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
.
4
,
A
u
g
u
s
t
201
8
,
p
p
.
2
3
3
8
~
2
3
5
0
I
SS
N:
2
0
8
8
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v8
i
4
.
p
p
2
3
3
8
-
2350
2338
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
Feature
Selec
tion
Appro
a
ch b
a
sed
o
n F
irefly
Alg
o
rit
h
m
and
Chi
-
squa
re
E
m
a
d M
o
ha
m
ed
M
a
s
hh
o
ur
1
,
E
na
s
M
.
F
.
E
l H
o
ub
y
2
,
K
ha
led
T
a
w
f
ik
Wa
s
s
if
3
,
A
k
ra
m
I
.
Sa
la
h
4
1
Co
m
p
u
ter
S
c
ien
c
e
De
p
a
rtm
e
n
t
,
M
o
d
e
r
n
A
c
a
d
e
m
y
f
o
r
Co
m
p
u
ter S
c
ien
c
e
,
Ca
iro
,
Eg
y
p
t
2
S
y
st
e
m
s &
In
f
o
r
m
a
ti
o
n
De
p
a
rtme
n
t
-
En
g
in
e
e
rin
g
Div
isio
n
,
Na
ti
o
n
a
l
Re
se
a
rc
h
Ce
n
tre,
G
iz
a
,
Eg
y
p
t
3
Co
m
p
u
ter
S
c
ien
c
e
De
p
a
rtm
e
n
t
,
F
a
c
u
lt
y
o
f
Co
m
p
u
ters
a
n
d
In
f
o
rm
a
t
io
n
,
Ca
iro
U
n
iv
e
rsity
,
G
iza
,
Eg
y
p
t
4
Co
m
p
u
ter
S
c
ien
c
e
De
p
a
rtm
e
n
t
,
F
a
c
u
lt
y
o
f
Co
m
p
u
ters
a
n
d
I
n
f
o
rm
a
ti
o
n
,
Ca
i
ro
Un
iv
e
rsit
y
,
G
iz
a
,
Eg
y
p
t
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
l 3
0
,
2
0
1
7
R
ev
i
s
ed
Dec
2
2
,
2
0
1
7
A
cc
ep
ted
Dec
2
9
,
2
0
1
7
Dim
e
n
sio
n
a
li
ty
p
ro
b
lem
is
a
w
e
ll
-
k
n
o
w
n
c
h
a
ll
e
n
g
in
g
issu
e
f
o
r
m
o
st
c
las
si
f
iers
in
w
h
ich
d
a
tas
e
ts
h
a
v
e
u
n
b
a
lan
c
e
d
n
u
m
b
e
r
o
f
sa
m
p
les
a
n
d
f
e
a
tu
re
s.
F
e
a
tu
re
s
m
a
y
c
o
n
tain
u
n
re
li
a
b
le
d
a
ta
w
h
ich
m
a
y
lea
d
t
h
e
c
las
si
f
ica
ti
o
n
p
ro
c
e
ss
to
p
ro
d
u
c
e
u
n
d
e
sira
b
le
re
su
lt
s.
F
e
a
tu
re
se
lec
ti
o
n
a
p
p
ro
a
c
h
is
c
o
n
si
d
e
re
d
a
so
l
u
ti
o
n
f
o
r
th
is
k
in
d
o
f
p
ro
b
lem
s.
In
th
is
p
a
p
e
ra
n
e
n
h
a
n
c
e
d
f
iref
l
y
a
lg
o
rit
h
m
is
p
ro
p
o
se
d
to
se
rv
e
a
s
a
f
e
a
tu
re
se
lec
ti
o
n
s
o
lu
ti
o
n
f
o
r
re
d
u
c
in
g
d
im
e
n
sio
n
a
li
ty
a
n
d
p
ick
in
g
th
e
m
o
st
in
f
o
rm
a
ti
v
e
f
e
a
tu
re
s
to
b
e
u
se
d
in
c
las
sif
ica
ti
o
n
.
T
h
e
m
a
in
p
u
r
p
o
se
o
f
th
e
p
ro
p
o
se
d
m
o
d
e
l
is
to
im
p
ro
v
e
th
e
c
las
sif
ica
ti
o
n
a
c
c
u
ra
c
y
th
ro
u
g
h
u
si
n
g
th
e
se
le
c
ted
f
e
a
tu
re
s
p
ro
d
u
c
e
d
f
ro
m
th
e
m
o
d
e
l,
th
u
s
c
las
sif
ica
ti
o
n
e
rr
o
rs
w
il
l
d
e
c
re
a
se
.
M
o
d
e
li
n
g
f
ire
f
l
y
in
th
is
re
se
a
rc
h
a
p
p
e
a
rs
th
ro
u
g
h
sim
u
latin
g
f
ire
f
l
y
p
o
siti
o
n
b
y
c
e
ll
c
h
i
-
sq
u
a
re
v
a
lu
e
w
h
ich
is
c
h
a
n
g
e
d
a
f
ter
e
v
e
r
y
m
o
v
e
,
a
n
d
si
m
u
latin
g
f
ir
e
f
l
y
in
ten
sity
b
y
c
a
lcu
latin
g
a
se
t
o
f
d
if
f
e
r
e
n
t
f
it
n
e
ss
f
u
n
c
ti
o
n
sa
s
a
w
e
ig
h
t
f
o
r
e
a
c
h
fe
a
tu
re
.
K
-
n
e
a
re
st
n
e
ig
h
b
o
r
a
n
d
Disc
rim
in
a
n
t
a
n
a
ly
sis
a
re
u
se
d
a
s
c
las
si
f
iers
to
tes
t
th
e
p
ro
p
o
se
d
f
iref
l
y
a
lg
o
rit
h
m
in
se
le
c
ti
n
g
f
e
a
tu
re
s.
Ex
p
e
rime
n
tal
re
su
lt
s
sh
o
w
e
d
th
a
t
th
e
p
r
o
p
o
se
d
e
n
h
a
n
c
e
d
a
lg
o
rit
h
m
b
a
se
d
o
n
f
iref
l
y
a
lg
o
rit
h
m
w
it
h
c
h
i
-
sq
u
a
re
a
n
d
d
if
f
e
r
e
n
t
f
it
n
e
ss
f
u
n
c
ti
o
n
s
c
a
n
p
ro
v
id
e
b
e
tt
e
r
re
su
lt
s
th
a
n
o
th
e
rs.
Re
su
lt
s sh
o
w
e
d
th
a
t
re
d
u
c
ti
o
n
o
f
d
a
tas
e
t
is
u
se
f
u
l
f
o
r
g
a
in
in
g
h
ig
h
e
r
a
c
c
u
ra
c
y
in
c
las
sif
ica
ti
o
n
.
K
ey
w
o
r
d
:
C
h
i
-
s
q
u
ar
e
Featu
r
e
s
elec
t
io
n
Fire
f
l
y
alg
o
r
it
h
m
Fit
n
e
s
s
f
u
n
ctio
n
S
w
ar
m
i
n
telli
g
e
n
ce
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
:
E
m
ad
Mo
h
a
m
ed
Ma
s
h
h
o
u
r
,
C
o
m
p
u
ter
s
cien
ce
d
ep
ar
t
m
e
n
t
,
Mo
d
er
n
A
ca
d
e
m
y
f
o
r
C
o
m
p
u
t
er
Scien
ce
,
C
air
o
,
E
g
y
p
t
.
E
m
ail:
e_
m
a
s
h
h
o
u
r
@
h
o
t
m
a
il.
co
m
1.
I
NT
RO
D
UCT
I
O
N
A
h
u
g
e
d
i
m
e
n
s
io
n
alit
y
p
r
o
b
le
m
is
a
k
i
n
d
o
f
p
r
o
b
lem
th
a
t
ap
p
ea
r
s
in
d
ataset
s
w
h
ich
n
ee
d
s
to
b
e
s
i
m
p
li
f
ied
o
r
r
ed
u
ce
d
.
It
co
n
tain
s
a
lar
g
e
n
u
m
b
er
o
f
f
ea
t
u
r
es
a
g
ain
s
t
s
m
al
l
n
u
m
b
er
o
f
s
a
m
p
les.
A
lar
g
e
n
u
m
b
er
o
f
f
ea
tu
r
e
s
ar
e
co
n
s
id
er
ed
a
h
u
g
e
ch
a
llen
g
e
f
o
r
a
n
y
class
if
icatio
n
p
r
o
ce
s
s
.
Us
in
g
th
e
w
h
o
le
f
ea
t
u
r
e
s
w
il
l e
n
f
o
r
ce
t
h
e
cla
s
s
i
f
ier
to
es
ti
m
ate
u
n
s
ee
n
d
ata
w
it
h
a
p
r
e
-
k
n
o
w
led
g
e
o
f
u
n
d
esira
b
le
f
ea
t
u
r
es,
w
h
ic
h
i
n
t
u
r
n
w
il
l
p
r
o
d
u
ce
a
p
o
o
r
p
er
f
o
r
m
an
ce
f
o
r
an
y
cla
s
s
i
f
ier
[
1
]
.
Featu
r
e
s
elec
tio
n
ca
n
b
e
u
s
ed
f
o
r
r
ed
u
ci
n
g
d
i
m
en
s
io
n
al
it
y
o
f
d
ata
s
ets,
i
n
o
r
d
er
to
r
ed
u
ce
co
m
p
u
tat
io
n
ti
m
e,
co
s
t
an
d
cla
s
s
i
f
ic
atio
n
er
r
o
r
.
