I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
8
,
No
.
6
,
Decem
b
er
201
8
,
p
p
.
5
4
1
5
~
5
4
2
4
I
SS
N:
2088
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
8
i
6
.
pp
5
4
1
5
-
5424
5415
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
co
r
e
.
co
m/
jo
u
r
n
a
ls
/in
d
ex
.
p
h
p
/
I
JE
C
E
I
m
pro
v
ing
H
iera
rchica
l Dec
isio
n
Appro
a
ch f
o
r
Sin
g
le I
m
a
g
e
Cla
ss
ificatio
n of
Pap S
m
ea
r
Dw
iza
Ria
na
1
,
Yudi
Ra
m
d
h
a
ni
2
,
Riz
k
i Tri P
ra
s
et
io
3
,
Ac
h
m
a
d Niz
a
r
H
ida
y
a
nto
4
1
S
T
M
IK Nu
sa
M
a
n
d
iri
,
In
d
o
n
e
sia
2,
3
Un
iv
e
rsitas
BS
I,
In
d
o
n
e
sia
4
Un
iv
e
rsitas
In
d
o
n
e
sia
,
In
d
o
n
e
sia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ma
r
2
8
,
2
0
1
8
R
ev
i
s
ed
J
u
l
2
7
,
2
0
1
8
A
cc
ep
ted
A
u
g
7
,
2
0
1
8
T
h
e
sin
g
le
im
a
g
e
c
las
sif
ic
a
ti
o
n
o
f
P
a
p
sm
e
a
rs
is
a
n
im
p
o
rtan
t
p
a
rt
o
f
th
e
e
a
rl
y
d
e
tec
ti
o
n
o
f
c
e
rv
ica
l
c
a
n
c
e
r
th
ro
u
g
h
P
a
p
sm
e
a
r
tes
ts.
Un
fo
rtu
n
a
tely
,
m
o
st
c
la
ss
i
f
ica
ti
o
n
p
ro
c
e
ss
e
s
stil
l
re
q
u
ire
a
c
c
u
ra
c
y
e
n
h
a
n
c
e
m
e
n
t,
e
sp
e
c
iall
y
to
c
o
m
p
lete
th
e
c
las
sif
ic
a
ti
o
n
in
se
v
e
n
c
las
se
s
a
n
d
to
g
e
t
a
q
u
a
li
f
ied
c
las
si
f
ica
ti
o
n
p
r
o
c
e
ss
.
In
a
d
d
it
i
o
n
,
a
t
tem
p
ts
to
im
p
ro
v
e
th
e
sin
g
le
im
a
g
e
c
las
si
f
ica
ti
o
n
o
f
P
a
p
sm
e
a
r
s
w
e
re
p
e
rf
o
rm
e
d
to
b
e
a
b
le
t
o
d
isti
n
g
u
i
sh
n
o
rm
a
l
a
n
d
a
b
n
o
rm
a
l
c
e
ll
s.
T
h
is
stu
d
y
p
ro
p
o
se
s
a
b
e
tt
e
r
a
p
p
ro
a
c
h
b
y
p
ro
v
i
d
in
g
d
if
fe
re
n
t
h
a
n
d
li
n
g
o
f
th
e
in
it
ial
d
a
ta
p
re
p
a
ra
ti
o
n
p
r
o
c
e
ss
in
th
e
fo
rm
o
f
th
e
d
istri
b
u
ti
o
n
f
o
r
train
in
g
d
a
ta
a
n
d
tes
ti
n
g
d
a
ta
so
th
a
t
it
re
su
lt
e
d
in
a
n
e
w
m
o
d
e
l
o
f
Hie
ra
rc
h
ial
De
c
isio
n
A
p
p
ro
a
c
h
(HD
A
)
w
h
ich
h
a
s
th
e
h
ig
h
e
r
lea
rn
in
g
ra
te
a
n
d
m
o
m
e
n
tu
m
v
a
l
u
e
s
in
th
e
p
ro
p
o
se
d
n
e
w
m
o
d
e
l.
T
h
is
stu
d
y
e
v
a
lu
a
ted
2
0
d
if
fe
re
n
t
f
e
a
tu
r
e
s
in
h
iera
rc
h
ica
l
d
e
c
isio
n
a
p
p
r
o
a
c
h
m
o
d
e
l
b
a
se
d
o
n
Ne
u
ra
l
Ne
tw
o
rk
(NN
)
a
n
d
g
e
n
e
ti
c
a
lg
o
rit
h
m
m
e
th
o
d
f
o
r
sin
g
le
i
m
a
g
e
c
las
si
f
ica
ti
o
n
o
f
P
a
p
sm
e
a
r
w
h
ic
h
re
su
lt
e
d
i
n
c
las
sif
ica
ti
o
n
e
x
p
e
ri
m
e
n
t
u
sin
g
v
a
lu
e
lea
rn
in
g
ra
te
o
f
0
.
3
a
n
d
m
o
m
e
n
tu
m
o
f
0
.
2
a
n
d
v
a
lu
e
o
f
lea
rn
in
g
ra
te
o
f
0.
5
a
n
d
m
o
m
e
n
tu
m
o
f
0
.
5
b
y
g
e
n
e
ra
ti
n
g
c
las
sif
ica
ti
o
n
o
f
7
c
las
se
s
(No
rm
a
l
In
term
e
d
iate
,
No
r
m
a
l
Co
lu
m
m
a
r,
M
il
d
(L
ig
h
t)
D
y
p
las
ia,
M
o
d
e
ra
te
D
y
p
las
ia,
S
e
rv
e
re
D
y
p
las
ia
a
n
d
Ca
rc
in
o
m
a
In
S
it
u
)
b
e
tt
e
r.
T
h
e
a
c
c
u
r
a
c
y
v
a
lu
e
e
n
h
a
n
c
e
m
e
n
e
t
w
e
re
a
lso
in
f
lu
e
n
c
e
d
b
y
th
e
a
p
p
li
c
a
ti
o
n
o
f
G
e
n
e
ti
c
A
l
g
o
rit
h
m
to
f
e
a
tu
re
se
le
c
ti
o
n
.
T
h
u
s,
f
ro
m
th
e
re
su
lt
s
o
f
m
o
d
e
l
tes
ti
n
g
,
it
c
a
n
b
e
c
o
n
c
lu
d
e
d
t
h
a
t
t
h
e
Hie
ra
rc
h
ica
l
De
c
isio
n
A
p
p
ro
a
c
h
(HD
A
)
m
e
th
o
d
f
o
r
P
a
p
S
m
e
a
r
i
m
a
g
e
c
las
sif
i
c
a
ti
o
n
c
a
n
b
e
u
se
d
a
s
a
re
fe
re
n
c
e
f
o
r
in
it
i
a
l
sc
re
e
n
in
g
p
ro
c
e
ss
to
a
n
a
ly
z
e
P
a
p
S
m
e
a
r
ima
g
e
c
las
si
f
ic
a
ti
o
n
.
K
ey
w
o
r
d
:
C
er
v
ical
c
a
n
ce
r
Gen
etic
a
l
g
o
r
ith
m
Hier
ar
ch
ical
Dec
is
io
n
A
p
p
r
o
ac
h
(
H
A
D)
Neu
r
al
Net
w
o
r
k
(
NN)
P
ap
s
m
ea
r
Co
p
y
rig
h
t
©
2
0
1
8
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts
re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
D
w
iza
R
ia
n
a
,
ST
MI
K
Nu
s
a
Ma
n
d
ir
i J
ak
ar
ta,
J
alan
Da
m
ai
n
o
8
J
ak
ar
ta
Selatan
,
I
n
d
o
n
esia
.
E
m
ail:
d
w
iza
@
n
u
s
a
m
a
n
d
ir
i.a
c.
id
1.
I
NT
RO
D
UCT
I
O
N
R
esear
ch
o
n
t
h
e
class
if
icatio
n
o
f
s
in
g
le
P
ap
s
m
ea
r
i
m
ag
e
h
a
s
b
ee
n
d
o
n
e.
T
h
is
atte
m
p
t
w
a
s
in
te
n
d
ed
to
d
ig
itize
th
e
in
tr
o
d
u
ctio
n
o
f
ea
r
l
y
d
etec
tio
n
o
f
ce
r
v
ical
ca
n
ce
r
.
As
k
n
o
w
n
t
h
at
o
n
e
t
y
p
e
o
f
m
ali
g
n
an
t
ca
n
ce
r
th
at
attac
k
s
w
o
m
e
n
ac
co
r
d
in
g
to
W
HO
b
o
d
y
w
it
h
th
e
m
as
s
iv
e
n
u
m
b
er
o
f
p
atien
t
s
in
I
n
d
o
n
esi
a
is
ce
r
v
ica
l
ca
n
ce
r
.
I
t
i
s
n
o
w
o
n
d
er
t
h
at
I
n
d
o
n
esia
b
ec
a
m
e
o
n
e
o
f
th
e
co
u
n
tr
ie
s
t
h
at
h
a
v
e
a
lo
t
o
f
ce
r
v
i
ca
l
ca
n
ce
r
p
atie
n
t
s
.
C
er
v
ical
ca
n
ce
r
is
g
en
er
all
y
ca
u
s
ed
b
y
a
v
ir
u
s
ca
lled
Hu
m
an
P
ap
illo
m
a
Vir
u
s
(
HP
V)
.
Sex
u
al
i
n
ter
co
u
r
s
e
b
ec
am
e
t
h
e
lar
g
est ca
s
e
o
f
H
P
V
[
1
]
.
P
ap
s
m
ea
r
is
a
m
et
h
o
d
o
f
ea
r
l
y
d
etec
tio
n
o
f
ce
r
v
ical
ca
n
c
er
.
T
h
e
p
r
o
ce
s
s
ap
p
lied
o
n
P
ap
s
m
ea
r
co
n
tin
u
o
u
s
l
y
a
n
d
co
n
s
is
te
n
tl
y
i
n
a
co
u
n
tr
y
w
i
ll
h
elp
p
r
ev
en
t
ea
r
l
y
ce
r
v
ical
ca
n
ce
r
.
T
h
is
m
et
h
o
d
w
a
s
p
er
f
o
r
m
ed
b
y
a
P
at
h
o
lo
g
is
t
i
n
a
cli
n
ical
p
at
h
o
lo
g
y
lab
o
r
ato
r
y
,
in
w
h
ich
test
s
w
er
e
p
er
f
o
r
m
ed
o
n
a
w
o
m
a
n
's
s
q
u
a
m
o
u
s
ep
ith
eli
u
m
.
T
h
e
r
esu
lt
s
o
f
p
ath
o
lo
g
i
s
t
'
s
ex
a
m
i
n
atio
n
w
it
h
a
P
ap
s
m
ea
r
w
ill
s
h
o
w
w
h
et
h
er
t
h
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
6
,
Decem
b
er
201
8
:
5
4
1
5
-
5
4
2
4
5416
w
o
m
an
h
as
n
o
r
m
al
o
r
ab
n
o
r
m
al
ce
lls
[
2
]
.
T
h
er
e
ar
e
v
ar
io
u
s
clas
s
if
icatio
n
s
i
n
P
ap
S
m
e
ar
,
b
u
t
i
n
th
i
s
s
t
u
d
y
,
P
ap
s
m
ea
r
i
m
ag
e
s
ar
e
class
i
f
ied
u
p
to
7
class
es
[
3
]
,
in
w
h
ich
t
h
e
f
ir
s
t
t
h
r
ee
class
e
s
ar
e
n
o
r
m
al
ce
ll
clas
s
ca
teg
o
r
ies
i
n
clu
d
i
n
g
No
r
m
al
Su
p
er
f
icia
l,
No
r
m
al
I
n
ter
m
ed
iate,
an
d
No
r
m
al
C
o
l
u
m
m
ar
w
h
ile
t
h
e
n
e
x
t
f
o
u
r
class
es
o
f
ab
n
o
r
m
al
ce
ll
ca
te
g
o
r
ies
ar
e:
M
ild
(
L
i
g
h
t)
D
y
p
la
s
ia,
Mo
d
er
ate
D
y
p
lasi
a,
D
y
p
la
s
ia
a
n
d
C
ar
ci
n
o
m
a
I
n
Sit
u
[
4
]
.
Gen
er
al
e
x
a
m
in
a
tio
n
u
s
ed
to
d
etec
t
ce
r
v
ical
ca
n
ce
r
i
n
P
ap
s
m
ea
r
m
et
h
o
d
is
to
p
r
ev
en
t
a
n
d
d
etec
t
t
h
e
p
r
esen
ce
o
f
p
r
e
-
ca
n
ce
r
a
n
d
ca
n
ce
r
s
it
u
atio
n
i
n
ce
r
v
ical
ce
ll
s
a
m
p
le
s
.
T
h
e
p
r
o
b
lem
o
f
P
ap
s
m
ea
r
i
m
a
g
e
class
i
f
icatio
n
is
ca
u
s
ed
b
y
P
ap
s
m
ea
r
i
m
ag
e
h
av
i
n
g
u
n
iq
u
e
ch
ar
ac
ter
is
tic
s
o
th
at
th
e
au
to
m
atic
id
en
ti
f
icat
io
n
o
f
P
ap
s
m
ea
r
i
m
a
g
e
is
a
ch
al
len
g
in
g
p
r
o
b
le
m
f
o
r
r
esear
ch
er
s
.
Dif
f
er
en
t
ce
ll
co
n
d
itio
n
s
an
d
s
tr
u
ct
u
r
es
w
i
t
h
h
ig
h
v
ar
iatio
n
s
o
f
i
m
a
g
e
co
n
d
i
tio
n
s
m
a
k
e
th
e
id
e
n
ti
f
icatio
n
a
n
d
class
i
f
icat
io
n
p
r
o
ce
s
s
o
f
t
h
e
P
ap
s
m
ea
r
i
m
a
g
e
n
ee
d
s
p
ec
ial
h
a
n
d
li
n
g
.
P
ar
ticu
lar
l
y
t
h
e
p
r
o
ce
s
s
o
f
P
ap
s
m
ea
r
i
m
a
g
e
cla
s
s
i
f
icatio
n
u
n
til
n
o
w
is
s
t
ill
ex
p
er
ien
ci
n
g
d
i
f
f
icu
lties
a
n
d
r
eq
u
ir
es tec
h
n
iq
u
e
s
an
d
m
et
h
o
d
s
o
f
class
i
f
icat
io
n
t
h
at
h
a
v
e
a
h
ig
h
ac
cu
r
ac
y
.
T
h
e
u
s
e
o
f
d
ata
m
in
i
n
g
s
o
f
ar
is
co
m
m
o
n
l
y
u
s
ed
to
o
b
tain
o
p
ti
m
al
i
n
f
o
r
m
atio
n
f
r
o
m
a
lar
g
e
g
r
o
u
p
o
f
lar
g
e
d
atab
ases
t
h
at
h
av
e
co
m
p
lex
i
t
y
[
5
]
.
