9
4
,
d
o
i:
1
0
.
1
1
4
5
/
2
9
3
9
6
7
2
.
2
9
3
9
7
8
5
.
[3
5
]
S.
S.
Dh
a
li
wa
l,
A
-
A
.
Na
h
i
d
,
a
n
d
R.
Ab
b
a
s,
"
Eff
e
c
ti
v
e
I
n
tru
si
o
n
De
tec
ti
o
n
S
y
ste
m
Us
in
g
XG
Bo
o
st
,
"
In
f
o
rm
a
ti
o
n
,
v
o
l.
9,
n
o
.
7
,
2
0
1
8
,
d
o
i:
1
0
.
3
3
9
0
/i
n
fo
9
0
7
0
1
4
9
.
[3
6
]
Y.
Ch
u
a
n
d
D.
R.
Co
re
y
,
"
RN
A
se
q
u
e
n
c
in
g
:
p
latfo
rm
se
lec
ti
o
n
,
e
x
p
e
rime
n
tal
d
e
sig
n
,
a
n
d
d
a
t
a
in
terp
re
tatio
n
,
"
Nu
c
leic
Acid
T
h
e
ra
p
e
u
ti
c
s
,
v
o
l.
2
2
,
n
o
.
4
pp:
2
7
1
-
2
7
4
,
2
0
1
2
,
d
o
i:
1
0
.
1
0
8
9
/n
a
t
.
2
0
1
2
.
0
3
6
7
.
[3
7
]
S
.
Ro
u
ss
e
a
u
x
et
a
l.
,
“
Ecto
p
ic
a
c
ti
v
a
ti
o
n
of
g
e
rm
li
n
e
a
n
d
p
lac
e
n
t
a
l
g
e
n
e
s
id
e
n
ti
fies
a
g
g
re
ss
iv
e
m
e
tas
tas
is
-
p
ro
n
e
Lu
n
g
c
a
n
c
e
rs
,”
S
c
ien
c
e
T
ra
n
sla
ti
o
n
a
l
M
e
d
icin
e
,
v
o
l.
22
,
n
o
.
5
,
2
0
1
3
,
d
o
i:
1
0
.
1
1
2
6
/sc
it
ra
n
slm
e
d
.
3
0
0
5
7
2
3
.
[3
8
]
A.
M
e
z
h
e
y
e
u
s
k
i
e
t.
a
l
,
"
M
u
lt
is
p
e
c
tral
ima
g
in
g
f
o
r
q
u
a
n
ti
tati
v
e
a
n
d
c
o
m
p
a
rtme
n
t
-
sp
e
c
ifi
c
imm
u
n
e
i
n
fil
trate
s
re
v
e
a
ls
d
isti
n
c
t
imm
u
n
e
p
r
o
fil
e
s
t
h
a
t
c
las
sify
L
u
n
g
c
a
n
c
e
r
p
a
ti
e
n
ts
,
"
J
P
a
th
o
l
,
v
o
l.
2
4
4
,
n
o
.
4,
p
p
.
4
2
1
-
4
3
1
,
2
0
1
8
,
d
o
i:
1
0
.
1
0
0
2
/
p
a
th
.
5
0
2
6
.
[3
9
]
D.
C
.
Be
ll
et
al
.,
"
DN
A
b
a
se
id
e
n
ti
fica
ti
o
n
by
e
lec
tro
n
m
icro
sc
o
p
y
,
"
M
icr
o
sc
M
icr
o
a
n
a
l
,
v
o
l
.
1
8
,
no.
5
,
p
p
.
10
49
-
10
5
3
,
2
0
1
2
,
d
o
i:
1
0
.
1
0
1
7
/S
1
4
3
1
9
2
7
6
1
2
0
1
2
6
1
5
.
[4
0
]
F.
Oz
so
lak
a
n
d
P
.
M
.
M
il
o
s
,
"
RNA
se
q
u
e
n
c
in
g
:
Ad
v
a
n
c
e
s,
c
h
a
ll
e
n
g
e
s
a
n
d
o
p
p
o
rtu
n
it
ies
,
"
N
a
t
Rev
Ge
n
e
t.
,
v
o
l
.
12
,
pp.
87
-
9
8
,
2
0
1
1
,
d
o
i:
1
0
.
1
0
3
8
/
n
rg
2
9
3
4
.
[4
1
]
L.
Wan
g
,
“
S
u
p
p
o
rt
Vec
to
r
M
a
c
h
in
e
s:
T
h
e
o
ry
a
n
d
Ap
p
li
c
a
ti
o
n
s
,
”
B
e
r
l
i
n
,
G
e
r
m
a
n
y
:
S
p
ri
n
g
e
r,
2
0
0
5
,
d
o
i:
1
0
.
1
0
0
7
/
b
9
5
4
3
9
.
[4
2
]
Z
-
H
.
Zh
o
u
a
n
d
J.
F
e
n
g
,
"
De
e
p
F
o
re
st:
to
wa
rd
s
an
a
lt
e
rn
a
ti
v
e
to
d
e
e
p
n
e
u
ra
l
n
e
two
rk
s
,
"
Pro
c
e
e
d
in
g
s
o
f
th
e
T
we
n
ty
-
S
ixth
I
n
ter
n
a
ti
o
n
a
l
J
o
i
n
t
Co
n
fer
e
n
c
e
o
n
Arti
fi
c
ia
l
In
telli
g
e
n
c
e
,
2
0
1
7
,
p
p
.
3
5
5
3
-
3
5
5
9
,
d
o
i:
1
0
.
2
4
9
6
3
/
i
jca
i.
2
0
1
7
/
4
9
7
.
Evaluation Warning : The document was created with Spire.PDF for Python.