Ma
n
y
r
esear
ch
er
s
u
s
ed
s
ta
tis
tica
l
tech
n
iq
u
es
f
o
r
f
ea
tu
r
e
s
elec
ti
o
n
,
b
u
t
f
e
w
o
f
th
e
m
ap
p
l
y
s
w
ar
m
i
n
telli
g
e
n
ce
alg
o
r
ith
m
s
f
o
r
f
ea
tu
r
e
s
elec
tio
n
.
A
p
p
l
y
i
n
g
s
w
ar
m
i
n
telli
g
e
n
c
e
alg
o
r
ith
m
s
b
ec
a
m
e
a
m
o
ti
v
a
tio
n
f
o
r
r
esear
ch
er
s
to
s
o
lv
e
d
im
e
n
s
io
n
alit
y
p
r
o
b
le
m
s
d
u
e
to
its
ca
p
ab
ilit
y
f
o
r
s
elec
ti
n
g
t
h
e
m
o
s
t
ap
p
r
o
p
r
iat
e
f
ea
tu
r
es
u
s
ed
f
o
r
class
i
f
icati
o
n
.
S
w
ar
m
i
n
tel
lig
e
n
ce
ap
p
r
o
ac
h
ap
p
ea
r
ed
in
1
9
8
9
b
y
Ger
r
ad
o
an
d
J
in
g
w
a
n
g
[
2
]
.
I
t
w
as
i
n
s
p
ir
ed
b
y
th
e
m
u
tu
al
b
eh
a
v
io
r
th
at
ap
p
ea
r
s
o
n
n
at
u
r
e,
in
c
lu
d
i
n
g
w
ater
an
d
o
th
er
cr
ea
tu
r
es
s
u
c
h
as
i
n
s
e
cts.
I
n
t
h
is
k
i
n
d
o
f
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:
2
0
8
8
-
8708
F
ea
tu
r
e
S
elec
tio
n
A
p
p
r
o
a
ch
b
a
s
ed
o
n
F
ir
efly
A
lg
o
r
ith
m
a
n
d
C
h
i
-
s
q
u
a
r
e
(
E
ma
d
Mo
h
a
med
Ma
s
h
h
o
u
r
)
2339
ap
p
r
o
ac
h
ea
ch
in
d
iv
id
u
al
o
r
in
s
ec
t is ca
lled
an
a
g
e
n
t.
E
ac
h
a
g
en
t
w
o
r
k
s
i
n
d
ep
en
d
en
tl
y
i
n
a
t
y
p
e
o
f
co
lo
n
y
,
b
u
t
th
is
b
eh
a
v
io
r
is
co
n
tr
o
lled
b
y
ce
r
tain
r
u
les.
T
h
e
y
co
o
p
er
ate
w
it
h
o
th
er
s
i
n
o
r
d
er
to
f
i
n
i
s
h
a
ce
r
tain
tas
k
.
T
h
ese
ag
en
t
s
ar
e
co
n
s
id
er
ed
to
b
e
a
p
o
p
u
latio
n
th
at
i
n
ter
ac
ts
w
it
h
ea
ch
o
th
er
in
d
if
f
er
e
n
t
w
a
y
s
ac
co
r
d
in
g
to
th
e
t
y
p
e
o
f
th
e
i
n
s
ec
t.
Fo
r
ex
a
m
p
le,
p
h
er
o
m
o
n
e
i
f
a
m
o
n
g
an
ts
,
w
a
g
g
le
d
an
ce
a
m
o
n
g
b
ee
s
to
id
en
tify
s
o
u
r
ce
o
f
f
o
o
d
an
d
d
is
tan
ce
,
i
n
ten
s
it
y
an
d
f
l
ash
i
n
g
lig
h
t
a
m
o
n
g
f
ir
e
f
lie
s
.
Fo
r
m
u
la
tin
g
an
d
s
i
m
u
lati
n
g
s
w
ar
m
in
te
lli
g
en
ce
b
eh
av
io
r
d
ep
en
d
o
n
th
e
n
a
tu
r
e
o
f
th
e
p
r
o
b
lem
b
ei
n
g
s
o
lv
ed
.
I
n
t
h
is
p
ap
er
a
w
ell
-
k
n
o
w
n
s
w
ar
m
in
telli
g
en
ce
al
g
o
r
ith
m
ca
lled
f
ir
ef
l
y
[
3
]
i
s
u
s
ed
f
o
r
f
ea
tu
r
e
s
elec
tio
n
.
T
h
e
f
ir
e
f
l
y
al
g
o
r
ith
m
p
r
o
v
ed
its
ca
p
ab
ilit
y
to
s
o
lv
e
co
m
p
lex
o
p
ti
m
izatio
n
p
r
o
b
le
m
s
.
An
en
h
a
n
ce
d
f
ir
ef
l
y
al
g
o
r
ith
m
i
s
p
r
o
p
o
s
ed
i
n
th
i
s
p
ap
er
to
r
ed
u
ce
f
ea
tu
r
es
an
d
s
elec
t
th
e
m
o
s
t
i
n
f
o
r
m
ati
v
e
f
ea
t
u
r
es
f
o
r
th
e
class
i
f
icatio
n
p
r
o
ce
s
s
.
T
h
e
m
o
d
i
f
icat
io
n
s
f
o
r
th
e
s
ta
n
d
ar
d
f
ir
ef
l
y
alg
o
r
it
h
m
ar
e
r
ep
r
esen
ted
th
r
o
u
g
h
co
n
s
id
er
in
g
th
e
p
o
s
it
io
n
o
f
f
ir
ef
l
y
a
s
ch
i
-
s
q
u
ar
e
v
al
u
e
as
s
ig
n
ed
f
o
r
ea
ch
v
al
u
e
i
n
th
e
f
ea
t
u
r
e
v
ec
to
r
.
A
n
d
a
s
et
o
f
d
if
f
er
e
n
t
f
it
n
e
s
s
f
u
n
ctio
n
s
s
u
ch
as
R
o
s
e
n
b
r
o
ck
,
Sp
h
er
e,
Ack
le
y
[
4
]
,
Xin
-
Sh
e
y
a
n
g
,
r
astrig
i
n
,
s
ch
w
e
f
el,
an
d
Salo
m
o
n
[
5
]
ar
e
u
s
ed
to
r
e
p
r
esen
t
i
n
ten
s
it
y
o
f
f
ir
e
f
lie
s
.
T
h
e
p
er
f
o
r
m
an
ce
o
f
t
h
e
p
r
o
p
o
s
ed
m
o
d
el
is
test
ed
u
s
i
n
g
t
h
e
K
-
Nea
r
est
Ne
ig
h
b
o
r
(
K
-
NN)
[
6
]
an
d
Dis
cr
i
m
i
n
a
n
t
An
al
y
s
i
s
(
D
A
)
[
7
]
class
i
f
i
er
s
to
m
ea
s
u
r
e
t
h
e
class
i
f
icatio
n
ac
cu
r
ac
y
u
s
in
g
t
h
e
s
elec
ted
f
ea
tu
r
es.
D
if
f
er
en
t
tec
h
n
iq
u
e
s
h
a
v
e
b
ee
n
ap
p
lied
f
o
r
f
ea
tu
r
e
s
elec
tio
n
an
d
class
i
f
icat
io
n
in
m
an
y
lit
er
atu
r
es.
A
n
u
m
b
er
o
f
r
elate
d
r
esear
ch
es
w
h
ic
h
ap
p
lied
f
ea
tu
r
e
s
elec
ti
o
n
tech
n
iq
u
es
a
n
d
class
if
ica
ti
o
n
ar
e
h
ig
h
li
g
h
ted
.
Sin
atab
a
k
h
i
et
a
l
.
[
8
]
p
r
o
p
o
s
ed
a
tech
n
iq
u
e
f
o
r
r
ed
u
cin
g
h
ig
h
d
i
m
e
n
s
io
n
alit
y
i
n
d
atase
ts
.
A
n
u
n
s
u
p
er
v
i
s
ed
g
en
e
s
elec
t
io
n
tec
h
n
iq
u
e
w
a
s
i
n
tr
o
d
u
ce
d
to
b
e
ap
p
lied
o
n
m
icr
o
ar
r
a
y
d
ataset
s
s
u
c
h
a
s
S
R
B
C
T
,
C
o
lo
n
,
p
r
o
s
tate
tu
m
o
r
,
leu
k
e
m
ia
an
d
lu
n
g
.
T
h
e
p
r
o
p
o
s
ed
tech
n
iq
u
e
u
tili
ze
d
an
t
co
lo
n
y
o
p
ti
m
i
za
tio
n
alg
o
r
ith
m
to
m
i
n
i
m
ize
t
h
e
r
ed
u
n
d
an
c
y
b
et
w
ee
n
g
e
n
es
a
n
d
i
n
cr
ea
s
e
r
ele
v
an
ce
o
f
g
en
e
s
.
T
h
e
y
tr
ied
v
ar
i
an
t
f
it
n
es
s
f
u
n
ct
io
n
s
th
at
m
a
y
i
m
p
r
o
v
e
th
e
cla
s
s
i
f
ic
atio
n
r
ate
an
d
s
elec
t
lo
w
er
n
u
m
b
er
o
f
g
e
n
es.