I
n
a
s
t
u
d
y
o
f
s
i
n
g
le
P
ap
s
m
ea
r
i
m
ag
e
clas
s
i
f
icatio
n
f
o
u
n
d
i
n
t
h
e
Her
l
ev
d
ataset
[
4
]
,
d
ata
m
i
n
in
g
w
as
u
s
ed
to
g
et
i
n
f
o
r
m
atio
n
f
r
o
m
2
0
f
ea
t
u
r
es i
n
t
h
e
d
ata
to
id
en
ti
f
y
p
at
h
o
lo
g
i
c
ca
s
es
o
f
ce
r
v
ical
ca
n
ce
r
.
T
h
e
p
r
ev
io
u
s
r
e
s
ea
r
ch
es
w
h
ich
ai
m
ed
to
id
e
n
ti
f
y
p
ath
o
lo
g
ical
ca
s
es
w
i
th
t
h
e
s
a
m
e
d
ataset
in
cl
u
d
e
th
e
s
t
u
d
y
o
f
class
i
f
icati
o
n
m
et
h
o
d
s
o
n
n
o
r
m
al
cla
s
s
i
m
ag
e
s
[
6
-
8
]
an
d
class
if
ica
tio
n
o
f
ab
n
o
r
m
al
class
e
s
[
9
]
.
B
esid
es
th
e
class
if
ica
tio
n
o
f
p
r
ev
io
u
s
r
esear
ch
f
o
r
m
s
,
s
o
m
e
r
esear
ch
er
s
ai
m
to
s
eg
m
e
n
t
th
e
P
ap
s
m
ea
r
i
m
a
g
e
[
6
]
,
[
1
0
]
.
E
v
en
th
e
e
f
f
o
r
t
to
id
en
ti
f
y
t
h
e
b
est
f
ea
t
u
r
es
to
s
o
lv
e
t
h
e
p
ath
o
lo
g
ical
ca
s
e
o
f
ce
r
v
ical
ca
n
ce
r
h
as
a
ls
o
b
ee
n
d
o
n
e.
Featu
r
e
[
1
1
]
an
d
tex
t
u
r
e
an
al
y
s
i
s
[
6
]
,
[
1
2
]
a
r
e
s
o
m
e
o
f
th
e
e
x
a
m
p
le
s
.
T
h
e
co
m
b
i
n
atio
n
o
f
s
e
v
er
al
f
ea
tu
r
es
(
2
0
f
ea
t
u
r
es)
r
ef
er
r
i
n
g
to
7
class
e
s
o
f
d
iv
er
s
e
ca
s
es
o
f
p
ath
o
lo
g
ical
ca
n
ce
r
,
ca
u
s
i
n
g
d
if
f
icu
lties
i
n
th
e
clas
s
if
ica
tio
n
f
o
r
7
class
es
in
th
i
s
P
ap
s
m
ea
r
i
m
a
g
e
w
h
er
e
it
r
e
m
ain
s
a
ch
alle
n
g
e
f
o
r
r
esear
ch
er
s
.
So
m
e
al
g
o
r
ith
m
s
ai
m
ed
at
s
elec
tin
g
f
ea
tu
r
es
s
u
ch
as
g
en
e
tic
alg
o
r
it
h
m
s
[
1
3
]
p
er
f
o
r
m
a
f
ea
t
u
r
e
s
elec
tio
n
p
r
o
ce
s
s
b
y
s
elec
ti
n
g
s
o
m
e
o
f
th
e
b
es
t
i
n
d
iv
id
u
als.
I
n
d
iv
id
u
al
ta
k
i
n
g
s
h
o
u
ld
b
e
d
o
n
e
r
an
d
o
m
l
y
a
n
d
p
r
o
p
o
r
tio
n
all
y
in
c
lu
d
i
n
g
t
h
e
p
r
o
p
o
r
ti
o
n
o
f
its
q
u
ali
t
y
.
T
h
e
p
r
o
p
o
s
ed
HDA
class
i
f
ica
t
io
n
m
o
d
el
o
n
s
in
g
le
P
ap
s
m
ea
r
im
a
g
e
w
as
s
tar
ted
[
1
4
]
f
r
o
m
w
h
e
n
th
e
P
ap
s
m
ea
r
cla
s
s
i
f
icat
io
n
m
o
d
el
o
f
f
er
ed
n
e
w
p
r
o
ce
s
s
s
ta
g
e
s
b
y
u
tili
zi
n
g
b
o
th
q
u
a
n
ti
tati
v
e
an
d
q
u
alitati
v
e
f
ea
t
u
r
es
t
h
at
w
a
s
u
tili
za
tio
n
o
f
I
m
p
o
r
tan
ce
P
er
f
o
r
m
a
n
ce
An
al
y
s
i
s
as
th
e
b
as
is
o
f
th
e
p
r
o
p
o
s
ed
m
u
lt
i
-
s
ta
g
e
class
i
f
icatio
n
.
T
h
e
r
esu
lts
o
f
th
is
s
tu
d
y
s
till
h
a
v
e
d
if
f
ic
u
lti
es
in
class
if
ica
tio
n
f
o
r
m
o
d
e
r
ate
d
y
s
p
las
ia
an
d
s
ev
er
e
d
y
s
p
la
s
ia
clas
s
[
1
4
]
.
T
h
e
n
ex
t
atte
m
p
t
to
clas
s
i
f
y
t
h
e
i
m
ag
e
o
f
ce
r
v
ical
ca
n
ce
r
was
to
ap
p
ly
th
e
Ge
n
etic
Alg
o
r
i
th
m
(
G
A
)
f
o
r
f
ea
tu
r
e
s
elec
tio
n
.
Fu
r
t
h
er
m
o
r
e,
to
class
i
f
y
h
ea
lt
h
y
ce
lls
an
d
ca
n
ce
r
ce
lls
,
w
e
u
s
e
d
SVM
alg
o
r
ith
m
(
Su
p
p
o
r
t
Vec
to
r
Ma
ch
in
e)
[
1
5
]
.
T
h
e
r
esu
lts
s
h
o
w
th
a
t
g
en
etic
alg
o
r
ith
m
is
a
b
etter
m
et
h
o
d
f
o
r
s
elec
tio
n
o
f
f
ea
t
u
r
es a
n
d
o
p
ti
m
izat
io
n
o
f
p
ar
a
m
eter
s
.
I
n
th
i
s
s
t
u
d
y
,
NN
w
as
s
elec
ted
as
a
to
o
l
o
f
an
al
y
s
i
s
o
n
P
ap
s
m
ea
r
i
m
a
g
e
d
ataset
u
s
ed
.
T
h
e
u
s
e
o
f
t
h
i
s
alg
o
r
ith
m
w
a
s
to
m
a
k
e
d
ata
p
r
ed
ictio
n
an
d
id
en
tify
p
ath
o
lo
g
ical
ca
s
es
o
f
ce
r
v
ical
ca
n
ce
r
to
b
e
h
an
d
led
.
T
h
e
u
s
e
o
f
NN
f
o
r
m
ed
ical
d
ata
class
if
icatio
n
is
co
m
m
o
n
l
y
u
s
ed
s
u
c
h
as
clas
s
if
icatio
n
t
o
p
r
e
d
ict
m
o
r
talit
y
p
r
ed
ictio
n
[
1
6
]
.
O
p
ti
m
izatio
n
o
n
NN
alg
o
r
ith
m
ca
n
b
e
d
o
n
e
w
it
h
th
e
ai
m
o
f
i
m
p
r
o
v
i
n
g
t
h
e
p
er
f
o
r
m
an
ce
o
f
NN
[
1
7
]
.
T
h
e
m
o
s
t
co
m
m
o
n
l
y
u
s
ed
o
p
ti
m
iza
tio
n
m
et
h
o
d
is
GA
,
P
ar
ticle
S
w
ar
m
Op
ti
m
iza
tio
n
(
P
SO)
,
an
d
An
t
C
o
lo
n
y
Op
ti
m
izatio
n
[
1
8
]
.
I
n
th
i
s
s
t
u
d
y
,
G
A
w
as
s
elec
ted
as
a
f
ea
tu
r
e
s
elec
tio
n
a
lg
o
r
it
h
m
.
G
A
i
s
o
n
e
o
f
alg
o
r
ith
m
s
t
h
at
ca
n
s
elec
t
a
r
elev
an
t
f
ea
t
u
r
e
s
u
b
s
e
t,
lear
n
in
g
r
ate,
m
o
m
e
n
t
u
m
,
a
n
d
in
itial
izatio
n
an
d
w
eig
h
t
o
p
tim
izatio
n
.
B
ased
o
n
th
e
p
r
ev
io
u
s
r
esear
c
h
[
1
9
]
,
w
e
f
o
cu
s
ed
t
h
i
s
r
esear
ch
to
i
m
p
r
o
v
e
clas
s
if
icatio
n
a
cc
u
r
ac
y
i
n
th
e
b
est
m
o
d
el
o
f
t
h
e
clas
s
i
f
icatio
n
r
esu
l
t
b
ased
o
n
th
e
HDA
m
o
d
el
f
o
r
s
i
n
g
le
-
ce
l
l
P
ap
s
m
ea
r
i
m
ag
e
class
i
f
icatio
n
.
T
h
e
co
m
p
ar
is
o
n
o
f
class
i
f
icatio
n
r
es
u
lt
s
w
a
s
d
o
n
e
b
y
u
s
in
g
NN
alg
o
r
ith
m
an
d
f
ea
t
u
r
e
o
p
tim
izatio
n
u
s
i
n
g
G
A
to
d
eter
m
in
e
t
h
e
in
cr
ea
s
e
o
f
ac
c
u
r
ac
y
.
T
h
e
r
esu
lts
s
h
o
w
t
h
at
th
e
r
e
is
a
s
ig
n
i
f
ica
n
t
in
cr
ea
s
e
o
f
ac
c
u
r
ac
y
f
r
o
m
t
h
e
p
r
o
p
o
s
ed
HDA
m
o
d
el.
I
n
th
i
s
p
ap
er
w
e
p
r
o
p
o
s
e
m
e
th
o
d
s
f
o
r
P
ap
s
m
ea
r
ce
ll
i
m
a
g
e
clas
s
i
f
icatio
n
a
i
m
ed
at
t
wo
s
p
ec
if
ic
o
b
j
ec
tiv
es:
a)
s
elec
tio
n
o
f
t
h
e
b
est
f
ea
t
u
r
es
o
n
2
0
f
ea
t
u
r
es
o
f
p
ap
s
m
ea
r
a
n
d
b
)
P
ap
s
m
ea
r
i
m
ag
e
c
la
s
s
if
ica
tio
n
ap
p
r
o
ac
h
u
s
in
g
h
ier
ar
ch
ial
d
ec
is
io
n
ap
p
r
o
ac
h
s
tag
e.
T
h
u
s
th
e
r
e
ar
e
t
w
o
m
ai
n
co
n
tr
ib
u
tio
n
s
in
o
u
r
p
ap
er
.
First,
f
ea
t
u
r
es
o
f
th
e
P
ap
s
m
ea
r
i
m
a
g
e
th
at
ar
e
n
o
t
r
elev
a
n
t
i
n
th
e
class
i
f
icatio
n
p
r
o
ce
s
s
ar
e
n
o
t
u
s
ed
lik
e
t
h
e
lo
n
g
est
d
ia
m
eter
n
u
cle
u
s
a
n
d
n
u
cle
u
s
r
o
u
n
d
n
e
s
s
.
Seco
n
d
,
th
e
u
s
e
s
o
f
th
e
h
ier
ar
c
h
ial
d
ec
is
io
n
a
p
p
r
o
ac
h
m
a
k
e
t
h
e
class
i
f
icatio
n
p
r
o
ce
s
s
m
o
r
e
ef
f
ec
tiv
e
a
n
d
in
cr
ea
s
e
th
e
ac
c
u
r
ac
y
o
f
clas
s
i
f
icatio
n
r
es
u
lt
s
.
I
n
th
is
w
a
y
th
e
au
to
m
at
ic
class
if
ica
tio
n
p
r
o
ce
s
s
to
h
elp
p
ath
o
lo
g
i
s
t a
llo
w
s
t
o
b
e
r
ea
lized
.
T
h
is
m
et
h
o
d
is
b
ased
o
n
f
ea
tu
r
e
s
elec
tio
n
f
o
r
les
s
r
elev
a
n
t
f
ea
tu
r
es
b
y
u
s
i
n
g
g
e
n
etic
al
g
o
r
ith
m
s
a
n
d
g
en
er
ate
s
r
elev
a
n
t
f
ea
t
u
r
es
to
b
e
u
s
ed
i
n
s
u
b
s
eq
u
e
n
t
cla
s
s
if
icatio
n
p
r
o
ce
s
s
es.
T
h
i
s
m
et
h
o
d
co
m
b
in
e
s
t
h
e
k
n
o
w
led
g
e
o
n
t
h
e
v
ar
iatio
n
s
o
f
clas
s
i
f
i
ca
tio
n
s
tag
e
s
b
et
w
e
en
P
ap
s
m
ea
r
a
n
d
h
ier
ar
c
h
ial
d
ec
is
io
n
ap
p
r
o
ac
h
class
b
y
o
p
ti
m
izi
n
g
t
h
e
v
alu
e
o
f
lear
n
i
n
g
r
ate
a
n
d
m
o
m
e
n
tu
m
o
n
NN
al
g
o
r
ith
m
.
B
as
ed
o
n
th
is
f
ac
t
w
e
p
r
o
p
o
s
e
a
m
e
th
o
d
t
h
at
ca
n
class
i
f
y
P
ap
s
m
ea
r
i
m
ag
e
i
n
to
7
class
e
s
w
h
ic
h
ar
e
3
n
o
r
m
al
cla
s
s
e
s
a
n
d
4
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
I
mp
r
o
vin
g
Hiera
r
ch
ica
l D
ec
is
io
n
A
p
p
r
o
a
ch
fo
r
S
i
n
g
le
I
ma
g
e
C
la
s
s
if
ica
tio
n
o
f P
a
p
S
mea
r
(
D
w
iz
a
R
ia
n
a
)
5417
ab
n
o
r
m
al
class
e
s
.
T
h
is
m
e
th
o
d
ex
p
lo
its
th
e
f
ea
tu
r
e
s
o
f
th
e
n
u
cle
u
s
an
d
c
y
to
p
las
m
t
h
r
o
u
g
h
f
ea
t
u
r
e
s
elec
tio
n
.
Fin
all
y
,
t
h
i
s
m
et
h
o
d
is
e
v
alu
a
t
ed
b
y
u
s
in
g
9
1
7
s
a
m
p
le
d
atas
et
an
d
h
as
2
0
f
ea
tu
r
e
s
,
d
iv
id
e
d
in
to
9
0
%
tr
ai
n
in
g
d
ata
an
d
1
0
%
test
in
g
d
ata.
T
h
e
ev
al
u
atio
n
p
r
o
ce
s
s
u
s
es
ap
p
licatio
n
s
b
u
ilt
to
s
u
p
p
o
r
t
th
e
p
r
o
p
o
s
ed
m
et
h
o
d
.