T
h
e
y
co
m
p
ar
e
d
th
eir
r
esu
lt
s
w
it
h
d
if
f
er
e
n
t
u
n
s
u
p
er
v
is
ed
a
n
d
s
u
p
er
v
is
ed
g
e
n
e
s
elec
tio
n
m
et
h
o
d
s
,
clas
s
i
f
icatio
n
ac
cu
r
ac
y
h
as
b
ee
n
m
ea
s
u
r
ed
b
ase
d
o
n
th
r
ee
d
if
f
er
en
t
clas
s
i
f
ier
s
w
h
ich
ar
e
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
i
n
e,
n
aï
v
e
B
ay
e
s
an
d
d
ec
is
io
n
tr
ee
.
Sh
ar
m
a
A
lo
k
et
a
l
.
[
9
]
in
tr
o
d
u
ce
d
a
te
ch
n
iq
u
e
f
o
r
f
ea
tu
r
e
s
elec
tio
n
b
ased
o
n
f
i
x
ed
p
o
in
t
alg
o
r
it
h
m
.
T
h
e
y
ap
p
lied
t
h
eir
tech
n
iq
u
e
o
n
h
u
m
a
n
ca
n
ce
r
d
atasets
u
s
in
g
m
icr
o
ar
r
a
y
g
e
n
e
ex
p
r
ess
io
n
.
T
h
e
u
s
a
g
e
o
f
f
ix
ed
p
o
in
t
alg
o
r
it
h
m
in
co
r
p
o
r
ated
w
it
h
P
C
A
(
p
r
in
cip
al
co
m
p
o
n
e
n
t
an
al
y
s
is
)
d
o
esn
’
t
n
ee
d
class
lab
els
f
o
r
f
ea
t
u
r
e
v
ec
to
r
s
.
On
co
n
tr
ar
y
an
e
ig
e
n
v
ec
to
r
is
co
m
p
u
ted
b
y
m
u
ltip
l
y
i
n
g
co
v
ar
i
an
ce
m
a
tr
ix
i
ter
ativ
el
y
to
s
e
le
ct
th
e
d
esire
d
g
e
n
es.
T
h
ey
ap
p
lied
th
eir
tech
n
iq
u
e
o
n
t
h
r
ee
p
u
b
lic
d
ata
s
ets
w
h
ich
ar
e
SR
B
C
T
,
AL
L
a
n
d
A
M
L
,
an
d
t
h
e
y
u
s
ed
J
4
.
8
an
d
NB
f
o
r
class
i
f
ica
tio
n
.
C
h
i
n
n
as
w
a
m
y
A
r
u
n
k
u
m
ar
a
n
d
R
a
m
ak
r
i
s
h
n
a
n
Sri
n
i
v
asa
n
[
1
0
]
p
r
o
p
o
s
ed
a
tech
n
iq
u
e
f
o
r
d
ev
elo
p
in
g
f
ea
t
u
r
e
s
elec
tio
n
p
r
o
ce
s
s
to
r
ed
u
ce
h
i
g
h
d
i
m
e
n
s
io
n
alit
y
d
atase
ts
.
T
h
eir
tech
n
iq
u
e
co
m
b
i
n
ed
co
r
r
elatio
n
co
ef
f
ici
en
t
w
it
h
p
ar
ticle
s
w
ar
m
o
p
ti
m
izatio
n
,
in
w
h
ic
h
co
r
r
elatio
n
co
ef
f
icie
n
t
w
a
s
u
s
ed
d
u
e
to
it
s
ca
p
ab
ilit
y
to
d
ete
ct
r
elatio
n
s
h
ip
b
et
w
ee
n
g
e
n
e
s
.
P
ar
ticle
s
w
ar
m
o
p
t
i
m
izati
o
n
w
as
u
s
ed
a
s
a
s
ea
r
ch
i
n
g
tec
h
n
iq
u
e
f
o
r
th
e
m
o
s
t
v
al
u
ab
le
g
e
n
es.
T
h
e
y
ap
p
l
ied
th
eir
tech
n
iq
u
e
o
n
t
h
r
ee
m
icr
o
ar
r
a
y
d
ataset
s
w
h
ic
h
ar
e
SR
B
C
T
,
L
y
m
p
h
o
m
a
an
d
ML
L
.
E
x
tr
e
m
e
lear
n
i
n
g
m
ac
h
i
n
es
clas
s
i
f
ier
w
a
s
u
s
ed
as
a
class
if
ier
f
o
r
ev
alu
a
tin
g
th
e
f
ea
tu
r
e
s
elec
t
i
o
n
p
r
o
ce
s
s
.
T
h
e
y
co
m
p
ar
ed
t
h
eir
r
es
u
lt
s
w
i
th
d
if
f
er
e
n
t
cla
s
s
i
f
ier
s
s
u
ch
as
j
4
8
,
r
an
d
o
m
f
o
r
est,
r
an
d
o
m
tr
ee
,
d
ec
is
io
n
s
t
u
m
p
a
n
d
g
en
et
i
c
p
r
o
g
r
a
m
m
in
g
.
P
ar
v
ee
n
An
is
th
a
n
a
et
a
l
.
[
1
1
]
p
r
o
p
o
s
ed
a
f
ea
tu
r
e
s
elec
tio
n
m
eth
o
d
to
eli
m
i
n
ate
r
ed
u
n
d
an
t
a
n
d
ir
r
elev
an
t
f
ea
t
u
r
es
f
r
o
m
d
at
asets
a
n
d
i
m
p
r
o
v
e
class
i
f
icatio
n
ac
c
u
r
ac
y
.
P
r
in
cip
al
co
m
p
o
n
e
n
t
a
n
al
y
s
i
s
(
P
C
A
)
,
r
o
u
g
h
P
C
A
,
u
n
s
u
p
er
v
i
s
ed
q
u
ic
k
r
ed
u
ct
s
alg
o
r
ith
m
a
n
d
e
m
p
ir
ical
d
is
tr
i
b
u
tio
n
r
an
k
i
n
g
ar
e
u
s
ed
f
o
r
f
ea
tu
r
e
s
elec
tio
n
p
r
o
ce
s
s
.
Fi
v
e
d
atasets
ar
e
test
ed
f
o
r
th
eir
tech
n
iq
u
es
w
h
ic
h
ar
e
lu
n
g
ca
n
ce
r
,
b
r
ea
s
t
ca
n
ce
r
,
d
iab
etes,
h
ea
r
t
an
d
ec
o
li.
A
n
u
m
b
er
o
f
class
if
ier
s
w
er
e
u
s
ed
s
u
c
h
a
s
J
R
ip
,
J
4
8
,
R
B
FN,
Naï
v
e
B
a
y
e
s
,
d
ec
is
io
n
tab
le
a
n
d
k
-
s
tar
.
K
G
Sri
n
iv
a
s
a
et
a
l
.
[
1
2
]
in
tr
o
d
u
ce
d
a
tech
n
iq
u
e
f
o
r
ex
tr
ac
tin
g
in
f
o
r
m
ati
v
e
f
ea
tu
r
es
f
o
r
class
if
icatio
n
u
s
in
g
f
u
zz
y
c
-
m
ea
n
s
clu
s
ter
in
g
.
T
h
e
clu
s
ter
ce
n
ter
is
cr
ea
ted
s
u
c
h
th
a
t
it
is
clo
s
er
to
f
ea
tu
r
es
w
ith
g
r
ea
ter
m
e
m
b
er
s
h
ip
.
Fo
u
r
d
atasets
w
er
e
test
ed
u
s
in
g
t
h
eir
tec
h
n
iq
u
e,
s
u
ch
a
s
p
h
y
s
ic
s
,
s
o
n
ar
,
d
er
m
a
to
lo
g
y
,
a
n
d
w
av
e
f
o
r
m
d
ata
s
ets.
T
w
o
clas
s
i
f
ier
w
er
e
u
s
ed
w
h
ic
h
ar
e
SVM
a
n
d
Ar
tif
icial
Neu
r
al
Net
w
o
r
k
(
ANN)
.
Me
i
-
L
i
n
g
H
u
an
g
et
a
l
.
[
1
3
]
in
tr
o
d
u
ce
d
a
f
r
a
m
e
w
o
r
k
f
o
r
s
o
l
v
i
n
g
t
h
e
p
r
o
b
le
m
o
f
d
ata
d
i
m
en
s
io
n
ali
t
y
b
y
ap
p
l
y
i
n
g
f
ea
t
u
r
e
s
elec
t
io
n
p
r
o
ce
s
s
o
n
d
atasets
.
T
h
ey
co
m
b
i
n
ed
SVM
w
ith
r
ec
u
r
s
iv
e
f
e
at
u
r
e
eli
m
i
n
atio
n
ap
p
r
o
ac
h
.
A
m
et
h
o
d
ca
lled
tag
u
c
h
i
p
ar
a
m
e
ter
o
p
tim
izatio
n
h
as
b
ee
n
u
s
ed
f
o
r
id
en
tify
in
g
t
h
e
p
ar
a
m
eter
v
alu
e.
T
h
e
y
u
s
ed
t
w
o
p
u
b
lic
d
atasets
w
h
ich
ar
e
d
er
m
ato
lo
g
y
an
d
zo
o
d
ataset.
Han
y
M.
Har
b
a
n
d
A
b
ee
r
S.
Desu
k
y
[
1
4
]
tr
ied
to
in
v
en
t
a
m
et
h
o
d
f
o
r
r
ed
u
ci
n
g
f
ea
t
u
r
es
.