T
h
e
r
em
i
n
d
er
o
f
th
i
s
p
ap
er
is
o
r
g
an
ized
as
f
o
llo
w
s
:
s
ec
tio
n
2
ab
o
u
t
r
elate
d
w
o
r
k
,
s
ec
ti
o
n
3
ab
o
u
t
r
esear
ch
m
et
h
o
d
u
s
ed
in
t
h
e
s
tu
d
y
.
Se
ctio
n
4
d
escr
ib
es
th
e
r
esu
lt
s
an
d
an
al
y
s
i
s
,
th
e
n
f
o
llo
w
ed
b
y
co
n
cl
u
s
io
n
s
an
d
f
u
r
t
h
er
r
esear
ch
p
lan
s
.
2.
RE
L
AT
E
D
WO
RK
T
h
e
au
to
m
a
tic
h
y
b
r
id
s
eg
m
e
n
tatio
n
class
i
f
ica
tio
n
ap
p
r
o
ac
h
to
s
elec
t
an
d
en
h
an
ce
t
h
e
s
eg
m
en
tatio
n
o
f
n
u
cle
u
s
ce
lls
f
o
r
P
ap
s
m
ea
r
test
i
m
ag
e
s
b
y
u
s
i
n
g
n
e
s
ted
h
ier
ar
ch
ical
p
o
r
tio
n
in
g
,
s
eg
m
e
n
tatio
n
le
v
el
s
elec
tio
n
,
a
n
d
SVM
clas
s
i
f
ier
w
a
s
alr
ea
d
y
p
er
f
o
r
m
ed
.
T
h
e
p
u
r
p
o
s
e
o
f
m
er
g
in
g
t
h
e
en
d
o
f
t
h
e
s
e
g
m
e
n
tat
io
n
i
s
to
av
o
id
o
v
er
s
eg
m
e
n
tatio
n
.
T
h
e
s
eg
m
e
n
tatio
n
w
as
d
o
n
e
w
it
h
m
o
r
p
h
o
lo
g
ical
al
g
o
r
ith
m
(
w
ater
s
h
ed
)
a
n
d
h
ier
ar
ch
ical
m
er
g
in
g
(
w
ater
f
a
ll
)
alg
o
r
ith
m
b
ased
o
n
s
p
ec
tr
al
in
f
o
r
m
atio
n
an
d
s
h
ap
e
i
n
f
o
r
m
atio
n
a
s
w
ell
a
s
class
in
f
o
r
m
at
io
n
.
S
VM
cla
s
s
i
f
ier
i
s
u
s
ed
to
s
ep
ar
ate
t
w
o
cl
ass
es
o
f
r
e
g
io
n
s
t
h
at
ar
e
t
h
e
n
u
cleu
s
a
n
d
n
o
t
t
h
e
n
u
cle
u
s
ar
ea
(
c
y
to
p
las
m
a
n
d
b
ac
k
g
r
o
u
n
d
)
b
y
u
s
in
g
a
f
ea
t
u
r
e
s
et
(
m
o
r
p
h
o
m
etr
ic,
ed
g
e
-
b
ase
d
,
an
d
co
n
v
e
x
h
u
ll
-
b
ased
)
.
T
h
e
r
esu
lts
o
f
s
e
g
m
e
n
tatio
n
a
n
d
class
i
f
icat
io
n
w
er
e
co
m
p
ar
ed
w
it
h
t
h
e
s
eg
m
e
n
tatio
n
p
r
o
v
id
ed
b
y
p
ath
o
lo
g
is
t
a
n
d
s
h
o
w
ed
i
m
p
r
o
v
e
m
e
n
t
in
t
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
[
2
0
]
.
Un
f
o
r
t
u
n
atel
y
,
th
i
s
r
esear
ch
h
as
n
o
t
y
e
t
r
ea
ch
ed
th
e
c
las
s
if
icatio
n
p
r
o
ce
s
s
o
f
P
ap
s
m
ea
r
i
m
ag
e.
GA
h
a
s
b
ee
n
u
s
ed
i
n
p
r
ev
io
u
s
r
esear
ch
a
n
d
is
co
n
s
id
er
ed
as
a
b
etter
m
et
h
o
d
f
o
r
f
ea
t
u
r
e
s
el
ec
tio
n
a
n
d
p
ar
am
eter
o
p
ti
m
izatio
n
i
n
P
ap
s
m
ea
r
i
m
a
g
e
s
o
n
t
h
e
s
a
m
e
d
ataset
[
1
5
]
.
Su
p
p
o
r
t
Vec
to
r
Ma
ch
in
e
(
SV
M)
A
l
g
o
r
ith
m
is
u
s
ed
f
o
r
clas
s
i
f
i
ca
tio
n
.
W
ith
th
is
s
tr
u
c
tu
r
e,
n
e
w
ce
lls
ca
n
b
e
cla
s
s
i
f
ied
b
y
o
b
s
er
v
in
g
t
h
e
b
es
t
f
ea
t
u
r
e
v
al
u
es
f
o
r
ca
n
ce
r
ce
ll
class
i
f
icatio
n
as
ca
n
ce
r
ce
lls
o
r
b
en
ig
n
ce
l
ls
.
U
n
f
o
r
t
u
n
a
tel
y
,
th
e
r
e
s
u
lt
s
s
h
o
w
th
at
t
h
e
ef
f
ec
ti
v
e
n
ess
o
f
th
i
s
m
eth
o
d
h
as
n
o
t g
i
v
en
t
h
e
h
ig
h
es
t a
cc
u
r
ac
y
f
o
r
th
e
clas
s
if
icatio
n
o
f
7
class
e
s
[
1
5
]
.
T
h
e
h
y
b
r
id
en
s
e
m
b
le
tec
h
n
iq
u
e
is
u
s
ed
f
o
r
P
ap
s
m
ea
r
i
m
a
g
e
class
i
f
icatio
n
w
i
th
t
h
e
ad
d
itio
n
o
f
n
e
w
d
ata
[
2
1
]
[
2
2
]
.
B
y
co
m
p
ar
i
n
g
t
h
e
m
eth
o
d
s
o
f
NN
a
n
d
SVM.
T
h
e
r
esear
ch
s
tag
e
s
a
r
e
n
o
t
th
o
r
o
u
g
h
l
y
co
n
d
u
cted
i
n
all
clas
s
co
n
d
iti
o
n
s
,
s
o
th
e
r
es
u
lts
o
b
tain
ed
a
p
p
ly
o
n
l
y
to
t
h
e
clas
s
ac
co
r
d
in
g
to
t
h
e
s
i
m
p
l
if
ied
s
tag
e
s
w
h
er
e
th
e
r
esear
ch
d
o
e
s
n
o
t
p
r
o
d
u
ce
a
cla
s
s
i
f
icatio
n
m
o
d
el
o
f
7
cla
s
s
es
b
u
t
o
n
l
y
p
r
esen
ts
clas
s
r
ec
all
d
ata
[
2
1
]
.
T
h
is
s
tu
d
y
co
m
p
ar
es
L
i
n
ea
r
Dis
cr
i
m
in
a
n
t
An
al
y
s
is
(
L
D
A
)
alg
o
r
ith
m
a
n
d
Naïv
e
B
a
y
es
a
lg
o
r
ith
m
to
o
b
tain
th
e
b
es
t
clas
s
i
f
icatio
n
r
esu
lt
s
.
T
h
e
r
esu
lt
o
f
cla
s
s
i
f
ic
atio
n
o
f
L
D
A
al
g
o
r
ith
m
h
as
p
o
o
r
ac
cu
r
ac
y
o
n
7
class
es
w
h
er
ea
s
f
o
r
No
r
m
a
l
an
d
A
b
n
o
r
m
al
c
lass
clas
s
i
f
ic
atio
n
,
th
e
r
esu
lt
h
a
s
g
o
o
d
en
o
u
g
h
ac
cu
r
ac
y
,
an
d
th
er
e
i
s
d
i
f
f
ic
u
lt
y
f
o
r
ab
n
o
r
m
al
clas
s
if
icatio
n
w
it
h
lo
w
ac
c
u
r
ac
y
v
al
u
e.
T
h
e
lo
w
ac
cu
r
ac
y
o
f
t
h
e
ab
n
o
r
m
al
class
a
f
f
ec
t
s
t
h
e
class
if
ica
tio
n
in
to
7
class
es [
2
3
]
.
T
h
e
r
esear
ch
th
at
tr
ied
to
o
v
er
co
m
e
t
h
e
d
if
f
ic
u
ltie
s
o
f
s
in
g
le
P
ap
s
m
ea
r
i
m
a
g
e
cla
s
s
i
f
i
ca
tio
n
in
7
class
es
w
as
d
o
n
e
b
y
[
2
4
]
.
T
h
is
s
t
u
d
y
o
b
s
er
v
ed
a
n
u
m
b
er
o
f
class
es
th
at
h
a
s
d
if
f
er
en
t
a
m
o
u
n
t
s
o
f
d
ata,
ie,
t
h
e
d
ataset
h
a
s
a
clas
s
w
it
h
a
n
u
m
b
er
o
f
d
if
f
er
e
n
t
a
n
d
u
n
b
alan
ce
d
class
es.
An
o
t
h
er
co
n
d
itio
n
is
th
at
th
e
d
ata
h
a
s
f
ea
t
u
r
es
t
h
at
ar
e
s
u
s
p
ec
ted
to
b
e
ir
r
elev
an
t,
s
o
it
is
s
ti
ll
d
i
f
f
icu
lt
to
clas
s
i
f
y
esp
ec
iall
y
ab
n
o
r
m
al
c
lass
e
s
.
T
o
h
an
d
le
th
e
clas
s
i
m
b
alan
ce
,
th
i
s
s
t
u
d
y
u
s
ed
en
s
e
m
b
le
m
eth
o
d
(
B
ag
g
i
n
g
)
.
Fo
r
h
a
n
d
li
n
g
d
ata
th
at
H
D
A
f
ea
t
u
r
es
an
d
HD
A
n
o
co
n
tr
ib
u
tio
n
,
we
m
ad
e
f
ea
t
u
r
e
s
elec
tio
n
o
f
Gr
ee
d
y
Fo
r
w
ar
d
Selectio
n
.
F
u
r
th
er
m
o
r
e,
Naï
v
e
B
ay
e
s
w
as
u
s
ed
as
lear
n
i
n
g
alg
o
r
ith
m
s
.
Alth
o
u
g
h
t
h
i
s
m
et
h
o
d
ca
n
h
a
n
d
le
i
m
b
a
lan
ce
class
es,
b
u
t
t
h
e
class
i
f
icatio
n
o
f
7
class
e
s
h
a
s
n
o
t a
ch
ie
v
ed
th
e
m
ax
i
m
u
m
r
es
u
lts
[
2
4
]
.
W
e
h
av
e
i
m
p
le
m
en
ted
P
ap
s
m
ea
r
class
i
f
icatio
n
alg
o
r
it
h
m
s
b
y
u
s
i
n
g
NN
cla
s
s
i
f
icatio
n
al
g
o
r
ith
m
an
d
f
ea
t
u
r
e
s
elec
tio
n
b
y
u
s
in
g
G
A
.
T
h
e
b
est
m
o
d
el
o
f
th
e
c
l
ass
i
f
icatio
n
r
es
u
lt
b
ec
a
m
e
th
e
Hier
ar
ch
ical
HD
A
m
o
d
el,
a
n
e
w
clas
s
i
f
icatio
n
ap
p
r
o
ac
h
f
o
r
P
ap
Sm
ea
r
i
m
a
g
e.
T
h
e
co
m
p
ar
is
o
n
o
f
class
if
i
ca
ti
o
n
r
esu
lt
s
b
y
u
s
i
n
g
NN
alg
o
r
ith
m
an
d
f
ea
t
u
r
e
o
p
ti
m
izatio
n
b
y
u
s
in
g
G
A
to
d
eter
m
in
e
t
h
e
in
cr
ea
s
e
o
f
ac
c
u
r
ac
y
w
a
s
co
n
d
u
cted
.
P
ap
s
m
ea
r
i
m
a
g
e
clas
s
i
f
icati
o
n
in
to
7
class
es
u
s
i
n
g
H
D
A
m
et
h
o
d
h
as
g
o
o
d
class
i
f
i
ca
tio
n
v
al
u
e
w
h
i
le
class
i
f
icatio
n
u
s
i
n
g
NN
al
g
o
r
i
th
m
a
n
d
f
ea
t
u
r
e
o
p
ti
m
iza
tio
n
u
s
i
n
g
G
A
h
a
v
e
lo
w
er
v
a
lu
e
co
m
p
ar
ed
to
HD
A
alg
o
r
ith
m
[
1
9
]
.
Ho
w
e
v
er
,
th
e
p
r
esen
t
s
t
u
d
y
is
a
n
i
m
p
r
o
v
e
m
en
t
o
f
th
e
r
esear
c
h
b
y
g
i
v
i
n
g
s
p
ec
ial
atten
tio
n
t
o
th
e
m
o
r
e
p
r
o
p
o
r
tio
n
al
in
itial
d
ata
-
s
h
ar
i
n
g
p
r
o
ce
s
s
b
y
u
s
i
n
g
s
p
lit
v
al
id
atio
n
m
et
h
o
d
t
h
at
i
m
p
r
o
v
es
t
h
e
p
r
o
ce
s
s
o
f
p
r
ev
io
u
s
r
esear
c
h
m
e
th
o
d
s
.
T
h
is
r
esu
lted
in
ac
c
u
r
ac
y
v
al
u
es
f
o
r
b
o
th
n
o
r
m
al
a
n
d
ab
n
o
r
m
al
clas
s
if
icatio
n
,
an
d
th
e
clas
s
i
f
icatio
n
o
f
7
clas
s
es e
x
p
er
ie
n
ce
d
a
s
ig
n
i
f
ica
n
t i
n
cr
ea
s
e.
3.
RE
S
E
ARCH
M
E
T
H
O
D
3
.
1
.
Da
t
a
Co
llect
io
n
A
t
t
h
i
s
s
ta
g
e,
w
e
d
eter
m
i
n
ed
th
e
d
ata
to
b
e
p
r
o
ce
s
s
ed
,
s
ea
r
ch
ed
f
o
r
av
ailab
le
d
ata,
o
b
tain
ed
th
e
ad
d
itio
n
al
d
ata
r
eq
u
ir
ed
,
an
d
in
teg
r
ated
all
d
ata
in
to
d
ata
s
ets
in
cl
u
d
in
g
v
ar
iab
les
r
eq
u
ir
ed
in
th
e
p
r
o
ce
s
s
.
T
h
e
d
ata
u
s
ed
f
o
r
tr
ain
in
g
a
n
d
tes
tin
g
is
s
ec
o
n
d
ar
y
d
ata
clas
s
i
f
i
ed
ca
r
ef
u
ll
y
b
y
c
y
to
-
tec
h
n
icia
n
s
a
n
d
d
o
cto
r
s
.
T
o
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
6
,
Decem
b
er
201
8
:
5
4
1
5
-
5
4
2
4
5418
i
m
p
r
o
v
e
t
h
e
clas
s
i
f
icatio
n
o
f
P
ap
s
m
ea
r
ce
ll
i
m
a
g
es
i
n
t
h
is
e
x
p
er
i
m
e
n
t
w
e
u
s
ed
Her
lev
9
1
7
d
ata
[
4
]
.
I
n
T
ab
le
1
,
it c
an
b
e
s
ee
n
th
at
th
e
2
0
f
ea
tu
r
e
s
f
o
u
n
d
in
t
h
e
d
a
taset f
ea
t
u
r
e
w
as o
p
ti
m
ized
b
y
u
s
i
n
g
G
A
.
T
ab
le
1
.
T
h
e
F
ea
tu
r
e
o
f
Her
lev
Data
s
et
[
4
]
N
a
me
O
f
F
e
a
t
u
r
e
N
a
me
O
f
F
e
a
t
u
r
e
N
a
me
O
f
F
e
a
t
u
r
e
N
a
me
O
f
F
e
a
t
u
r
e
N
u
c
l
e
u
s A
r
e
a
o
r
K
e
r
n
e
_
A
N
u
c
l
e
u
s Sh
o
r
t
e
st
D
i
a
me
t
e
r
o
r
K
e
r
n
e
S
h
o
r
t
C
y
t
o
p
l
a
sm L
o
n
g
e
st
D
i
a
me
t
e
r
o
r
C
y
t
o
L
o
n
g
N
u
c
l
e
u
s R
e
a
l
t
i
v
e
P
o
si
t
i
o
n
o
r
K
e
r
n
e
P
o
s
C
y
t
o
p
l
a
sm A
r
e
a
o
r
C
y
t
o
_
A
N
u
c
l
e
u
s L
o
n
g
e
st
D
i
a
me
t
e
r
o
r
K
e
r
n
e
L
o
n
g
C
y
t
o
p
l
a
sm El
o
n
g
a
t
i
o
n
o
r
C
y
t
o
El
o
n
g
N
u
c
l
e
u
s
M
a
x
i
m
u
m o
r
K
e
r
n
e
M
a
x
N
/
C
r
a
t
i
o
o
r
K
/
C
N
u
c
l
e
u
s E
l
o
n
g
a
t
i
o
n
o
r
K
e
r
n
e
El
o
n
g
C
y
t
o
p
l
a
sm R
o
u
n
d
n
e
ss o
r
C
y
t
o
R
u
n
d
N
u
c
l
e
u
s
M
i
n
i
m
u
m o
r
K
e
r
n
e
M
i
n
N
u
c
l
e
u
s B
r
i
g
h
t
n
e
ss o
r
K
e
r
n
e
_
Y
c
o
l
N
u
c
l
e
u
s R
o
u
n
d
n
e
ss o
r
K
e
r
n
e
R
u
n
d
N
u
c
l
e
u
s Pe
r
i
me
t
e
r
o
r
K
e
r
n
e
P
e
r
i
C
y
t
o
p
l
a
sm M
a
x
i
mu
m o
r
C
y
t
o
M
a
x
C
y
t
o
p
l
a
sm B
r
i
g
h
t
n
e
ss
o
r
C
y
t
o
_
Y
c
o
l
C
y
t
o
p
l
a
sm Sh
o
r
t
e
st
D
i
a
me
t
e
r
o
r
C
y
t
o
S
h
o
r
t
C
y
t
o
p
l
a
sm Pe
r
i
me
t
e
r
o
r
C
y
t
o
P
e
r
i
C
y
t
o
p
l
a
sm M
i
n
i
m
u
m o
r
C
y
t
o
M
i
n
3
.
2
.
P
r
o
po
s
ed
M
e
t
ho
d
A
t
t
h
i
s
s
ta
g
e
t
h
e
d
ata
w
a
s
a
n
al
y
ze
d
an
d
g
r
o
u
p
ed
i
n
to
v
ar
iab
les
th
at
ar
e
r
elate
d
to
ea
ch
o
t
h
er
.
Af
ter
th
e
d
ata
w
as a
n
al
y
ze
d
,
t
h
e
m
o
d
els ac
co
r
d
in
g
to
th
e
d
ata
t
y
p
e
w
er
e
ap
p
lied
.
Data
s
h
ar
i
n
g
i
n
t
o
tr
ain
in
g
d
ata
an
d
test
d
ata
w
as
also
r
eq
u
ir
ed
f
o
r
m
o
d
elin
g
.
T
h
is
s
tu
d
y
w
i
ll
s
elec
t
an
d
ap
p
ly
ap
p
r
o
p
r
iate
tech
n
iq
u
es
f
o
r
P
ap
s
m
ea
r
i
m
a
g
e
clas
s
i
f
icatio
n
.
T
h
e
f
ir
s
t
s
ta
g
e
i
n
t
h
is
s
t
u
d
y
w
a
s
t
o
d
iv
id
e
th
e
P
ap
s
m
ea
r
ce
l
l d
ataset
i
n
to
t
w
o
p
ar
ts
ie,
tr
an
in
g
d
ata
an
d
test
in
g
d
ata.
T
h
e
n
ex
t
s
tep
w
a
s
to
p
er
f
o
r
m
th
e
b
est
f
ea
t
u
r
e
s
elec
tio
n
in
th
e
P
ap
s
m
ea
r
i
m
a
g
e
d
ataset
b
y
u
s
i
n
g
GA
,
an
d
th
en
t
h
e
s
elec
ted
f
ea
t
u
r
e
w
a
s
class
i
f
ied
b
y
u
s
in
g
NN
alg
o
r
ith
m
.
T
h
e
b
est
m
o
d
el
f
r
o
m
t
h
e
clas
s
if
icatio
n
r
esu
lt
w
a
s
u
s
ed
as
H
D
A
m
o
d
el,
s
o
a
n
e
w
cla
s
s
i
f
icatio
n
m
e
th
o
d
ap
p
r
o
ac
h
w
a
s
p
r
o
p
o
s
ed
f
o
r
P
a
p
s
m
e
ar
i
m
a
g
e.
T
h
e
r
esu
lts
o
f
t
h
e
m
o
d
el
c
lass
i
f
icatio
n
w
ill
b
e
m
ea
s
u
r
ed
w
it
h
a
n
ac
cu
r
ac
y
v
alu
e.
T
h
e
r
esear
c
h
d
esi
g
n
ca
n
b
e
s
ee
n
i
n
F
i
g
u
r
e
1
.
I
n
i
t
i
a
l
P
o
p
u
l
a
t
i
o
n
G
e
n
e
t
i
c
A
l
g
o
r
i
t
h
m
F
e
a
t
u
r
e
S
e
l
e
c
t
i
o
n
F
i
t
n
e
s
s
E
v
a
l
u
a
t
i
o
n
I
n
d
i
v
i
d
u
a
l
S
e
l
e
c
t
i
o
n
C
r
o
s
s
o
v
e
r
a
n
d
M
u
t
a
t
i
o
n
I
n
i
t
i
a
l
P
o
p
u
l
a
t
i
o
n
E
v
a
l
u
a
t
i
o
n
a
n
d
V
a
l
i
d
a
t
i
o
n
P
r
o
b
l
e
m
I
d
e
n
t
i
f
i
c
a
t
i
o
n
T
r
a
i
n
i
n
g
D
a
t
a
70
%
T
e
s
t
i
n
g
D
a
t
a
30
%
I
n
i
t
i
a
l
D
a
t
a
P
r
o
c
e
s
s
i
n
g
2
0
F
e
a
t
u
r
e
H
a
r
l
e
v
D
a
t
a
s
e
t
P
a
r
a
m
e
t
e
r
O
p
t
i
m
i
z
e
d
N
e
u
r
a
l
N
e
t
w
o
r
k
E
x
p
e
r
i
m
e
n
t
s
a
n
d
M
o
d
e
l
T
e
s
t
i
n
g
C
l
a
s
s
i
f
i
c
a
t
i
o
n
R
e
s
u
l
t
s
N
o
r
m
a
l
A
b
n
o
r
m
a
l
C
l
a
s
s
1
C
l
a
s
s
2
C
l
a
s
s
3
C
l
a
s
s
4
C
l
a
s
s
5
a
n
d
6
C
l
a
s
s
7
C
l
a
s
s
5
C
l
a
s
s
6
H
i
e
r
a
r
c
h
i
c
a
l
D
e
c
i
s
i
o
n
A
p
p
r
o
a
c
h
(
H
D
A
)
Fig
u
r
e
1
.
R
esear
c
h
De
s
ig
n
a)
I
n
itial Da
ta
P
r
o
ce
s
s
in
g
I
n
th
i
s
s
ta
g
e,
d
ata
s
elec
tio
n
was
co
n
d
u
cted
.
T
h
e
d
ata
w
a
s
c
lean
ed
an
d
tr
an
s
f
o
r
m
ed
in
to
t
h
e
d
esire
d
s
h
ap
e
s
o
t
h
at
it
ca
n
b
e
d
o
n
e
in
p
r
ep
ar
atio
n
o
f
m
o
d
el
m
a
k
in
g
.
A
t
th
i
s
s
tag
e,
e
x
p
lo
r
atio
n
o
f
t
h
e
d
ataset
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
I
mp
r
o
vin
g
Hiera
r
ch
ica
l D
ec
is
io
n
A
p
p
r
o
a
ch
fo
r
S
i
n
g
le
I
ma
g
e
C
la
s
s
if
ica
tio
n
o
f P
a
p
S
mea
r
(
D
w
iz
a
R
ia
n
a
)
5419
p
r
o
v
id
ed
is
r
eq
u
ir
ed
.
First
o
f
all,
it
i
s
k
n
o
w
n
t
h
at
t
h
e
m
ai
n
g
o
al
to
b
e
ac
h
ie
v
ed
is
to
k
n
o
w
t
h
e
b
es
t
class
i
f
icatio
n
r
es
u
lt
o
f
P
ap
s
m
ea
r
ce
ll
i
m
a
g
e.
T
h
is
s
t
u
d
y
u
s
e
d
Her
lev
d
ata
s
et
w
i
th
t
h
e
r
ec
o
r
d
s
o
f
9
1
7
.
T
o
test
th
e
m
o
d
el
d
ev
elo
p
ed
,
th
e
d
ata
w
o
u
ld
b
e
d
iv
id
ed
in
to
t
w
o
p
ar
ts
,
n
a
m
el
y
tr
ain
in
g
d
ata
a
n
d
d
ata
tes
ti
n
g
.
T
h
e
d
ata
tr
ain
in
g
w
as
u
s
ed
f
o
r
m
o
d
el
d
ev
elo
p
m
e
n
t
w
h
ile
d
ata
te
s
tin
g
w
as
u
s
ed
f
o
r
m
o
d
el
test
i
n
g
.
I
t
is
k
n
o
w
n
t
h
at
th
e
a
m
o
u
n
t
o
f
d
ata
is
9
1
7
w
it
h
a
d
i
v
is
io
n
o
f
7
0
%
(
6
4
2
)
u
s
ed
f
o
r
tr
ain
i
n
g
d
ata
a
n
d
3
0
%
(
2
7
5
)
u
s
ed
f
o
r
d
ata
test
i
n
g
.
T
h
e
n
ex
t
s
ta
g
e
w
as
to
s
elec
t
d
ata
th
at
w
u
o
ld
b
e
u
s
e
d
as
tr
ain
in
g
d
ata
an
d
d
ata
test
in
g
b
y
u
s
i
n
g
s
p
lit
v
alid
atio
n
.
F
u
r
th
er
m
o
r
e,
th
e
f
ea
tu
r
e
s
elec
tio
n
m
et
h
o
d
w
as
p
er
f
o
r
m
ed
i
n
th
i
s
r
esear
ch
wh
ich
i
s
G
A
m
et
h
o
d
.
GA
cr
ea
te
a
p
o
p
u
lat
io
n
co
m
p
o
s
ed
o
f
m
a
n
y
in
d
i
v
id
u
al
s
t
h
at
ev
o
lv
e
ac
co
r
d
i
n
g
to
ce
r
tain
s
elec
tio
n
r
u
le
s
t
h
at
h
av
e
o
p
ti
m
izatio
n
d
eter
m
in
a
ti
o
n
an
d
v
al
u
e.
b)
E
x
p
er
i
m
e
n
ts
a
n
d
Mo
d
el
T
esti
n
g
A
t
t
h
is
s
tag
e
t
h
e
p
r
o
p
o
s
ed
m
o
d
el
w
ill
b
e
test
ed
to
s
ee
th
e
r
esu
lts
o
f
a
r
u
le
t
h
at
w
ill
b
e
u
tili
ze
d
in
d
ec
is
io
n
m
a
k
i
n
g
.
T
h
is
r
esear
ch
w
ill
co
n
d
u
ct
e
x
p
er
i
m
e
n
ts
o
n
th
e
clas
s
i
f
icatio
n
o
f
d
ata
m
i
n
in
g
u
s
i
n
g
NN
alg
o
r
ith
m
.
T
h
e
m
o
d
eli
n
g
w
ill
b
e
d
o
n
e
b
y
u
s
in
g
R
ap
id
m
i
n
er
s
o
f
t
w
ar
e.
T
h
e
m
o
d
els
th
a
t h
a
v
e
b
ee
n
o
b
tain
ed
ar
e
tr
an
s
f
o
r
m
ed
i
n
to
t
h
e
p
r
o
g
r
a
m
m
i
n
g
lan
g
u
ag
e
o
f
Vi
s
u
al
B
asi
c
.
Net
2
0
1
7
,
an
d
m
o
d
elin
g
tr
a
n
s
lat
i
o
n
o
f
r
esear
ch
d
esig
n
th
at
h
a
s
b
ee
n
d
o
n
e
b
e
f
o
r
e
ar
e
p
e
r
f
o
r
m
ed
b
ec
au
s
e
t
h
e
m
o
d
el
o
f
HD
A
ca
n
n
o
t
b
e
d
o
n
e
o
n
s
o
f
t
w
ar
e
R
ap
id
m
i
n
er
p
r
o
g
r
a
m
m
in
g
.
c)
E
v
alu
a
tio
n
a
n
d
Valid
atio
n
A
t
t
h
is
s
ta
g
e
a
n
e
v
al
u
atio
n
o
f
t
h
e
m
o
d
el
d
eter
m
in
ed
to
f
i
n
d
o
u
t
th
e
le
v
el
o
f
m
o
d
el
ac
cu
r
ac
y
w
as
d
o
n
e.
T
h
e
ev
alu
atio
n
w
a
s
p
er
f
o
r
m
ed
b
y
u
s
in
g
t
h
e
co
n
f
u
s
io
n
m
atr
i
x
tab
le
to
d
eter
m
i
n
e
th
e
al
g
o
r
ith
m
p
er
f
o
r
m
a
n
ce
m
ea
s
u
r
e
m
en
t
o
n
th
e
cla
s
s
i
f
icatio
n
al
g
o
r
ith
m
m
o
d
el.