T
h
e
y
u
s
ed
p
ar
ticle
s
w
ar
m
o
p
ti
m
izat
io
n
f
o
r
i
m
p
le
m
en
tin
g
f
ea
t
u
r
e
s
elec
tio
n
m
et
h
o
d
,
th
r
ee
m
ed
ical
d
atasets
w
er
e
u
s
ed
:
d
er
m
ato
lo
g
y
,
b
r
ea
s
t c
a
n
ce
r
a
n
d
h
ea
r
t
s
ta
tlo
g
d
ataset
s
.
P
SO
is
u
s
ed
a
s
e
ar
ch
in
g
m
et
h
o
d
f
o
r
f
ea
t
u
r
es
a
n
d
C
FS
i
s
u
s
ed
f
o
r
m
ea
s
u
r
i
n
g
th
e
u
s
ef
u
l
n
es
s
o
f
ea
ch
f
ea
tu
r
e.
F
iv
e
c
lass
if
ier
s
w
er
e
u
s
ed
f
o
r
ev
alu
a
tin
g
f
ea
t
u
r
es
w
h
ich
ar
e
NB
,
B
ay
esia
n
,
r
ad
ial
b
asis
f
u
n
ctio
n
n
et
w
o
r
k
(
R
B
F),
d
ec
is
i
o
n
T
r
ee
an
d
K
-
NN.
T
h
ey
co
m
p
ar
ed
th
e
ir
tech
n
iq
u
e
w
it
h
g
en
et
ic
alg
o
r
it
h
m
an
d
d
if
f
er
e
n
t
co
m
b
i
n
at
io
n
b
et
w
ee
n
P
SO
an
d
d
if
f
er
en
t
class
i
f
ier
s
.
P
in
ar
y
ild
ir
i
m
[
1
5
]
p
r
o
p
o
s
ed
d
if
f
er
en
t
co
m
b
in
at
io
n
s
o
f
f
ea
t
u
r
e
s
ele
ctio
n
m
eth
o
d
s
a
n
d
class
i
f
icatio
n
tec
h
n
iq
u
es
to
s
elec
t
in
f
o
r
m
at
iv
e
f
ea
t
u
r
es
f
r
o
m
h
ig
h
d
i
m
e
n
s
io
n
alit
y
d
atas
et.
I
n
t
h
is
r
esear
c
h
f
ea
t
u
r
e
s
elec
tio
n
m
e
t
h
o
d
s
s
u
c
h
as
C
f
s
S
u
b
s
et
E
v
al,
P
r
in
ci
p
al
C
o
m
p
o
n
e
n
ts
,
C
o
n
s
i
s
ten
c
y
Su
b
s
e
t
E
v
a
l,
I
n
f
o
Gain
Attr
ib
u
te
E
v
a
l,
O
n
e
R
A
ttrib
u
te
E
v
al
an
d
R
elie
f
Att
r
ib
u
te
E
v
a
l
w
er
e
co
m
p
ar
ed
.
Hep
atitis
d
ata
s
et
w
a
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2088
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
4
,
A
u
g
u
s
t 2
0
1
8
:
2
3
3
8
–
2
3
5
0
2340
u
s
ed
a
s
a
ca
s
e
s
tu
d
y
d
u
e
to
it
s
s
er
io
u
s
h
ea
lt
h
p
r
o
b
le
m
.
Fo
u
r
d
if
f
er
e
n
t
clas
s
if
ier
s
w
er
e
u
s
ed
w
h
ic
h
ar
e
J
4
8
,
NB
,
I
B
K
an
d
d
ec
is
io
n
tab
le
.
Na
n
c
y
P
et
a
l
.
[
1
6
]
in
tr
o
d
u
ce
d
a
s
t
u
d
y
to
ex
p
lo
r
e
a
s
et
o
f
f
e
atu
r
e
s
e
lectio
n
an
d
class
i
f
icatio
n
m
e
th
o
d
s
ap
p
lie
d
o
n
h
ep
atitis
d
ata
s
et.
Fo
r
f
e
atu
r
e
s
elec
tio
n
f
is
h
er
f
i
lter
in
g
,
r
elief
f
i
lter
i
n
g
a
n
d
s
tep
d
is
c
w
er
e
u
s
ed
.
Fo
r
cla
s
s
i
f
icatio
n
m
o
r
e
t
h
a
n
1
0
clas
s
if
ica
tio
n
al
g
o
r
ith
m
s
h
as
b
ee
n
u
s
ed
,
n
u
m
b
er
o
f
f
ea
t
u
r
es
s
elec
ted
f
o
r
th
r
ee
m
e
th
o
d
s
w
as
6
f
o
r
f
i
s
h
er
f
ilter
i
n
g
,
9
f
o
r
r
elief
f
i
lter
in
g
a
n
d
4
f
o
r
s
tep
d
is
c
.
Sm
ita
C
h
o
r
m
u
n
g
e
an
d
Su
d
ar
s
o
n
J
en
a
[
1
7
]
p
r
o
p
o
s
ed
a
f
ea
tu
r
e
s
ele
ctio
n
al
g
o
r
ith
m
b
ased
o
n
i
n
f
o
r
m
atio
n
g
ai
n
.
T
h
e
y
ap
p
lied
a
f
ilter
m
et
h
o
d
an
d
t
h
en
ap
p
lied
i
n
f
o
r
m
atio
n
g
ai
n
m
ea
s
u
r
e
f
o
r
th
e
s
u
b
s
et
p
r
o
d
u
ce
d
.
T
w
o
d
i
f
f
er
e
n
t
class
i
f
ier
s
w
er
e
u
s
ed
Naïv
e
b
a
y
es
a
n
d
I
B
K
an
d
m
ed
ical
d
atasets
w
er
e
u
s
ed
a
s
a
ca
s
e
s
t
u
d
y
s
u
ch
a
s
S
R
B
C
T
.
Hig
h
p
er
ce
n
tag
e
w
a
s
p
r
o
d
u
ce
d
b
y
I
n
f
o
r
m
atio
n
g
ai
n
co
m
p
ar
ed
w
it
h
R
elie
f
a
n
d
C
HI
-
s
q
u
ar
e
m
eth
o
d
s
.
R
esear
ch
er
s
f
r
o
m
[
8
]
to
[
1
7
]
ar
e
u
s
i
n
g
o
n
e
o
r
m
o
r
e
o
f
o
u
r
d
atasets
,
t
h
er
ef
o
r
e,
a
co
m
p
ar
i
s
o
n
o
f
o
u
r
r
esu
lt
s
w
ith
t
h
e
ab
o
v
e
r
esea
r
ch
es
w
ill
b
e
co
n
d
u
cted
an
d
d
em
o
n
s
tr
ated
in
s
ec
tio
n
5
.
1
.
Featu
r
e
s
elec
tio
n
s
o
lu
tio
n
w
as
ap
p
lied
o
n
d
i
f
f
er
en
t
k
in
d
s
o
f
d
atase
ts
,
f
o
r
ex
a
m
p
le
in
Si
n
g
h
an
d
C
h
h
i
k
ar
a
[
1
8
]
p
r
o
p
o
s
ed
a
m
o
d
el
f
o
r
d
etec
tin
g
f
ea
t
u
r
es
o
f
i
m
ag
es
e
x
tr
ac
ted
f
r
o
m
d
is
cr
ete
w
a
v
elet
tr
an
s
f
o
r
m
(
DW
T
)
a
n
d
d
is
cr
ete
co
s
in
e
tr
an
s
f
o
r
m
(
DC
T
)
u
s
i
n
g
f
ir
ef
l
y
a
lg
o
r
it
h
m
co
m
b
i
n
ed
w
it
h
SVM
class
if
ier
.
W
h
il
s
t
L
o
n
g
Z
h
a
n
g
et
a
l
.
[
1
9
]
d
etec
ted
th
e
m
o
s
t
i
n
f
o
r
m
ati
v
e
f
ea
t
u
r
es
in
m
ed
ical
d
atase
t
s
u
s
i
n
g
f
ir
ef
l
y
alg
o
r
it
h
m
b
as
ed
o
n
d
is
ta
n
ce
w
it
h
m
u
tu
al
i
n
f
o
r
m
atio
n
cr
iter
io
n
.
T
h
ey
u
s
ed
K
-
NN
an
d
SVM
as
class
i
f
ier
s
to
m
ea
s
u
r
e
th
e
p
er
f
o
r
m
a
n
ce
o
f
t
h
e
p
r
o
p
o
s
ed
tech
n
iq
u
e
.
V.
S
u
b
h
a
an
d
D.
Mu
r
u
g
an
[
20
]
in
tr
o
d
u
ce
d
a
tech
n
iq
u
e
f
o
r
s
o
lv
in
g
t
h
e
h
i
g
h
d
i
m
en
s
io
n
al
it
y
p
r
o
b
le
m
f
o
r
ca
r
d
io
to
co
g
r
am
(
C
T
G)
d
ata.
Fire
f
l
y
alg
o
r
it
h
m
w
as
u
s
ed
w
i
t
h
a
n
o
v
e
l
ap
p
r
o
ac
h
ca
lled
o
p
p
o
s
itio
n
b
ase
lear
n
in
g
(
OB
L
)
.
E
n
n
y
I
Sela,
et
a
l
[
21
]
ex
tr
ac
t
f
ea
tu
r
es
f
r
o
m
X
-
R
a
y
i
m
a
g
es,
r
esear
ch
er
s
d
ev
elo
p
ed
an
alg
o
r
ith
m
to
e
x
tr
ac
t
f
ea
t
u
r
e
o
f
i
m
a
g
es
p
r
o
d
u
ce
d
f
r
o
m
h
u
m
an
b
o
d
y
.
Sa
m
p
le
s
ex
tr
ac
ted
ar
e
f
o
r
X
-
R
a
y
d
en
ta
l
b
o
n
e
to
id
en
t
if
y
w
o
m
e
n
w
it
h
lo
w
s
k
ele
tal
B
MD
,
J
4
.
8
is
u
s
ed
to
ev
al
u
ate
t
h
e
f
ea
t
u
r
es
e
x
tr
ac
ted
f
r
o
m
f
ea
tu
r
e
s
elec
tio
n
alg
o
r
it
h
m
,
r
esu
lts
p
r
o
v
ed
th
at
t
h
eir
tech
n
iq
u
e
ac
h
iev
e
h
i
g
h
ac
c
u
r
ac
y
,
s
en
s
iti
v
it
y
,
a
n
d
s
p
ec
if
icit
y
.
Ad
i
Su
r
y
ap
u
tr
a
P
ar
a
m
ita
[
2
2
]
p
r
o
p
o
s
ed
an
alg
o
r
ith
m
f
o
r
f
ea
t
u
r
e
s
elec
tio
n
ap
p
lied
o
n
in
ter
n
et
tr
a
f
f
ic
d
ata.
T
h
e
y
u
s
ed
P
C
A
f
o
r
ex
tr
ac
tin
g
d
is
cr
i
m
i
n
an
t
f
ea
t
u
r
es
in
d
ata.
Fu
zz
y
c
-
m
ea
n
i
s
u
s
ed
to
i
m
p
r
o
v
e
K
-
N
N
class
i
f
ier
p
er
f
o
r
m
a
n
ce
.
B
y
d
is
tr
ib
u
ti
n
g
a
n
d
g
r
o
u
p
in
g
d
ata
in
to
clu
s
ter
s
.
R
es
u
lts
p
r
o
v
ed
th
at
w
h
e
n
u
s
in
g
P
C
A
as
a
f
ea
t
u
r
e
s
elec
tio
n
s
o
lu
t
io
n
w
it
h
K
-
NN
a
n
d
f
u
zz
y
C
-
Me
an
,
i
t
o
u
tp
er
f
o
r
m
o
th
er
tech
n
iq
u
es.
T
h
e
r
em
ai
n
d
er
o
f
th
e
p
ap
er
is
o
r
g
an
ized
as
f
o
llo
w
s
:
Sec
tio
n
2
r
ev
ie
w
s
ar
ti
f
icial
f
ir
ef
l
y
al
g
o
r
ith
m
i
n
g
en
er
al.
Sect
io
n
3
p
r
esen
ts
t
h
e
p
r
o
p
o
s
ed
s
o
lu
tio
n
i
n
th
i
s
r
es
e
ar
ch
f
o
llo
w
ed
b
y
th
e
e
n
h
an
c
ed
f
ir
ef
l
y
alg
o
r
it
h
m
.
Sectio
n
4
p
r
ese
n
ts
r
es
u
lts
an
d
an
al
y
s
i
s
.
Sect
io
n
5
p
r
ese
n
ts
e
x
p
er
i
m
e
n
t
d
is
c
u
s
s
io
n
f
o
llo
w
e
d
b
y
a
co
m
p
ar
ati
v
e
tab
le
th
at
co
m
p
ar
es o
u
r
ap
p
r
o
ac
h
w
it
h
o
th
er
r
esear
ch
e
s
.
Sec
tio
n
6
p
r
esen
ts
co
n
cl
u
s
io
n
&
f
u
tu
r
e
w
o
r
k
.
2.
ARTI
F
I
CI
AL
F
I
R
E
F
L
Y
AL
G
O
RI
T
H
M
Fire
f
l
y
al
g
o
r
ith
m
is
co
n
s
id
er
e
d
to
b
e
a
m
eta
-
h
e
u
r
is
t
ic
alg
o
r
i
th
m
th
a
t
w
as
i
n
s
p
ir
ed
b
y
t
h
e
b
eh
av
io
r
o
f
f
las
h
i
n
g
li
g
h
ts
o
f
r
ea
l
f
ir
ef
lies
.
T
h
e
alg
o
r
ith
m
p
er
f
o
r
m
an
ce
is
b
ased
o
n
t
h
e
r
ea
l
b
eh
a
v
io
r
o
f
f
ir
ef
lies
th
at
r
elies
o
n
th
e
attr
ac
tio
n
b
et
w
ee
n
a
f
ir
ef
l
y
a
n
d
an
o
th
er
o
n
b
asis
o
f
th
eir
b
r
ig
h
tn
e
s
s
.
Fo
r
m
u
lat
in
g
t
h
e
r
ea
l
f
ir
e
f
l
y
b
eh
av
io
r
in
to
an
al
g
o
r
ith
m
m
u
s
t
f
o
llo
w
t
h
r
ee
r
u
les
w
h
ic
h
g
o
v
er
n
h
o
w
t
h
e
r
ea
l
f
ir
e
f
lie
s
ac
t
in
r
ea
l
s
p
ac
e.
T
h
ese
r
u
les ar
e
as f
o
llo
w
s
:
a.
T
h
e
f
ir
ef
l
y
is
a
u
n
i
s
ex
.
So
,
all
th
e
f
ir
ef
lie
s
w
ill b
e
attr
ac
ted
t
o
ea
ch
o
th
er
r
eg
ar
d
less
o
f
t
h
ei
r
s
ex
.
b.
A
ttra
ct
iv
e
n
es
s
is
p
r
o
p
o
r
tio
n
al
to
b
r
ig
h
tn
es
s
.
T
h
er
ef
o
r
e,
f
o
r
an
y
f
las
h
lig
h
ti
n
g
b
et
w
ee
n
t
w
o
f
ir
e
f
lie
s
,
th
e
less
b
r
i
g
h
t
o
n
e
w
ill
m
o
v
e
to
th
e
b
r
ig
h
ter
o
n
e.
T
h
e
at
tr
ac
tiv
e
n
ess
d
ec
r
ea
s
es
a
s
t
h
e
d
is
tan
ce
i
n
cr
ea
s
e
s
b
et
w
ee
n
t
w
o
f
ir
ef
lie
s
.
T
h
e
f
ir
ef
lies
w
ill
m
o
v
e
r
a
n
d
o
m
l
y
i
n
ca
s
e
th
er
e
i
s
n
o
t
a
f
ir
ef
l
y
t
h
at
is
b
r
ig
h
ter
th
a
n
th
e
o
th
er
.
c.
Fire
f
l
y
b
r
i
g
h
t
n
e
s
s
is
i
n
f
lu
e
n
ce
d
o
r
d
eter
m
in
ed
b
y
th
e
lan
d
s
ca
p
e
o
f
th
e
f
i
tn
e
s
s
f
u
n
ctio
n
.
I
n
t
h
e
ma
x
i
m
izatio
n
p
r
o
b
le
m
,
b
r
ig
h
t
n
es
s
ca
n
s
i
m
p
l
y
b
e
p
r
o
p
o
r
tio
n
al
to
th
e
v
al
u
e
o
f
t
h
e
f
i
tn
e
s
s
f
u
n
ctio
n
[
23
].
T
h
e
f
ir
ef
l
y
a
lg
o
r
it
h
m
r
elies
o
n
t
w
o
i
m
p
o
r
tan
t
f
ac
to
r
s
:
t
h
e
lig
h
t
in
te
n
s
it
y
a
n
d
th
e
attr
ac
tiv
e
n
ess
b
et
w
ee
n
f
ir
e
f
lies
[
24
]
.
L
ig
h
t
in
ten
s
it
y
v
ar
ies
i
n
ea
ch
s
o
u
r
ce
ac
co
r
d
in
g
to
th
e
b
r
ig
h
t
n
ess
o
f
th
e
f
ir
e
f
l
y
,
w
h
ic
h
is
r
ep
r
esen
ted
an
d
ca
lc
u
lated
w
i
t
h
a
k
i
n
d
o
f
f
itn
e
s
s
f
u
n
ctio
n
.
B
r
ig
h
t
n
es
s
t
h
at
r
elies
o
n
li
g
h
t
i
n
ten
s
it
y
d
eter
m
in
e
s
attr
ac
tiv
e
n
es
s
.
Attr
ac
tiv
e
n
es
s
o
f
ea
ch
f
ir
ef
l
y
is
ca
lc
u
lated
u
s
in
g
t
h
e
f
o
llo
w
in
g
E
q
u
a
tio
n
(
1
)
[
2
4
]
.
(
r
)
=
(
1
)
W
h
er
e
β
0
r
ep
r
esen
t
s
t
h
e
attr
a
ctiv
e
n
ess
at
d
is
ta
n
ce
(
r
)
=0
an
d
s
o
m
e
ti
m
es
f
o
r
m
at
h
e
m
at
ical
co
m
p
u
ta
tio
n
i
s
co
n
s
id
er
ed
as
1
.
γ
s
y
m
b
o
l
r
ep
r
esen
ts
h
o
w
m
u
c
h
t
h
e
lig
h
t
ab
s
o
r
p
tio
n
is
.
r
is
th
e
d
is
ta
n
ce
b
et
w
ee
n
a
n
y
t
w
o
f
ir
ef
lies
i
a
n
d
j
at
d
if
f
er
e
n
t
p
o
s
itio
n
s
.
Fire
f
l
ies
ar
e
a
l
w
a
y
s
in
m
o
v
i
n
g
s
tat
u
s
f
r
o
m
p
o
s
itio
n
to
p
o
s
itio
n
.