T
h
e
m
ea
s
u
r
ed
p
er
f
o
r
m
an
ce
is
A
cc
u
r
ac
y
.
T
h
e
v
alid
atio
n
p
er
f
o
r
m
ed
u
s
e
d
th
e
d
ata
t
h
at
h
ad
b
ee
n
d
iv
i
d
ed
m
a
n
u
all
y
i
n
to
test
i
n
g
d
at
a
an
d
tr
ain
i
n
g
d
ata.
T
h
e
m
o
d
el
p
er
f
o
r
m
a
n
ce
w
ill
b
e
co
m
p
ar
ed
w
ith
NN
al
g
o
r
ith
m
b
y
p
er
f
o
r
m
i
n
g
f
ea
t
u
r
e
o
p
ti
m
izatio
n
b
y
u
s
i
n
g
GA
an
d
co
m
p
ar
ed
w
it
h
Ne
u
r
al
Net
w
r
o
k
alg
o
r
it
h
m
w
it
h
o
u
t
d
o
in
g
o
p
ti
m
iza
tio
n
.
A
cc
u
r
ac
y
w
a
s
u
s
ed
to
co
m
p
a
r
e
t
h
e
r
esu
l
ts
s
o
t
h
at
th
e
r
esu
lts
o
b
tai
n
ed
ar
e
m
o
r
e
ac
c
u
r
ate.
4.
RE
SU
L
T
S
A
ND
AN
AL
Y
SI
S
I
n
th
i
s
r
esear
ch
,
w
e
w
i
ll
p
er
f
o
r
m
f
ea
tu
r
e
s
elec
tio
n
e
x
p
er
i
m
en
ts
b
y
u
s
i
n
g
G
A
a
n
d
P
ap
s
m
ea
r
class
i
f
icatio
n
b
y
u
s
i
n
g
NN
al
g
o
r
ith
m
.
T
h
e
e
x
p
er
i
m
e
n
ts
w
e
r
e
co
n
d
u
cted
b
y
u
s
in
g
Her
lev
d
ataset
w
h
er
e
t
h
e
in
itial
d
ata
p
r
o
ce
s
s
i
n
g
h
ad
b
ee
n
d
o
n
e
w
it
h
t
h
e
d
is
tr
ib
u
tio
n
o
f
tr
ain
i
n
g
a
n
d
test
in
g
d
ata.
I
n
t
h
is
s
ec
tio
n
w
e
w
ill
s
h
o
w
t
h
e
e
x
p
er
i
m
e
n
tal
r
es
u
lt
s
b
y
u
s
i
n
g
th
e
NN
al
g
o
r
ith
m
a
n
d
f
ea
t
u
r
e
s
elec
t
io
n
u
s
i
n
g
G
A
b
y
u
s
i
n
g
2
0
attr
ib
u
tes s
h
o
w
n
i
n
T
ab
le
2
in
th
e
Her
lev
d
atase
t.
I
n
th
e
ea
r
l
y
s
ta
g
es
o
f
t
h
is
r
e
s
ea
r
ch
,
th
e
p
r
o
ce
s
s
o
f
s
ep
ar
atio
n
o
f
tr
a
n
i
n
g
d
ata
a
n
d
te
s
tin
g
d
ata
w
a
s
co
n
d
u
cted
,
an
d
th
e
f
ea
tu
r
e
s
el
ec
tio
n
u
s
in
g
Ge
n
etic
A
l
g
o
r
tih
m
w
as
t
h
en
p
er
f
o
r
m
ed
.
T
h
e
b
est
attr
ib
u
te
w
ill
b
e
u
s
ed
a
s
t
h
e
P
a
p
s
m
ea
r
clas
s
i
f
i
ca
tio
n
m
o
d
el
u
s
i
n
g
NN
m
et
h
o
d
.
T
h
e
class
i
f
icatio
n
p
r
o
ce
s
s
u
s
in
g
N
N
al
g
o
r
ith
m
w
a
s
d
o
n
e
b
y
o
p
ti
m
iz
in
g
th
e
b
est
v
alu
e
o
f
NN
al
g
o
r
ith
m
p
ar
am
eter
w
it
h
th
e
v
alu
e
o
f
L
ea
r
n
i
n
g
R
ate
an
d
Mo
m
en
t
u
m
i
n
to
2
m
o
d
els.
T
h
e
f
ir
s
t
m
o
d
el
u
s
ed
th
e
lear
n
i
n
g
r
ate
(
lr
)
v
a
lu
e
o
f
0
.
3
an
d
m
o
m
e
n
t
u
m
(
m
)
o
f
0
.
2
w
h
ile
t
h
e
s
ec
o
n
d
m
o
d
el
u
s
es
th
e
lear
n
i
n
g
r
ate
(
lr
)
v
alu
e
o
f
0
.
5
an
d
m
o
m
e
n
tu
m
(
m
)
o
f
0
.
5
.
Fu
r
th
er
m
o
r
e,
th
e
h
ig
h
e
s
t
ac
c
u
r
ac
y
v
al
u
e
an
a
l
y
s
is
w
as
u
s
ed
f
o
r
t
h
e
HD
A
m
o
d
el.
Fro
m
t
h
e
r
es
u
lts
,
it
is
k
n
o
w
n
t
h
at
t
h
e
v
al
u
e
o
f
l
ea
r
n
in
g
r
ate
an
d
m
o
m
e
n
t
u
m
g
r
ea
tl
y
af
f
ec
t
s
th
e
ac
c
u
r
ac
y
o
f
t
h
e
class
if
icatio
n
.
T
ab
le
2
.
C
lass
if
icatio
n
R
es
u
lt
o
f
NN
A
lg
o
r
it
h
m
an
d
G
A
No
Ty
p
e
O
f
C
l
a
ssi
f
i
c
a
t
i
o
n
NN
G
A
+
NN
(
0
.
3
lr
a
n
d
0
.
2
m)
G
A
+
NN
(
0
.
5
l
r
a
n
d
0
.
5
m)
1
7
c
l
a
sse
s
6
4
.
0
0
%
7
0
.
1
8
%
6
6
.
9
1
%
2
N
o
r
mal
&
A
b
n
o
r
mal
9
3
.
1
2
%
9
6
.
0
1
%
9
7
.
1
0
%
3
N
o
r
mal
1
,
2
,
3
9
7
.
2
2
%
9
8
.
6
1
%
1
0
0
%
4
A
b
n
o
r
mal
4
,
5
&
6
,
7
5
7
.
1
4
%
7
4
.
8
8
%
7
3
.
4
0
%
5
A
b
n
o
r
mal
5
&
6
7
4
.
7
6
%
8
5
.
4
4
%
8
4
.
4
7
%
I
n
T
ab
le
2
,
th
e
class
i
f
icatio
n
co
m
p
ar
is
o
n
r
es
u
lt
s
h
o
w
s
t
h
at
t
h
e
class
i
f
icatio
n
w
it
h
7
class
es
u
s
i
n
g
NN
alg
o
r
ith
m
w
it
h
t
h
e
ac
c
u
r
ac
y
v
alu
e
o
f
6
4
.
0
0
%
af
ter
u
s
i
n
g
f
e
atu
r
e
s
elec
tio
n
b
y
u
s
i
n
g
G
A
a
n
d
clas
s
i
f
icatio
n
b
y
u
s
i
n
g
NN
alg
o
r
i
t
h
m
w
it
h
th
e
lear
n
i
n
g
r
ate
o
f
0
.
3
an
d
m
o
m
en
tu
m
o
f
0
.
2
ex
p
er
ie
n
ce
s
t
h
e
i
m
p
r
o
v
e
m
en
t
o
f
ac
cu
r
ac
y
w
it
h
a
v
al
u
e
o
f
7
0
.
1
8
%
an
d
w
i
t
h
t
h
e
v
alu
e
o
f
l
ea
r
n
in
g
r
ate
o
f
0
.
5
an
d
m
o
m
en
tu
m
o
f
0
.
5
b
u
t
p
r
o
d
u
ce
s
an
ac
c
u
r
ac
y
v
al
u
e
o
f
6
6
.
9
1
%
w
h
er
e
t
h
e
ac
cu
r
ac
y
r
es
u
lts
s
ti
ll
lo
o
k
le
s
s
.
T
h
u
s
,
t
h
e
p
r
o
ce
s
s
o
f
class
i
f
icatio
n
u
s
i
n
g
t
h
e
HD
A
m
o
d
el
b
y
ta
k
i
n
g
th
e
b
est
m
o
d
el
in
ea
c
h
cla
s
s
i
f
icat
io
n
w
a
s
d
o
n
e.
F
r
o
m
t
h
e
b
est
class
i
f
icatio
n
r
es
u
lt o
f
ea
ch
c
l
ass
,
t
h
e
b
est
m
o
d
el
w
as ta
k
e
n
f
o
r
th
e
f
o
r
m
at
io
n
o
f
HD
A
m
o
d
el.
Fro
m
th
e
m
o
d
el
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
6
,
Decem
b
er
201
8
:
5
4
1
5
-
5
4
2
4
5420
w
h
o
s
e
h
i
g
h
e
s
t
ac
cu
r
ac
y
v
alu
e
h
ad
b
ee
n
k
n
o
w
n
,
th
e
b
es
t
f
ea
t
u
r
e
s
ep
ar
atio
n
u
s
i
n
g
G
A
w
a
s
p
er
f
o
r
m
ed
w
i
th
th
e
d
is
tr
ib
u
tio
n
o
f
d
ata
s
h
o
w
n
in
T
ab
le
3
.
T
h
is
s
ta
g
e
w
ill
p
er
f
o
r
m
a
n
ev
al
u
atio
n
th
at
ai
m
s
to
d
e
ter
m
i
n
e
t
h
e
le
v
el
o
f
ac
c
u
r
a
c
y
o
f
th
e
class
i
f
icatio
n
te
s
ti
n
g
r
es
u
lts
u
s
in
g
NN
al
g
o
r
it
h
m
an
d
f
ea
tu
r
e
s
elec
tio
n
u
s
i
n
g
G
A
b
y
co
u
n
tin
g
t
h
e
a
m
o
u
n
t
o
f
test
i
n
g
d
ata
t
h
at
ca
n
b
e
c
lass
i
f
ied
co
r
r
ec
tl
y
.
T
h
e
test
w
as
d
o
n
e
b
y
u
s
in
g
r
ap
id
m
in
er
s
o
f
t
w
ar
e
to
g
et
t
h
e
b
est
m
o
d
el
a
n
d
g
et
t
h
e
r
esu
lt
o
f
a
cc
u
r
ac
y
v
alu
e.
Af
ter
o
b
tain
i
n
g
th
e
b
est
m
o
d
el
o
f
th
e
r
es
u
lt
s
o
b
tain
ed
,
th
en
t
h
e
p
r
o
ce
s
s
o
f
clas
s
i
f
icatio
n
u
s
i
n
g
HD
A
w
as
co
n
d
u
cted
to
o
b
tain
cla
s
s
i
f
icatio
n
r
es
u
lt
s
w
it
h
7
class
e
s
b
y
u
s
in
g
Vis
u
a
l
S
tu
d
io
p
r
o
g
r
a
m
2
0
1
7
.
B
ased
o
n
t
h
e
r
esear
c
h
t
h
at
h
a
s
b
ee
n
clas
s
i
f
ied
w
it
h
7
clas
s
e
s
,
it
h
a
s
t
h
e
h
i
g
h
e
s
t
ac
cu
r
ac
y
o
f
7
0
.
1
8
%
in
w
h
ic
h
th
e
ac
c
u
r
ac
y
p
r
o
d
u
ce
d
h
as
n
o
t
b
ee
n
o
p
ti
m
al,
s
o
t
h
e
n
w
e
p
r
o
p
o
s
ed
class
if
icatio
n
p
r
o
ce
s
s
u
s
i
n
g
HD
A
m
o
d
el.
I
t
w
as
p
er
f
o
r
m
ed
b
y
s
ep
ar
ati
n
g
th
e
clas
s
i
f
icatio
n
m
o
d
el
in
t
o
s
o
m
e
o
f
t
h
e
b
est
m
o
d
el
s
in
cl
u
d
in
g
:
No
r
m
al
an
d
A
b
n
o
r
m
al
C
la
s
s
i
f
icat
io
n
w
it
h
th
e
ac
cu
r
ac
y
o
f
9
7
.
1
0
%,
No
r
m
al
C
las
s
i
f
icatio
n
1
,
2
,
3
w
ith
t
h
e
ac
cu
r
ac
y
o
f
1
0
0
%,
A
b
n
o
r
m
al
C
las
s
i
f
icatio
n
4
,
5
+6
,
7
w
i
t
h
th
e
ac
cu
r
ac
y
o
f
7
4
,
8
8
%,
b
y
r
ef
er
r
in
g
to
T
ab
le
3
.
C
lass
5
+6
w
as
m
ad
e
in
to
o
n
e
b
ec
au
s
e
t
h
er
e
w
er
e
cla
s
s
i
f
icat
io
n
d
i
f
f
icu
l
ti
es
f
o
r
t
h
e
Mo
d
er
ate
D
y
s
p
la
s
ia
class
a
n
d
Sev
er
e
D
y
s
p
las
ia
[
1
4
]
.
T
h
e
f
in
al
s
tep
w
a
s
to
class
i
f
y
cla
s
s
5
an
d
6
w
it
h
t
h
e
ac
cu
r
ac
y
o
f
85
.
4
4
3
%.
T
h
e
HDA
m
o
d
el
h
i
g
h
l
y
d
ep
e
n
d
s
o
n
th
e
m
o
d
el
th
at
h
a
s
b
ee
n
d
er
iv
ed
f
r
o
m
th
e
cla
s
s
i
f
icat
i
o
n
o
f
ea
ch
class
to
b
e
th
e
r
ef
er
en
ce
m
o
d
e
l
f
o
r
m
ak
i
n
g
th
e
HD
A
al
g
o
r
it
h
m
.
T
h
er
ef
o
r
e,
ea
ch
o
f
th
e
b
e
s
t
f
ea
t
u
r
es
t
h
at
h
a
v
e
b
ee
n
s
elec
ted
b
y
u
s
i
n
g
G
A
is
p
r
esen
ted
in
T
ab
le
3
a
s
a
r
ep
r
esen
tat
io
n
o
f
HD
A
m
o
d
el
f
o
r
m
atio
n
.
E
ac
h
clas
s
h
as a
d
if
f
er
e
n
t H
id
d
en
la
y
er
d
ep
en
d
in
g
o
n
th
e
n
u
m
b
er
o
f
f
ea
tu
r
es selec
ted
an
d
th
e
m
o
s
t r
el
ev
an
t
f
ea
t
u
r
e
to
th
e
ac
cu
r
ac
y
v
alu
e.
HD
A
al
g
o
r
ith
m
m
o
d
el
n
o
t
o
n
l
y
a
f
f
ec
ts
t
h
e
ac
c
u
r
ac
y
o
f
ea
c
h
clas
s
b
u
t
also
a
f
f
ec
ts
t
h
e
w
e
ig
h
t
v
a
lu
e
o
f
ea
ch
n
o
d
e
w
h
er
e
n
o
d
es
ar
e
o
b
tain
ed
f
r
o
m
ea
c
h
attr
ib
u
te
th
at
h
as
b
ee
n
s
elec
ted
.
E
ac
h
w
ei
g
h
t
h
a
s
d
if
f
er
e
n
t
v
alu
e
s
.
T
ab
le
3
.
Selecte
d
A
ttrib
u
tes
U
s
in
g
G
A
No
N
o
r
mal
a
n
d
a
b
n
o
r
mal
C
l
a
ssi
f
i
c
a
t
i
o
n
C
l
a
ssi
f
i
c
a
t
i
o
n
1
,
2
,
3
C
l
a
ssi
f
i
c
a
t
i
o
n
o
f
C
l
a
ss
4
,
5
,
6
,
7
C
l
a
ss
5
,
6
1
K
e
r
n
e
_
Y
c
o
l
K
e
r
n
e
_
A
C
y
t
o
_
A
K
e
r
n
e
_
A
2
C
y
t
o
_
Y
c
o
l
C
y
t
o
_
A
K
/
C
K
e
r
n
e
_
Y
c
o
l
3
K
e
r
n
e
S
h
o
r
t
K
e
r
n
e
_
Y
c
o
l
C
y
t
o
_
Y
c
o
l
C
y
t
o
_
Y
c
o
l
4
K
e
r
n
e
L
o
n
g
C
y
t
o
_
Y
c
o
l
K
e
r
n
e
L
o
n
g
K
e
r
n
e
S
h
o
r
t
5
C
y
t
o
L
o
n
g
C
y
t
o
El
o
n
g
K
e
r
n
e
M
a
x
K
e
r
n
e
L
o
n
g
6
C
y
t
o
R
u
n
d
C
y
t
o
R
u
n
d
K
e
r
n
e
M
i
n
C
y
t
o
S
h
o
r
t
7
C
y
t
o
P
e
r
i
C
y
t
o
P
e
r
i
C
y
t
o
M
a
x
C
y
t
o
L
o
n
g
8
K
e
r
n
e
P
o
s
K
e
r
n
e
P
o
s
C
y
t
o
R
u
n
d
9
K
e
r
n
e
M
a
x
C
y
t
o
M
i
n
K
e
r
n
e
P
e
r
i
10
K
e
r
n
e
M
i
n
C
y
t
o
P
e
r
i
11
K
e
r
n
e
M
a
x
12
C
y
t
o
M
a
x
13
C
y
t
o
M
i
n
4
.
1
.
Appl
ica
t
io
n De
v
elo
p
m
e
nt
o
f
H
iera
rc
hy
M
o
del
Fro
m
t
h
e
r
es
u
lt
s
o
b
tai
n
ed
,
t
h
en
t
h
e
b
est
m
o
d
el
w
a
s
i
m
p
le
m
en
ted
in
Vi
s
u
al
St
u
d
io
.
Net
2
0
1
7
ap
p
licatio
n
f
o
r
th
e
clas
s
if
ica
tio
n
o
f
7
cla
s
s
es.
T
h
e
m
o
d
elin
g
s
ta
g
e
u
s
ed
t
h
e
Vi
s
u
al
Stu
d
io
.
Net
2
0
1
7
ap
p
licatio
n
w
ith
in
ter
f
ac
e
d
is
p
lay
in
Fi
g
u
r
e
2
(
a)
u
s
i
n
g
ea
ch
attr
ib
u
te
in
p
u
t
an
d
2
(
b
)
in
ter
f
ac
e
v
ie
w
s
f
o
r
class
i
f
icatio
n
u
s
in
g
d
ataset
s
with
m
u
l
tip
le
in
p
u
ts
.
T
h
e
n
ex
t
s
tep
w
a
s
th
e
clas
s
i
f
ic
atio
n
m
o
d
elin
g
i
m
p
le
m
e
n
tatio
n
o
f
7
class
es
w
it
h
th
e
f
o
llo
w
i
n
g
s
tag
e
s
:
n
o
r
m
al
a
n
d
ab
n
o
r
m
al
clas
s
if
ic
atio
n
m
o
d
el,
n
o
r
m
al
cla
s
s
i
f
ica
tio
n
m
o
d
el
1
,
2
,
3
,
ab
n
o
r
m
al
c
lass
i
f
icatio
n
m
o
d
el
4
,
5
an
d
6
,
7
,
an
d
class
if
icati
o
n
o
f
clas
s
5
an
d
6
w
i
th
t
h
e
f
o
llo
w
i
n
g
e
x
p
lan
a
tio
n
:
A
t
th
i
s
s
tag
e,
t
h
e
m
o
d
eli
n
g
f
o
r
th
e
n
o
r
m
al
a
n
d
ab
n
o
r
m
al
class
i
f
icatio
n
w
a
s
p
er
f
o
r
m
ed
b
y
u
s
i
n
g
t
h
e
p
r
o
ce
d
u
r
e
d
escr
i
b
ed
in
t
h
e
f
o
llo
w
i
n
g
s
tag
e
s
o
f
t
h
e
p
r
o
g
r
a
m
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
I
mp
r
o
vin
g
Hiera
r
ch
ica
l D
ec
is
io
n
A
p
p
r
o
a
ch
fo
r
S
i
n
g
le
I
ma
g
e
C
la
s
s
if
ica
tio
n
o
f P
a
p
S
mea
r
(
D
w
iz
a
R
ia
n
a
)
5421
(
a)
(
b
)
Fig
u
r
e
2
.
A
p
p
licatio
n
i
n
ter
f
ac
e
o
f
th
e
clas
s
i
f
icatio
n
o
f
7
clas
s
es
No
r
m
a
l a
n
d
ab
n
o
r
m
al
cla
s
s
i
f
i
ca
tio
n
alg
o
r
it
h
m
I
n
p
u
t
: H
id
d
en
la
y
er
W
eig
h
t o
f
ea
c
h
attr
ib
u
te,
Ma
x
an
d
Mi
n
W
eig
h
t o
f
ea
c
h
attr
ib
u
te,
Ou
tp
u
t W
eig
h
t o
f
ea
ch
attr
ib
u
te.
Ou
tp
u
t
: Cl
as
s
if
icatio
n
d
ataset
o
f
No
r
m
al
an
d
A
b
n
o
r
m
al
P
ap
s
m
ea
r
i
m
a
g
e.
P
r
o
ce
s
s
:
a.
S
ta
r
t.
A
ttrib
u
te
n
o
r
m
aliz
atio
n
.
No
r
m
aliza
t
io
n
=
(
(
d
ata
-
min
)
/(
ma
x
-
min
)
)
*
(
1
-
(
-
1
)
)
+(
-
1
)
;
P
er
f
o
r
m
n
o
r
m
aliza
t
io
n
o
n
ea
ch
at
tr
ib
u
t
e
*
Min
i
m
u
m
an
d
m
ax
i
m
u
m
v
alu
e
o
n
tr
ai
n
in
g
attr
ib
u
te.
b.
C
alcu
late
t
h
e
w
ei
g
h
t
o
f
ea
c
h
h
id
d
en
la
y
er
/
n
o
d
e
w
it
h
as
m
u
ch
w
ei
g
h
t
as
t
h
e
h
id
d
en
la
y
er
in
t
h
e
n
o
r
m
al
an
d
ab
n
o
r
m
al
cla
s
s
i
f
icatio
n
m
o
d
el.
B
eg
in
b
y
ca
lcu
la
ti
n
g
ea
c
h
h
id
d
en
la
y
er
o
b
tain
ed
f
r
o
m
t
h
e
m
u
ltip
licatio
n
o
f
attr
ib
u
t
es
t
h
at
h
av
e
b
ee
n
n
o
r
m
alize
d
w
it
h
ea
ch
w
eig
h
t
t
h
at
h
as
b
ee
n
d
eter
m
in
ed
i
n
s
elec
ted
attr
ib
u
tes
u
s
i
n
g
G
A
.
Fu
r
t
h
er
m
o
r
e,
ca
lcu
late
t
h
e
w
e
ig
h
t
o
f
ea
c
h
attr
ib
u
te
o
n
n
o
r
m
al
an
d
ab
n
o
r
m
al
clas
s
f
r
o
m
t
h
e
ca
lcu
latio
n
o
f
th
e
i
n
itial
w
e
ig
h
t.
C
alc
u
late
th
e
o
u
tp
u
t
w
e
ig
h
t
s
o
f
ea
ch
No
r
m
al
an
d
A
b
n
o
r
m
a
l Cl
a
s
s
o
u
tp
u
t v
al
u
e.
No
d
e
1
=
(
n
o
r
m
aisa
s
i_
attr
ib
u
te
*
attr
ib
u
te
w
ei
g
h
t)
+
b
ias
No
d
e
W
eig
h
t 1
=1
/(
1
+
E
x
p
(
-
n
o
d
e1
)
)
c.
C
alcu
late
th
e
w
ei
g
h
ts
o
f
ea
ch
n
o
r
m
al
a
n
d
ab
n
o
r
m
al
clas
s
i
f
ic
atio
n
.
C
alcu
late
ea
ch
clas
s
i
f
icatio
n
w
ei
g
h
t
o
b
tain
ed
f
r
o
m
t
h
e
m
u
lt
ip
licatio
n
o
f
ea
ch
w
ei
g
h
t
o
f
t
h
e
h
id
d
en
la
y
er
w
it
h
t
h
e
w
ei
g
h
t
o
f
t
h
e
n
o
d
es
s
p
ec
if
ied
in
ca
lc
u
latio
n
w
ei
g
h
t
o
f
h
id
d
en
la
y
er
.
Hid
d
en
la
y
er
s
ar
e
o
b
tain
ed
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
6
,
Decem
b
er
201
8
:
5
4
1
5
-
5
4
2
4
5422
f
r
o
m
th
e
m
u
ltip
licatio
n
o
f
attr
ib
u
tes
t
h
at
h
a
v
e
b
ee
n
n
o
r
m
alize
d
w
it
h
ea
ch
w
e
ig
h
t
t
h
at
h
as
b
ee
n
d
eter
m
in
ed
i
n
s
elec
ted
attr
ib
u
t
es
u
s
in
g
G
A
.
C
las
s
i
f
icatio
n
=(
h
id
d
en
_
la
y
er
_
w
ei
g
h
ts
*
n
o
d
e
w
eig
h
t)
+t
h
r
es
h
o
ld
C
las
s
i
f
icatio
n
w
ei
g
h
t
=
1
/(
1
+E
x
p
(
-
clas
s
i
f
icatio
n
)
)
d.
C
o
m
p
ar
e
th
e
class
if
ica
tio
n
weig
h
t
th
at
h
a
s
b
ee
n
c
alcu
lated
w
it
h
th
e
n
o
r
m
al
an
d
ab
o
n
o
r
m
al
w
ei
g
h
t.
I
f
th
e
n
o
r
m
al
w
ei
g
h
t
is
g
r
ea
ter
th
an
t
h
e
ab
n
o
r
m
al
w
ei
g
h
t,
th
e
clas
s
i
f
icatio
n
r
e
s
u
l
ts
ar
e
n
o
r
m
al,
b
u
t
o
th
er
w
is
e
t
h
e
class
if
icatio
n
b
e
co
m
e
s
ab
n
o
r
m
al.
C
las
s
i
f
icat
io
n
=i
f
n
o
r
m
al
w
ei
g
h
t>a
b
n
o
r
m
a
l
w
eig
h
t
R
es
u
lt=
n
o
r
m
a
l
I
f
n
o
t
R
es
u
lt=ab
n
o
r
m
al
e.
I
n
th
e
n
ex
t
s
ta
g
e,
p
er
f
o
r
m
t
h
e
s
a
m
e
p
r
o
ce
s
s
f
r
o
m
s
ta
g
e
a
-
d
b
y
p
er
f
o
r
m
i
n
g
ca
lc
u
lat
io
n
s
i
n
ea
ch
class
i
f
icatio
n
in
c
lu
d
i
n
g
:
n
o
r
m
al
class
ca
ls
s
i
f
icatio
n
1
,
2
,
3
,
A
b
n
o
r
m
al
clas
s
4
,
5
an
d
6
,
7
an
d
ab
n
o
r
m
al
class
5
an
d
6
.
4
.
2
.
Co
m
pa
ri
s
o
n
Resu
lt
s
o
f
Acc
ura
cy
Va
lues
T
ab
le
4
s
h
o
w
s
th
a
t
th
e
HD
A
m
o
d
el
h
a
s
a
s
u
p
er
io
r
ac
cu
r
ac
y
v
al
u
e
co
m
p
ar
ed
to
th
e
cla
s
s
i
f
icatio
n
alg
o
r
ith
m
r
e
s
u
lt
s
h
o
w
n
i
n
T
ab
le
4
.
T
h
e
r
esu
lts
o
b
tain
ed
f
r
o
m
t
h
e
r
esear
ch
s
h
o
w
s
th
at
t
h
e
class
i
f
icatio
n
m
o
d
el
o
f
HD
A
a
n
d
NN
a
lg
o
r
it
h
m
h
a
s
s
u
p
er
io
r
ac
cu
r
ac
y
co
m
p
ar
ed
to
o
th
er
class
if
icatio
n
al
g
o
r
it
h
m
s
.
Af
ter
d
o
in
g
t
h
e
r
esear
ch
,
th
e
clas
s
i
f
icatio
n
r
esu
lt
s
o
f
4
class
e
s
th
a
t
b
ec
am
e
th
e
m
ai
n
g
o
al
w
er
e
co
m
p
ar
ed
to
s
ee
w
h
ic
h
alg
o
r
ith
m
a
n
d
w
h
ic
h
m
e
th
o
d
i
s
b
est f
o
r
th
e
cla
s
s
i
f
icatio
n
in
t
o
7
class
e
s.
B
ased
o
n
th
e
test
th
at
h
a
s
b
e
en
o
b
tain
ed
o
n
th
e
P
ap
s
m
ea
r
i
m
a
g
e
d
ataset,
it
i
s
k
n
o
w
n
t
h
at
th
e
NN
an
d
HDA
al
g
o
r
ith
m
s
h
av
e
t
h
e
h
i
g
h
e
s
t
ac
cu
r
ac
y
w
i
th
t
h
e
v
alu
e
o
f
8
7
.
0
2
%
w
h
e
n
co
m
p
ar
ed
w
it
h
o
th
er
class
i
f
icatio
n
alg
o
r
it
h
m
s
.
T
ab
le
4
.