A
cc
o
r
d
in
g
to
t
h
e
f
ac
t
o
f
attr
a
ctiv
e
n
ess
b
et
w
ee
n
f
ir
ef
lie
s
i
s
r
elate
d
to
th
e
d
is
tan
ce
b
et
w
ee
n
t
h
e
m
.
He
n
ce
,
th
e
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:
2
0
8
8
-
8708
F
ea
tu
r
e
S
elec
tio
n
A
p
p
r
o
a
ch
b
a
s
ed
o
n
F
ir
efly
A
lg
o
r
ith
m
a
n
d
C
h
i
-
s
q
u
a
r
e
(
E
ma
d
Mo
h
a
med
Ma
s
h
h
o
u
r
)
2341
d
is
tan
ce
b
et
w
ee
n
an
y
t
w
o
f
ir
ef
lies
i
an
d
j
i
s
co
m
p
u
ted
t
h
r
o
u
g
h
a
w
ell
-
k
n
o
w
n
d
is
ta
n
ce
la
w
ca
lled
E
u
clid
ea
n
,
w
h
ic
h
is
ca
lc
u
lated
as
f
o
llo
w
s
[
24
]:
√
∑
(
2
)
W
h
er
e
d
r
ep
r
esen
ts
t
h
e
d
i
m
e
n
s
io
n
alit
y
o
f
t
h
e
p
r
o
b
le
m
,
an
d
x
i,
k
is
th
e
k
th
co
m
p
o
n
e
n
t
o
f
t
h
e
p
o
s
itio
n
o
f
f
ir
e
f
l
y
i
.
Af
ter
ca
lc
u
lati
n
g
t
h
e
d
is
ta
n
ce
b
et
w
ee
n
th
e
t
w
o
f
ir
e
f
lies
,
s
u
p
p
o
s
e
th
e
f
ir
e
f
l
y
i
is
le
s
s
b
r
ig
h
t
n
es
s
th
a
n
f
ir
ef
l
y
j
,
s
o
th
e
attr
ac
ti
v
en
e
s
s
b
et
w
ee
n
t
h
e
m
o
cc
u
r
s
w
h
ile
m
o
v
in
g
t
h
e
f
ir
e
f
l
y
i
to
t
h
e
f
ir
e
f
l
y
j.
T
h
e
f
o
llo
w
i
n
g
E
q
u
atio
n
(
3
)
[
2
4
]
co
n
tr
o
ls
th
is
k
in
d
o
f
m
o
v
e
m
e
n
t a
n
d
it i
s
r
ep
r
esen
ted
as f
o
llo
w
s
:
(
)
(
3
)
w
h
er
e
t
r
ep
r
esen
ts
th
e
n
u
m
b
er
o
f
iter
atio
n
s
,
a
n
d
th
e
co
e
f
f
ici
en
t
α
r
ep
r
esen
ts
a
r
a
n
d
o
m
n
u
m
b
er
co
n
tr
o
llin
g
t
h
e
s
ize
o
f
t
h
e
r
an
d
o
m
w
al
k
,
an
d
r
a
n
d
r
ep
r
esen
ts
a
r
an
d
o
m
n
u
m
b
er
g
en
e
r
ato
r
w
h
ich
f
all
s
b
et
w
ee
n
[
0
,
1
]
.
T
h
e
f
ir
ef
l
y
w
it
h
lo
w
b
r
i
g
h
tn
e
s
s
m
o
v
es
to
th
e
h
ig
h
er
o
n
e
a
f
ter
c
o
n
s
id
er
in
g
t
h
r
ee
ter
m
s
[
24
]
.
T
h
e
f
ir
s
t
ter
m
is
t
h
e
cu
r
r
en
t
p
o
s
itio
n
o
f
t
h
e
lo
w
b
r
ig
h
tn
e
s
s
f
ir
e
f
l
y
.
Seco
n
d
ter
m
is
th
e
m
o
v
e
m
e
n
t
to
w
ar
d
th
e
f
ir
ef
l
y
w
it
h
h
i
g
h
er
br
ig
h
t
n
es
s
b
y
th
e
attr
ac
tio
n
co
ef
f
icie
n
t
β.
F
in
al
l
y
,
t
h
e
last
ter
m
is
a
k
i
n
d
o
f
r
an
d
o
m
w
a
lk
ca
lcu
la
ted
b
y
a
r
an
d
o
m
g
e
n
er
ato
r
m
u
ltip
lied
b
y
α.
3.
P
RO
P
O
S
E
D
M
E
T
H
O
DS
3
.
1
.
P
ro
po
s
ed
f
ra
m
ew
o
rk
o
f
f
iref
ly
ba
s
ed
f
ea
t
ure
s
elec
t
io
n
I
n
th
i
s
r
esear
ch
,
a
f
r
a
m
e
w
o
r
k
h
as
b
ee
n
d
ev
elo
p
ed
to
s
elec
t
th
e
m
o
s
t
i
n
f
o
r
m
ati
v
e
s
u
b
s
et
o
f
f
ea
t
u
r
es
f
r
o
m
d
i
f
f
er
en
t
d
ata
s
ets
b
ase
d
o
n
f
ir
e
f
l
y
al
g
o
r
ith
m
.
T
h
e
u
s
ed
f
ir
e
f
l
y
alg
o
r
it
h
m
h
as
b
ee
n
m
o
d
i
f
ied
an
d
co
m
b
i
n
ed
w
it
h
d
i
f
f
er
e
n
t
tec
h
n
iq
u
e
s
to
i
m
p
r
o
v
e
f
ea
tu
r
e
s
el
ec
tio
n
p
r
o
ce
s
s
w
h
ic
h
i
n
t
u
r
n
m
a
y
ac
h
ie
v
e
h
i
g
h
e
s
t
class
i
f
icatio
n
ac
c
u
r
ac
y
a
s
p
o
s
s
ib
le.
T
h
e
m
o
d
i
f
ied
f
ir
ef
l
y
alg
o
r
ith
m
i
s
b
ased
o
n
an
as
s
u
m
p
tio
n
th
at
ea
c
h
d
ataset
co
n
tai
n
s
a
n
u
m
b
er
o
f
f
ea
tu
r
es
n
an
d
a
n
u
m
b
er
o
f
s
a
m
p
les
d.
E
ac
h
f
ea
tu
r
e
i
is
a
v
e
cto
r
o
f
v
alu
e
s
(
V
i
,
k
)
,
w
h
er
e
(
k=1
,
2
,
3
,
….
d
)
f
o
r
d
if
f
e
r
en
t
s
a
m
p
le
s
S
=(
s
1
,s
2
,s
3
,…..,s
d,
)
,
(
i=
1
,
2
,
3
,
…,
n
)
fo
r
d
iffer
en
t
fea
tu
r
es
.
Mo
d
elin
g
f
ea
t
u
r
es
to
f
ir
ef
lie
s
is
r
ep
r
esen
ted
b
y
cr
ea
tin
g
n
f
ir
ef
lies
f
1
,
f
2
,
f
3
,
f
4
…f
n
.
Fo
r
e
ac
h
cr
ea
ted
f
ir
ef
l
y
(f
i,
)
,
(
i=1
,
2
,
3
,
….
,
n
)
,
a
v
ec
to
r
(
x
i,
k
)
o
f
ch
i
-
s
q
u
ar
e
v
al
u
es
is
ca
l
cu
lated
a
s
a
m
ap
p
in
g
v
ec
to
r
t
o
th
e
co
r
r
esp
o
n
d
i
n
g
v
ec
to
r
(
V
i
,
k
)
in
t
h
e
o
r
i
g
in
a
l
d
ataset
to
r
ep
r
esen
t
f
ir
ef
l
y
p
o
s
itio
n
,
w
h
er
e
(
k=1
,
2
,
3
,
….
d
)
to
r
ep
r
esen
t
a
s
et
o
f
d
if
f
er
e
n
t
p
o
s
itio
n
s
f
o
r
a
f
ir
e
f
l
y
/f
ea
t
u
r
e
in
d
i
f
f
er
e
n
t
s
a
m
p
les.
T
h
is
r
esear
ch
ai
m
s
to
ap
p
ly
t
h
e
f
ir
ef
l
y
f
r
a
m
e
w
o
r
k
o
n
m
icr
o
ar
r
ay
s
d
atase
ts
a
n
d
o
th
er
k
i
n
d
o
f
d
ataset
s
b
y
s
i
m
u
lati
n
g
e
x
is
ti
n
g
f
ea
t
u
r
es
a
s
a
n
u
m
b
er
o
f
f
ir
ef
l
ies,
ea
ch
f
ir
ef
l
y
(
f
ea
t
u
r
e)
h
as
its
o
w
n
p
o
s
itio
n
an
d
i
n
te
n
s
it
y
.
D
y
n
a
m
ic
p
ar
a
m
eter
s
s
u
ch
a
s
γ
,
α
,
β,
n
u
m
b
er
o
f
iter
atio
n
s
a
n
d
p
o
p
u
latio
n
s
ize
(
n
p
o
p
)
h
av
e
b
ee
n
d
eter
m
i
n
e
d
b
y
d
i
f
f
er
en
t
e
x
p
er
i
m
e
n
t
s
to
ac
h
ie
v
e
t
h
e
h
ig
h
es
t
p
er
f
o
r
m
a
n
ce
f
o
r
f
ea
t
u
r
e
s
elec
tio
n
an
d
cla
s
s
i
f
icatio
n
.
T
h
e
p
r
o
p
o
s
ed
f
r
am
e
w
o
r
k
i
n
t
h
is
r
es
ea
r
ch
co
n
tai
n
s
s
i
x
p
h
ases
w
h
ic
h
ar
e
as
f
o
llo
w
(
1
)
p
r
e
-
p
r
o
ce
s
s
in
g
p
h
ase
i
s
r
e
s
p
o
n
s
ib
le
f
o
r
d
ataset
f
iltra
tio
n
f
r
o
m
n
o
is
y
d
ata,
se
ar
ch
i
n
g
f
o
r
m
is
s
ed
v
al
u
es
in
d
atasets
a
n
d
f
illi
n
g
it
w
i
th
r
el
iab
le
v
al
u
es;
(
2
)
r
an
k
i
n
g
p
h
as
e
is
r
esp
o
n
s
ib
le
f
o
r
s
o
r
tin
g
th
e
o
r
ig
i
n
al
d
ataset
in
d
escen
d
i
n
g
o
r
d
er
ac
co
r
d
i
n
g
to
its
e
v
alu
a
tio
n
v
a
lu
e;
(
3
)
f
ir
ef
l
y
p
o
s
itio
n
ca
lcu
latio
n
p
h
a
s
e
i
s
f
o
r
d
eter
m
in
in
g
f
i
r
e
f
l
y
p
o
s
itio
n
v
a
l
u
es
f
o
r
d
i
f
f
er
e
n
t
f
ir
e
f
lies
;
(
4
)
f
ir
ef
l
y
in
te
n
s
it
y
ca
lcu
latio
n
p
h
a
s
e
i
s
r
esp
o
n
s
i
b
le
o
f
ca
lc
u
lati
n
g
i
n
te
n
s
it
y
v
alu
es
f
o
r
d
i
f
f
er
en
t
f
ir
e
f
lie
s
;
(
5
)
f
ir
ef
l
y
p
r
o
ce
s
s
in
g
p
h
ase
i
s
u
s
ed
f
o
r
s
elec
ti
n
g
th
e
h
i
g
h
e
s
t
i
n
f
o
r
m
ati
v
e
f
ea
t
u
r
es
f
r
o
m
t
h
e
r
an
k
ed
f
ea
tu
r
e
s
t
h
r
o
u
g
h
ap
p
l
y
i
n
g
th
e
m
o
d
i
f
ied
f
ir
ef
l
y
alg
o
r
it
h
m
,
an
d
f
in
a
ll
y
(
6
)
class
i
f
icatio
n
p
h
a
s
e
f
o
r
e
v
alu
a
tin
g
t
h
e
ab
ili
t
y
o
f
s
elec
ted
f
ea
tu
r
e
s
i
n
class
i
f
icatio
n
.
I
f
th
e
clas
s
i
f
ica
tio
n
ac
cu
r
ac
y
i
s
ac
ce
p
tab
le,
th
en
a
s
et
o
f
f
ea
t
u
r
es
ar
e
s
u
i
tab
le
to
class
i
f
y
f
u
t
u
r
e
u
n
s
ee
n
d
ata,
o
t
h
er
w
is
e
t
h
e
p
r
o
ce
s
s
is
r
ep
ea
ted
w
ith
o
t
h
er
c
r
iter
ia
s
u
c
h
as
o
t
h
er
r
an
k
in
g
m
et
h
o
d
s
,
o
r
f
itn
e
s
s
f
u
n
ctio
n
to
i
m
p
r
o
v
e
t
h
e
ac
c
u
r
ac
y
.
T
h
e
p
r
o
ce
s
s
co
n
t
in
u
es
u
n
til
r
ea
c
h
i
n
g
th
e
cr
iter
ia
t
h
at
ac
h
iev
e
t
h
e
h
ig
h
es
t
p
o
s
s
ib
le
ac
cu
r
ac
y
.
T
h
e
f
o
llo
w
i
n
g
s
ec
tio
n
s
d
is
cu
s
s
d
i
f
f
er
e
n
t
f
r
a
m
e
w
o
r
k
p
h
a
s
es
:
3
.
1
.
1
.
P
re
-
pro
ce
s
s
ing
ph
a
s
e
Hu
g
e
d
atasets
o
f
te
n
s
u
f
f
er
f
r
o
m
n
o
is
y
a
n
d
m
i
s
s
ed
d
ata
v
al
u
es
t
h
at
m
a
y
af
f
ec
t
a
n
y
class
i
f
ier
n
eg
at
iv
el
y
.
I
n
t
h
i
s
r
esear
ch
t
h
e
u
s
ed
d
atasets
s
u
f
f
er
f
r
o
m
m
i
s
s
ed
v
al
u
es
;
th
i
s
p
r
o
b
le
m
m
a
y
lead
an
y
clas
s
i
f
ier
to
u
n
r
eliab
le
r
esu
lt
s
.
I
n
th
is
p
h
ase,
ea
ch
f
ea
tu
r
e
h
a
s
b
ee
n
s
ca
n
n
ed
f
o
r
d
if
f
er
en
t
d
atasets
s
e
ar
ch
in
g
f
o
r
m
i
s
s
ed
v
alu
e
s
; a
n
d
f
illi
n
g
it b
y
co
n
s
id
er
in
g
t
h
e
av
er
a
g
e
v
al
u
e
o
f
t
h
e
w
h
o
le
f
ea
t
u
r
e
v
ec
to
r
.
3
.
1
.
2
.
Ra
n
k
ing
p
ha
s
e
I
n
th
i
s
p
h
a
s
e,
d
if
f
er
en
t
s
tati
s
tical
ap
p
r
o
ac
h
es
h
a
v
e
b
ee
n
u
s
ed
to
r
an
k
t
h
e
f
ea
t
u
r
es.
A
v
alu
e
i
s
ca
lcu
lated
f
o
r
ea
ch
f
ea
t
u
r
e
ac
co
r
d
in
g
to
a
s
p
ec
if
ic
cr
iter
io
n
to
r
an
k
t
h
e
m
.
I
n
th
i
s
r
esear
c
h
,
T
-
test
an
d
r
elie
f
f
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2088
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
4
,
A
u
g
u
s
t 2
0
1
8
:
2
3
3
8
–
2
3
5
0
2342
tech
n
iq
u
es
w
er
e
u
s
ed
to
r
an
k
th
e
p
r
e
-
p
r
o
ce
s
s
ed
d
ataset
to
p
ick
th
e
h
i
g
h
e
s
t
r
an
k
ed
f
ea
t
u
r
e
s
f
ir
s
t.
T
h
e
r
an
k
ed
f
ea
t
u
r
es
b
y
th
e
t
w
o
d
if
f
er
e
n
t
t
ec
h
n
iq
u
es
ar
e
in
tr
o
d
u
ce
d
to
t
h
e
s
u
b
s
eq
u
en
t
p
h
a
s
es.
I
n
ca
s
e
o
f
h
u
g
e
d
atase
t
s
u
c
h
as
m
icr
o
ar
r
a
y
s
,
t
h
e
h
i
g
h
e
s
t
r
a
n
k
ed
f
ea
t
u
r
es
w
h
ich
ar
e
t
h
e
m
o
s
t
in
f
o
r
m
ati
v
e
f
ea
t
u
r
es
ar
e
s
e
lect
ed
as
ca
n
d
id
ate
f
o
r
f
ir
ef
l
y
p
r
o
ce
s
s
i
n
g
.
3
.
1
.
2
.
1
.
T
-
t
est
m
et
ho
d
T
-
test
is
co
n
s
id
er
ed
as
a
well
-
k
n
o
w
n
r
a
n
k
i
n
g
f
ea
t
u
r
e
m
et
h
o
d
.
T
-
test
is
u
s
ed
to
m
ea
s
u
r
e
t
h
e
d
if
f
er
e
n
ce
b
et
w
ee
n
t
w
o
Ga
u
s
s
ian
d
is
tr
ib
u
tio
n
s
.
T
h
e
s
tan
d
ar
d
T
-
test
is
u
s
ed
to
r
an
k
d
ata
s
ets
w
it
h
t
w
o
cla
s
s
es
;
in
t
h
e
ca
s
e
o
f
t
h
is
r
esear
c
h
t
h
e
d
atasets
m
a
y
v
ar
y
to
b
e
m
u
lti
class
d
atase
ts
.
A
m
o
d
if
icat
io
n
w
as
d
o
n
e
b
y
[
2
5
]
in
o
r
d
er
to
ca
lc
u
late
th
e
d
if
f
er
en
ce
b
et
w
ee
n
o
n
e
clas
s
a
n
d
t
h
e
ce
n
ter
o
f
all
clas
s
es.
C
alc
u
latio
n
s
ar
e
f
o
r
m
u
lated
th
r
o
u
g
h
E
q
u
atio
n
s
(
4
)
-
(
8
)
[
2
5
]
.
TS
i
=ma
x
{
|
̅
̅
|
}
(
4
)
W
h
er
e
̅
∑
̅
⁄
(
5
)
̅
∑
⁄
(
6
)
∑
∑
(
̅
)
(
7
)
√
⁄
⁄
(
8
)
3
.
1
.
2
.
2
.
RE
L
I
E
F
F
m
et
ho
d
R
elie
f
f
i
s
a
k
in
d
o
f
m
eth
o
d
w
h
ic
h
ca
n
b
e
u
s
ed
f
o
r
f
ea
t
u
r
e
r
an
k
i
n
g
.