C
o
m
p
ar
is
o
n
o
f
A
cc
u
r
ac
y
Valu
e
s
A
l
g
o
r
i
t
h
m
A
c
c
u
r
a
c
y
P
r
o
p
o
se
d
M
e
t
h
o
d
8
7
,
0
2
%
D
e
c
i
si
o
n
T
r
e
e
+
H
D
A
[
1
4
]
8
3
,
2
6
G
A
+
H
D
A
N
o
n
O
p
t
i
mi
z
e
d
N
N
[
1
9
]
7
9
,
7
8
%
H
y
b
r
i
d
En
se
m
b
l
e
L
e
a
r
n
i
n
g
[
2
2
]
78
,
0
0
%
G
F
+
B
a
g
g
i
n
g
+
N
a
ï
v
e
B
a
y
e
s [
2
4
]
6
3
,
2
5
%
G
A
+
L
D
A
[
2
3
]
6
2
,
9
2
%
G
A
+
N
a
ï
v
e
B
a
y
e
s [
2
3
]
6
2
,
1
6
5.
CO
NCLU
SI
O
N
P
ap
s
m
ea
r
i
m
a
g
e
cla
s
s
i
f
icati
o
n
b
y
u
s
i
n
g
HD
A
m
et
h
o
d
with
t
h
e
cla
s
s
i
f
icatio
n
tes
t
i
n
to
7
class
e
s
(
n
o
r
m
al
s
u
p
er
f
icial,
n
o
r
m
a
l
in
ter
m
ed
iate,
n
o
r
m
al
co
lu
m
m
ar
,
m
ild
(
lig
h
t)
d
y
p
las
ia,
m
o
d
er
ate
d
y
p
lasi
a,
s
er
v
er
e
d
y
p
la
s
ia
an
d
ca
r
ci
n
o
m
a
i
n
s
it
u
)
h
a
s
th
e
h
i
g
h
est
ac
c
u
r
ac
y
v
alu
e
o
f
8
7
.
0
2
%.
T
h
e
r
esu
lts
o
b
tain
ed
f
r
o
m
t
h
e
HD
A
m
o
d
el
f
o
r
P
ap
s
m
ea
r
i
m
ag
e
cla
s
s
i
f
icat
io
n
i
n
to
7
class
es
w
er
e
co
m
p
ar
ed
to
th
e
class
i
f
icatio
n
r
esu
lt
s
u
s
i
n
g
th
e
NN
alg
o
r
it
h
m
a
n
d
f
ea
tu
r
e
o
p
ti
m
izatio
n
u
s
i
n
g
G
A
to
im
p
r
o
v
e
ac
cu
r
ac
y
.
I
n
th
i
s
w
o
r
k
w
e
p
r
o
p
o
s
e
a
class
i
f
icatio
n
m
eth
o
d
o
lo
g
y
i
n
a
s
in
g
le
ce
l
l
P
ap
s
m
ea
r
i
m
a
g
e.
T
h
is
task
is
p
ar
tic
u
lar
l
y
u
s
ef
u
l
f
o
r
n
o
r
m
al
a
n
d
ab
n
o
r
m
al
ce
ll
i
m
a
g
e
clas
s
i
f
ica
tio
n
i
n
ea
ch
cla
s
s
.
W
e
ca
n
co
m
e
o
u
t
w
it
h
th
e
f
ac
t
th
at
th
e
p
r
o
p
o
s
ed
m
et
h
o
d
h
as
n
o
t r
ea
ch
ed
a
v
er
y
h
i
g
h
le
v
el
o
f
ac
c
u
r
ac
y
.
Ho
w
e
v
er
,
w
e
n
ee
d
a
m
o
r
e
p
r
ac
tical,
p
r
ac
tical
al
ter
n
ati
v
e
m
et
h
o
d
to
class
i
f
y
P
ap
s
m
ea
r
i
m
ag
e
s
m
o
r
e
ac
cu
r
ately
.
As
f
u
t
u
r
e
w
o
r
k
,
w
e
i
n
te
n
d
to
ex
p
an
d
o
u
r
m
eth
o
d
u
s
i
n
g
h
y
b
r
id
m
o
d
eli
n
g
class
i
f
ica
tio
n
.
I
n
h
o
p
es
it
ca
n
f
u
r
t
h
er
i
m
p
r
o
v
e
th
e
ac
cu
r
ac
y
ac
h
ie
v
ed
.
T
h
u
s
,
f
r
o
m
t
h
e
r
esu
lts
o
f
o
u
r
m
o
d
el
te
s
ti
n
g
,
it
ca
n
b
e
co
n
clu
d
ed
th
at
th
e
H
D
A
m
e
th
o
d
f
o
r
P
ap
s
m
ea
r
i
m
a
g
e
clas
s
if
icati
o
n
ca
n
b
e
u
s
ed
as
a
r
ef
er
en
ce
f
o
r
i
n
itial
s
cr
ee
n
i
n
g
p
r
o
ce
s
s
to
an
a
l
y
ze
P
ap
s
m
e
ar
i
m
a
g
e
cla
s
s
i
f
icatio
n
.
F
u
r
th
er
r
esear
ch
w
il
l
b
e
d
o
n
e
b
y
m
a
k
i
n
g
w
eb
-
b
ased
a
p
p
licatio
n
s
,
a
n
d
t
h
e
p
er
f
o
r
m
a
n
ce
m
ea
s
u
r
e
m
en
t
o
f
w
eb
-
b
as
ed
ap
p
licatio
n
s
w
i
l
l
b
e
co
n
d
u
cted
b
y
u
s
er
s
w
h
o
ar
e
p
ath
o
lo
g
is
t
s
an
d
r
esear
ch
er
s
in
th
e
f
ield
o
f
ce
r
v
ical
ca
n
ce
r
.
ACK
NO
WL
E
D
G
E
M
E
NT
S
Au
t
h
o
r
s
w
o
u
ld
lik
e
to
t
h
an
k
R
I
ST
E
KDI
KT
I
.
T
h
is
r
esear
ch
w
as
s
u
p
p
o
r
ted
b
y
T
h
e
Min
is
tr
y
o
f
R
esear
ch
,
T
e
ch
n
o
lo
g
y
,
a
n
d
Hig
h
er
E
d
u
ca
tio
n
,
I
n
d
o
n
e
s
ia,
f
o
r
s
u
p
p
o
r
tin
g
th
i
s
r
esear
ch
th
r
o
u
g
h
T
h
e
P
asca
Do
cto
r
al
R
esear
ch
Gr
an
t
(
2
0
1
7
)
.
T
h
is
w
o
r
k
i
s
u
s
i
n
g
t
h
e
d
ata
f
r
o
m
:
P
ap
s
m
ea
r
B
e
n
ch
m
ar
k
Data
f
o
r
P
atter
n
C
las
s
i
f
icatio
n
J
.
J
an
tzen
,
J
.
N
o
r
u
p
,
G.
Do
u
n
ias,
a
n
d
B
.
B
j
e
r
r
eg
aa
r
d
,
Un
iv
er
s
i
t
y
Dep
t.
o
f
P
ath
o
lo
g
y
Her
le
v
R
in
g
v
ej
7
5
,
DK
-
2
7
3
0
Her
lev
,
Den
m
ar
k
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
I
mp
r
o
vin
g
Hiera
r
ch
ica
l D
ec
is
io
n
A
p
p
r
o
a
ch
fo
r
S
i
n
g
le
I
ma
g
e
C
la
s
s
if
ica
tio
n
o
f P
a
p
S
mea
r
(
D
w
iz
a
R
ia
n
a
)
5423
RE
F
E
R
E
NC
E
S
[1
]
"
W
o
rld
'
s
h
e
a
lt
h
m
in
isters
re
n
e
w
c
o
m
m
it
m
e
n
t
to
c
a
n
c
e
r
p
re
v
e
n
ti
o
n
a
n
d
c
o
n
tr
o
l.
,
"
M
a
y
2
0
1
7
.
[
O
n
li
n
e
]
.
A
v
a
il
a
b
le:
h
tt
p
:
//
ww
w
.
w
h
o
.
in
t
/ca
n
c
e
r/me
d
ia/n
e
w
s/c
a
n
c
e
rp
re
v
e
n
ti
o
n
-
re
so
lu
ti
o
n
/en
/
.
[2
]
D.
Rian
a
,
W
.
H.
D
w
i,
D.
E.
De
w
i
a
n
d
T
.
L
.
R.
M
e
n
g
k
o
,
"
S
e
g
m
e
n
tas
i
L
u
a
s
Nu
k
leu
s
S
e
l
No
rm
a
l
S
u
p
e
rf
isial
P
a
p
S
m
e
a
r
M
e
n
g
g
u
n
a
k
a
n
Op
e
ra
si
Ka
n
a
l
W
a
rn
a
Da
n
De
tek
si
T
e
p
i,
"
S
e
min
a
r
Na
si
o
n
a
l
I
n
o
v
a
si
d
a
n
T
e
k
n
o
lo
g
i
(
S
NIT
),
2
0
1
2
.
[3
]
J.
Ja
n
tze
n
,
J.
No
ru
p
,
G
.
Do
u
n
ia
s
a
n
d
B.
Bjerre
g
a
a
rd
,
"
P
a
p
-
sm
e
a
r
Be
n
c
h
m
a
rk
D
a
ta
F
o
r
P
a
tt
e
r
n
Clas
sif
ic
a
ti
o
n
,
"
2
0
0
5
.
[4
]
E
.
M
a
rti
n
,
"
P
a
p
-
S
m
e
a
r
Clas
si
f
ica
t
io
n
,
"
T
e
c
h
n
ica
l
Un
ive
rs
it
y
o
f
De
n
ma
rk
-
DTU,
2
0
0
3
.
[5
]
P
r
u
e
n
g
k
a
rn
,
Ra
tch
a
k
o
o
n
,
K
o
k
W
a
i
W
o
n
g
a
n
d
Ch
u
n
Ch
e
F
u
n
g
,
"
A
re
v
i
e
w
o
f
d
a
ta
m
in
in
g
tec
h
n
i
q
u
e
s
a
n
d
a
p
p
li
c
a
ti
o
n
s,"
J
o
u
rn
a
l
o
f
A
d
v
a
n
c
e
d
Co
mp
u
ta
t
io
n
a
l
I
n
telli
g
e
n
c
e
a
n
d
I
n
telli
g
e
n
t
In
fo
rm
a
t
ics
2
1
,
v
o
l.
1
,
p
p
.
3
1
-
4
8
,
2
0
1
7
.
[6
]
D.
Rian
a
,
D.
H.
W
id
y
a
n
to
ro
a
n
d
T
.
L
.
M
e
n
g
k
o
,
"
E
x
trac
ti
o
n
a
n
d
c
las
si
f
ica
ti
o
n
tex
tu
re
o
f
in
f
l
a
m
m
a
to
ry
c
e
ll
s
a
n
d
n
u
c
lei
i
n
n
o
rm
a
l
P
a
p
sm
e
a
r
i
m
a
g
e
s,"
in
ICI
-
BM
E
,
Ba
n
d
u
n
g
,
2
0
1
5
.
[7
]
D.
Rian
a
,
D.
E
.
O.
De
w
i,
D.
H.
W
id
y
a
n
to
ro
a
n
d
T
.
L
.
r.
M
e
n
g
k
o
,
"
Co
lo
r
c
a
n
a
ls
m
o
d
if
ica
ti
o
n
w
it
h
c
a
n
n
y
e
d
g
e
d
e
tec
ti
o
n
a
n
d
m
o
rp
h
o
l
o
g
ica
l
re
c
o
n
stru
c
t
io
n
f
o
r
c
e
ll
n
u
c
leu
s
se
g
m
e
n
tatio
n
a
n
d
a
re
a
m
e
a
su
re
m
e
n
t
in
n
o
rm
a
l
P
a
p
s
m
e
a
r
i
m
a
g
e
s,
"
in
AIP
,
Ba
n
d
u
n
g
,
2
0
1
4
.
[8
]
D.
Rian
a
,
D.
H.
W
id
y
a
n
to
ro
a
n
d
T
.
L
.
R.
M
e
n
g
k
o
,
"
In
f
la
m
m
a
to
r
y
c
e
ll
e
x
trac
ti
o
n
a
n
d
n
u
c
lei
d
e
tec
ti
o
n
i
n
P
a
p
sm
e
a
r
im
a
g
e
s,
"
In
t.
J
.
e
-
He
a
lt
h
M
e
d
.
Co
mm
u
n
,
v
o
l
.
6
,
p
p
.
2
7
-
4
3
,
2
0
1
5
.
[9
]
D.
Rian
a
,
D.
E.
O.
D
e
w
i,
D.
H
.
W
id
y
a
n
to
ro
a
n
d
T
.
L
.
R.
M
e
n
g
k
o
,
"
S
e
g
m
e
n
tatio
n
a
n
d
A
re
a
M
e
a
su
re
m
e
n
t
in
A
b
n
o
rm
a
l
P
a
p
sm
e
a
r
I
m
a
g
e
s
Us
in
g
Co
l
o
r
Ca
n
a
ls
M
o
d
if
ica
ti
o
n
w
it
h
Ca
n
n
y
Ed
g
e
De
tec
ti
o
n
,
"
in
In
I
n
ter
n
a
t
io
n
a
l
Co
n
fer
e
n
c
e
o
n
W
o
me
n
’s He
a
lt
h
i
n
S
c
ien
c
e
&
En
g
in
e
e
rin
g
,
Ba
n
d
u
n
g
,
2
0
1
2
.
[1
0
]
R.
Ku
rn
iaw
a
n
,
A
.
Ku
rn
iaw
a
rd
h
a
n
i
a
n
d
I.
M
u
h
im
m
a
h
,
"
In
f
la
m
m
a
to
ry
Ce
ll
Ex
tra
c
ti
o
n
in
P
a
p
sm
e
a
r
I
m
a
g
e
s:
A
Co
m
b
in
a
ti
o
n
o
f
Dista
n
c
e
Crit
e
rio
n
a
n
d
Im
a
g
e
T
ra
n
s
f
o
rm
a
ti
o
n
A
p
p
ro
a
c
h
,
"
T
EL
KOM
NIKA
T
e
le
c
o
mm
u
n
ica
t
io
n
,
Co
mp
u
t
in
g
,
El
e
c
tro
n
ics
a
n
d
Co
n
t
ro
l,
V
o
l
1
6
(5
)
:
2
0
4
8
-
2
0
5
6
;
2
0
1
8
.
[1
1
]
J.
H
y
e
o
n
,
H.
-
J.
Ch
o
i
a
n
d
B.
D.
L
e
e
,
"
Di
a
g
n
o
sin
g
Ce
rv
ic
a
l
Ce
ll
Im
a
g
e
s
Us
in
g
P
re
-
train
e
d
Co
n
v
o
lu
ti
o
n
a
l
Ne
u
ra
l
Ne
tw
o
rk
a
s F
e
a
tu
re
Ex
trac
to
r,
"
in
in
Bi
g
D
a
ta
a
n
d
S
ma
rt C
o
mp
u
ti
n
g
(
Bi
g
Co
mp
)
,
2
0
1
7
.
[1
2
]
D.
Ka
sh
y
a
p
,
A
.
S
o
m
a
n
i
a
n
d
J.
S
h
e
k
h
a
r,
"
Ce
r
v
ica
l
C
a
n
c
e
r
D
e
tec
ti
o
n
A
n
d
Clas
sif
i
c
a
ti
o
n
Us
in
g
In
d
e
p
e
n
d
e
n
t
L
e
v
e
l
S
e
ts A
n
d
M
u
lt
i
S
V
M
s,"
i
n
9
th
I
n
t
.
Co
n
f.
T
e
lec
o
mm
u
n
.
S
i
g
n
a
l
Pro
c
e
ss
,
2
0
1
6
.
[1
3
]
N.
L
a
ss
o
u
a
o
u
i,
L
.
Ha
m
a
m
i
a
n
d
N.
No
u
a
li
,
"
M
o
rp
h
o
l
o
g
ica
l
d
e
sc
rip
ti
o
n
o
f
c
e
rv
ica
l
c
e
ll
i
m
a
g
e
s
f
o
r
t
h
e
p
a
th
o
lo
g
ica
l
re
c
o
g
n
it
io
n
,
"
In
t.
J
.
M
e
d
.
He
a
lt
h
,
V
o
ls
.
1
,
No
5
,
p
p
.
3
1
3
-
3
1
6
,
2
0
0
7
.
[1
4
]
D.
Rian
a
,
"
Hie
ra
rc
h
ica
l
De
c
isio
n
A
p
p
ro
a
c
h
Be
rd
a
sa
rk
a
n
Im
p
o
rtan
c
e
P
e
rf
o
rm
a
n
c
e
A
n
a
l
y
sis
Un
tu
k
Kla
sif
i
k
a
s
Cit
ra
T
u
n
g
g
a
l
P
a
p
S
m
e
a
r
M
e
n
g
g
u
n
a
k
a
n
F
it
u
r
K
u
a
n
ti
tatif
d
a
n
Ku
a
li
tatif
,
"
2
0
1
0
.
[
On
li
n
e
]
.
A
v
a
il
a
b
le:
h
tt
p
:
//
ww
w
.
d
ig
il
ib
.
u
i.
a
c
.
id
/
d
e
tail?id
=
3
1
7
0
2
&
lo
k
a
si=
1
2
.
[
A
c
c
e
ss
e
d
2
0
M
a
rc
h
2
0
1
8
]
.
[1
5
]
T
.
K.
M
a
n
so
o
ri,
A
.
S
u
m
a
n
a
n
d
S
.
K.
M
ish
ra
,
"
F
e
a
tu
re
S
e
lec
ti
o
n
b
y
Ge
n
e
ti
c
A
lg
o
rit
h
m
a
n
d
S
VM
Cl
a
ss
if
i
c
a
ti
o
n
f
o
r
Ca
n
c
e
r
De
tec
t
io
n
,
"
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Ad
v
a
n
c
e
d
Res
e
a
rc
h
in
C
o
mp
u
ter
S
c
ien
c
e
a
n
d
S
o
ft
wa
re
En
g
i
n
e
e
rin
g
V
o
l
4
,
pp
.
3
5
7
-
3
6
5
,
2
0
1
4
.
[1
6
]
M
.
H.
A
.
Ya
z
id
,
S
.
T
a
li
b
,
M
.
H.
S
a
tri
a
a
n
d
A
.
A
.
G
h
a
z
i,
"
Ne
u
ra
l
Ne
tw
o
rk
o
n
M
o
rtali
ty
P
re
d
ictio
n
f
o
r
th
e
P
a
ti
e
n
t
A
d
m
it
ted
w
it
h
A
DH
F
(
A
c
u
te
De
c
o
m
p
e
n
sa
ted
He
a
rt
F
a
il
u
re
),
"
in
2
0
1
7
4
th
I
n
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
El
e
c
trica
l
En
g
i
n
e
e
rin
g
,
Co
m
p
u
ter
S
c
ie
n
c
e
a
n
d
I
n
fo
rm
a
ti
c
s
,
Yo
g
y
a
k
a
rta,
2
0
1
7
.
[1
7
]
M
.
F
.
M
o
h
a
m
m
e
d
a
n
d
T
.
H.
R
a
ss
e
m
,
"
A
n
En
se
m
b
le
o
f
En
h
a
n
c
e
d
F
u
z
z
y
M
in
M
a
x
Ne
u
ra
l
N
e
t
w
o
rk
s
f
o
r
Da
ta
Clas
sif
ic
a
ti
o
n
,
"
T
EL
KOM
NIKA
T
e
lec
o
mm
u
n
ica
ti
o
n
,
Co
mp
u
ti
n
g
,
E
lec
tro
n
ics
a
n
d
C
o
n
tr
o
l,
Vo
l
1
5
(
2
);
2
0
1
7
.
[1
8
]
A
.
G
.
K
a
re
g
o
w
d
a
,
A
.
M
a
n
ju
n
a
th
a
n
d
M
.
Ja
y
a
r
a
m
,
"
A
p
p
li
c
a
ti
o
n
of
Ge
n
e
ti
c
A
lg
o
rit
h
m
Op
ti
m
ize
d
Ne
u
ra
l
Ne
t
w
o
rk
Co
n
n
e
c
ti
o
n
W
e
ig
h
ts
f
o
r
M
e
d
ica
l
Dia
g
n
o
sis
Of
P
im
a
In
d
ian
s
Dia
b
e
tes
,
"
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
n
S
o
ft
Co
mp
u
ti
n
g
(
IJ
S
C
),
p
p
.
1
5
-
2
3
,
2
0
1
1
.
[1
9
]
Y.
Ra
m
d
h
a
n
i
a
n
d
D.
Ria
n
a
,
"
Hie
ra
rc
h
ica
l
De
c
isio
n
A
p
p
ro
a
c
h
B
a
se
d
o
n
Ne
u
ra
l
Ne
tw
o
rk
a
n
d
G
e
n
e
ti
c
A
lg
o
rit
h
m
M
e
th
o
d
f
o
r
S
i
n
g
le Im
a
g
e
Clas
si
f
i
c
a
ti
o
n
o
f
P
a
p
S
m
e
a
r,
"
in
In
fo
rm
a
t
ic a
n
d
Co
m
p
u
ti
n
g
(
ICIC)
,
Ja
y
a
p
u
ra
,
2
0
1
7
.
[2
0
]
M
.
Oro
z
c
o
-
M
o
n
tea
g
u
d
o
,
C.
M
ih
a
i,
H.
S
a
h
li
a
n
d
A
.
T
a
b
o
a
d
a
-
Crisp
i,
"
Co
m
b
in
e
d
Hie
ra
rc
h
ica
l
W
a
ters
h
e
d
S
e
g
m
e
n
tatio
n
a
n
d
S
V
M
C
las
sif
ic
a
ti
o
n
f
o
r
P
a
p
S
m
e
a
r
Ce
ll
Nu
c
leu
s E
x
trac
ti
o
n
,
"
2
0
1
2
.
[2
1
]
E.
J.
M
a
riarp
u
t
h
a
m
a
n
d
A
.
S
tep
h
e
n
,
"
No
m
in
a
ted
T
e
x
tu
re
Ba
se
d
Ce
rv
ica
l
Ca
n
c
e
r
Clas
si
f
ica
ti
o
n
,
"
Co
mp
u
t
a
ti
o
n
a
l
a
n
d
M
a
th
e
ma
ti
c
a
l
M
e
th
o
d
s i
n
M
e
d
icin
e
,
p
p
.
1
-
1
0
,
2
0
1
5
.
[2
2
]
A
.
S
a
r
wa
r,
V
.
S
h
a
rm
a
a
n
d
R.
Gu
p
ta,
"
Hy
b
rid
e
n
se
m
b
le
lea
rn
in
g
tec
h
n
iq
u
e
f
o
r
sc
re
e
n
in
g
o
f
c
e
r
v
ic
a
l
c
a
n
c
e
r
u
sin
g
P
a
p
a
n
ico
lao
u
sm
e
a
r
i
m
a
g
e
a
n
a
l
y
s
is,"
Per
so
n
a
li
ze
d
M
e
d
ici
n
e
Un
ive
rs
e
,
p
p
.
1
-
9
,
2
0
1
5
.
[2
3
]
Y.
Ra
m
d
h
a
n
i,
"
Ko
m
p
a
ra
si
A
lg
o
rit
m
a
LD
A
D
AN
N
a
iv
e
Ba
y
e
s
De
n
g
a
n
Op
ti
m
a
si
F
it
u
r
Un
tu
k
Kla
sif
i
k
a
si
Cit
ra
T
u
n
g
g
a
l
P
a
p
S
m
e
a
r,
"
In
fo
rm
a
ti
k
a
,
Vo
ls.
III
,
No
,
2
,
p
p
.
4
3
4
-
4
4
1
,
2
0
1
5
.
[2
4
]
D.
Rian
a
,
A
.
N.
Hid
a
y
a
n
to
a
n
d
F
i
tri
y
a
n
i,
"
In
teg
ra
ti
o
n
o
f
Ba
g
g
in
g
a
n
d
g
re
e
d
y
f
o
rw
a
rd
se
le
c
ti
o
n
o
n
i
m
a
g
e
p
a
p
s
m
e
a
r
c
las
si
f
ica
ti
o
n
u
sin
g
Na
ïv
e
Ba
y
e
s,
"
in
Cy
b
e
r a
n
d
I
T
S
e
rv
ice
M
a
n
a
g
e
me
n
t
(
CIT
S
M
)
,
Ba
li
,
2
0
1
7
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
6
,
Decem
b
er
201
8
:
5
4
1
5
-
5
4
2
4
5424
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
Dw
iz
a
Rian
a
w
a
s
b
o
rn
in
In
d
o
n
e
sia
in
1
9
7
0
.
S
h
e
is
a
a
ss
o
c
iate
p
r
o
f
e
ss
o
r
in
S
T
M
IK
Nu
sa
M
a
n
d
iri
Ja
k
a
rta
a
n
d
p
rin
c
ip
a
l
o
f
th
e
M
a
g
ister
Ilm
u
Ko
m
p
u
ter
a
t
S
T
M
IK
Nu
sa
M
a
n
d
ir
i.
S
h
e
d
i
d
h
e
r
BA
in
M
a
th
e
m
a
ti
c
a
t
th
e
Un
iv
e
rsit
y
o
f
S
riw
ij
a
y
a
,
In
d
o
n
e
sia
,
M
a
g
ister
o
f
M
a
n
a
g
e
m
e
n
t
at
Un
iv
e
rsit
y
o
f
Bu
d
i
L
u
h
u
r,
In
d
o
n
e
sia
,
M
a
g
iste
r
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
t
Un
iv
e
rsit
y
o
f
In
d
o
n
e
sia
a
n
d
P
h
D
i
n
El
e
c
tro
n
ica
l
En
g
i
n
e
e
rin
g
in
In
f
o
r
m
a
ti
c
s
a
t
In
stit
u
t
T
e
k
n
o
lo
g
i
Ba
n
d
u
n
g
,
I
n
d
o
n
e
sia
.
He
r
re
se
a
rc
h
in
th
e
a
re
a
o
f
Co
m
p
u
ter S
c
ien
c
e
,
Bio
m
e
d
ica
l
En
g
in
e
e
rin
g
,
Da
ta M
in
i
n
g
,
a
n
d
I
n
f
o
rm
a
ti
o
n
S
y
ste
m
.
Yu
d
i
Ra
m
d
h
a
n
i
w
a
s
b
o
rn
in
In
d
o
n
e
sia
i
n
1
9
9
0
.
He
d
i
d
h
e
r
BA
in
I
n
f
o
rm
a
ti
o
n
T
e
c
h
n
ica
l
a
t
th
e
Un
iv
e
rsit
y
BS
I,
In
d
o
n
e
sia
,
M
a
g
ister
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
t
S
T
M
IK
Nu
sa
M
a
n
d
iri
Ja
k
a
rta
,
In
d
o
n
e
sia
.
He
r
re
se
a
rc
h
in
th
e
a
re
a
o
f
Co
m
p
u
ter
S
c
ien
c
e
,
Bio
m
e
d
ica
l
En
g
in
e
e
rin
g
,
In
f
o
rm
a
ti
o
n
S
y
st
e
m
,
a
n
d
Big
Da
ta A
n
a
l
y
sis.
Rizk
i
T
ri
P
ra
se
ti
o
w
a
s
b
o
rn
in
I
n
d
o
n
e
sia
i
n
1
9
8
9
.
He
is
a
L
e
c
tu
re
r
in
U
n
iv
e
rsitas
BS
I.
He
d
i
d
h
is
Ba
c
h
e
lo
r
o
f
S
c
ien
c
e
in
In
f
o
rm
a
ti
o
n
S
y
ste
m
a
t
Un
iv
e
rsitas
B
S
I,
In
d
o
n
e
sia
a
n
d
M
a
g
ister
o
f
Co
m
p
u
ter
S
c
ien
c
e
in
S
T
M
IK
Nu
sa
M
a
n
d
iri
Ja
k
a
rta,
In
d
o
n
e
sia
.
Hi
s
re
se
a
rc
h
in
th
e
a
re
a
o
f
S
o
f
t
w
a
r
e
En
g
in
e
e
rin
g
,
Big
Da
ta an
d
Da
ta
M
in
i
n
g
,
a
n
d
M
o
b
il
e
C
o
m
p
u
ti
n
g
.
A
c
h
m
a
d
Niz
a
r
Hid
a
y
a
n
to
is
th
e
V
ice
De
a
n
f
o
r
Re
so
u
rc
e
s,
V
e
n
tu
re
s,
a
n
d
G
e
n
e
ra
l
A
d
m
in
istratio
n
,
F
a
c
u
lt
y
o
f
Co
m
p
u
ter
S
c
ien
c
e
,
Un
iv
e
rsitas
In
d
o
n
e
sia
.
He
re
c
e
iv
e
d
h
is
P
h
D
i
n
Co
m
p
u
ter
S
c
ie
n
c
e
f
ro
m
Un
iv
e
r
sitas
In
d
o
n
e
sia
.
His
re
se
a
r
c
h
in
tere
sts
a
re
r
e
late
d
to
in
f
o
rm
a
ti
o
n
m
a
n
a
g
e
m
e
n
t,
IT
d
if
f
u
sio
n
a
n
d
a
d
o
p
ti
o
n
,
e
-
c
o
m
m
e
rc
e
,
e
-
g
o
v
e
rn
m
e
n
t,
in
f
o
r
m
a
t
io
n
sy
ste
m
s
se
c
u
rit
y
,
c
h
a
n
g
e
m
a
n
a
g
e
m
e
n
t,
k
n
o
w
led
g
e
m
a
n
a
g
e
m
e
n
t
a
n
d
i
n
f
o
rm
a
ti
o
n
re
tri
e
v
a
l.
Evaluation Warning : The document was created with Spire.PDF for Python.