I
t
ev
al
u
ates
ea
c
h
f
ea
tu
r
e
an
d
ass
i
g
n
s
a
k
i
n
d
o
f
w
ei
g
h
t
f
o
r
e
ac
h
f
ea
t
u
r
e
.
T
h
is
w
ei
g
h
t
i
s
as
s
ig
n
ed
ac
co
r
d
in
g
to
th
e
ca
p
ab
ilit
y
f
o
r
th
i
s
f
ea
t
u
r
e
to
d
is
ti
n
g
u
is
h
b
et
w
ee
n
clas
s
es
.
R
E
L
I
E
FF
tech
n
iq
u
e
i
s
u
s
ed
f
o
r
b
in
ar
y
a
n
d
m
u
lticlas
s
p
r
o
b
le
m
s
.
T
h
is
m
et
h
o
d
is
m
o
r
e
r
o
b
u
s
t a
n
d
ca
n
d
ea
l
w
i
th
in
co
m
p
lete
an
d
n
o
is
y
d
ata
[
2
6
].
3
.
1
.
3
.
F
ire
f
ly
po
s
it
io
n c
a
lcula
t
io
n pha
s
e
T
h
e
ch
an
g
i
n
g
o
f
p
o
s
itio
n
i
n
f
ir
ef
lies
r
elie
s
o
n
th
e
i
n
te
n
s
it
y
an
d
th
e
m
o
v
e
m
e
n
t
o
f
t
h
e
lo
w
i
n
te
n
s
it
y
f
ir
ef
l
y
to
t
h
e
h
i
g
h
e
s
t
i
n
te
n
s
it
y
o
n
e.
I
n
th
i
s
r
esear
c
h
,
th
e
f
i
r
ef
l
y
p
o
s
it
io
n
w
i
ll
b
e
r
ep
r
esen
ted
u
s
in
g
ce
l
l
ch
i
-
s
q
u
ar
e
[
2
7
]
.
I
n
th
is
p
h
ase,
d
eter
m
i
n
in
g
p
o
s
itio
n
s
f
o
r
d
if
f
e
r
en
t
f
ea
t
u
r
es
ƒ
i
(
w
h
er
e
i=1
,
2
…,
n
)
is
d
o
n
e
b
y
ca
lcu
lati
n
g
n
ce
ll
c
h
i
-
s
q
u
ar
e
v
ec
to
r
s
x
i.
E
ac
h
f
ea
t
u
r
e
v
ec
to
r
v
alu
e
V
i,
k
i
s
as
s
i
g
n
ed
a
r
elev
a
n
t
v
ec
to
r
v
alu
e
x
i,
k
o
b
tain
ed
b
y
ca
lc
u
lati
n
g
ch
i
-
s
q
u
ar
e
f
o
r
ea
ch
f
ea
t
u
r
e
v
al
u
e
i
n
v
ec
to
r
V
i,
k
.
T
h
e
cr
ea
ted
ch
i
-
s
q
u
ar
e
v
ec
to
r
s
ar
e
to
r
ep
r
esen
t
th
e
p
o
s
itio
n
v
al
u
es
o
f
th
e
f
ir
ef
l
ies.
T
h
is
is
d
o
n
e
b
y
m
ea
s
u
r
i
n
g
ea
ch
tab
le
ce
ll
an
d
test
s
w
h
et
h
er
it
i
s
d
if
f
er
e
n
t f
r
o
m
its
e
x
p
ec
ted
v
al
u
e
th
r
o
u
g
h
o
u
t t
h
e
w
h
o
le
d
atas
et
u
s
i
n
g
E
q
u
atio
n
(
9
)
[
2
7
].
x
2
=
∑
(
(V
i
-
E
i
)
2
/E
i
)
(
9
)
W
h
er
e
(
Vi)
is
t
h
e
o
b
s
er
v
ed
v
a
lu
e
i
n
t
h
e
f
ea
t
u
r
e
v
ec
to
r
,
an
d
(
E
)
s
tan
d
s
f
o
r
t
h
e
e
x
p
ec
ted
v
al
u
e
f
o
r
ea
ch
ce
l
l
o
r
v
alu
e
i
n
t
h
e
f
ea
tu
r
e
v
ec
to
r
.
3
.
1
.
4
.
F
ire
f
ly
inte
ns
it
y
ca
lcula
t
io
n pha
s
e
I
n
th
is
p
h
ase,
ea
c
h
f
ir
ef
l
y
(
f
i
)
is
as
s
ig
n
ed
a
li
g
h
t
i
n
te
n
s
i
t
y
v
a
lu
e
(
L
i
)
ca
lcu
lated
b
y
a
f
it
n
es
s
f
u
n
ct
io
n
.
I
n
ten
s
it
y
is
u
s
ed
to
co
m
p
ar
e
b
et
w
ee
n
f
ir
ef
l
ies
i
n
o
r
d
er
to
d
e
cid
e
w
h
ich
h
a
v
e
th
e
lo
w
er
i
n
t
en
s
it
y
to
m
o
v
e
w
i
t
h
a
co
n
tr
o
lled
m
o
v
e
m
en
t
u
s
i
n
g
E
q
u
atio
n
(
3
)
.
T
h
e
f
ir
ef
l
y
w
it
h
lo
w
er
i
n
ten
s
it
y
u
p
d
ates
i
t’
s
i
n
ten
s
it
y
a
f
ter
ea
ch
m
o
v
e
m
e
n
t
t
h
r
o
u
g
h
a
s
e
t
o
f
it
er
atio
n
s
.
I
n
t
h
is
r
e
s
ea
r
ch
s
ev
en
d
i
f
f
er
e
n
t
f
it
n
es
s
f
u
n
ctio
n
s
h
av
e
b
ee
n
tr
ied
to
r
ep
r
esen
t
i
n
te
n
s
it
y
,
t
h
e
g
o
al
o
f
u
til
izin
g
m
o
r
e
t
h
a
n
f
it
n
es
s
f
u
n
ct
io
n
is
to
s
ea
r
c
h
f
o
r
t
h
e
b
est
f
it
n
es
s
f
u
n
ct
io
n
th
at
ca
n
s
i
m
u
late
f
ir
ef
l
y
in
te
n
s
it
y
to
h
elp
in
s
elec
ti
n
g
th
e
m
o
s
t
in
f
o
r
m
ati
v
e
f
ea
t
u
r
es
t
h
at
ca
n
b
e
u
s
ed
to
m
i
n
i
m
ize
clas
s
i
f
icatio
n
s
er
r
o
r
s
,
T
a
b
le
1
r
ev
ie
w
s
t
h
e
u
s
ed
f
it
n
ess
f
u
n
ctio
n
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:
2
0
8
8
-
8708
F
ea
tu
r
e
S
elec
tio
n
A
p
p
r
o
a
ch
b
a
s
ed
o
n
F
ir
efly
A
lg
o
r
ith
m
a
n
d
C
h
i
-
s
q
u
a
r
e
(
E
ma
d
Mo
h
a
med
Ma
s
h
h
o
u
r
)
2343
T
ab
le
1
.
Dif
f
er
en
t
F
itn
e
s
s
F
u
n
ctio
n
s
u
s
ed
f
o
r
Si
m
u
lati
n
g
F
ir
ef
l
y
I
n
te
n
s
it
y
Fi
t
n
e
ss F
u
n
c
t
i
o
n
Eq
u
a
t
i
o
n
R
o
se
n
b
r
o
c
k
[
4
]
ƒ
R
o
s
e
n
b
r
o
c
k
(x
1
,
….
.
,
x
n
)
=
∑
(
)
(
1
0
)
A
c
k
l
e
y
[
4
]
ƒ(
〈
〉
)
=
2
0
+
e
-
2
0
e
x
p
(
√
∑
)
(
∑
)
(
1
1
)
S
p
h
e
r
e
[
4
]
S
p
h
e
r
e
(
x
1
,
….
.
,
x
n
)
=
∑
(
1
2
)
R
a
st
r
i
g
i
n
[
5
]
ƒ
(
x
)
=
1
0
n
+
∑
(
1
3
)
S
a
l
o
mo
n
[
5
]
ƒ
(
x
)
=
-
c
o
s
(
√
∑
)
√
∑
+1
(
1
4
)
S
c
h
w
e
f
e
l
[
5
]
ƒ
(
x
)
=
4
1
8
.
9
8
2
9
d
-
∑
√
|
|
(
1
5
)
X
i
n
-
S
h
e
Y
a
n
g
[
5
]
ƒ
(
x
)
=
-
∑
|
|
∑
(
1
6
)
3
.
1
.
5
.
F
iref
ly
pro
ce
s
s
ing
ph
a
s
e
I
n
th
i
s
p
h
ase,
t
h
e
f
ir
e
f
l
y
alg
o
r
ith
m
is
ap
p
lied
as
a
f
ea
t
u
r
e
s
e
lectio
n
p
r
o
ce
s
s
.
T
h
e
p
r
o
ce
s
s
b
eg
in
s
w
it
h
co
m
p
ar
i
n
g
t
w
o
r
a
n
d
o
m
f
ir
e
f
li
es
in
te
n
s
it
y
w
i
th
ea
c
h
o
th
er
,
t
h
e
o
n
e
w
it
h
lo
w
er
lig
h
t
i
n
te
n
s
it
y
w
ill
m
o
v
e
to
th
e
h
ig
h
er
f
ir
e
f
l
y
,
d
is
ta
n
ce
(
r
)
b
et
w
ee
n
th
e
m
w
ill
b
e
ca
lcu
la
ted
u
s
i
n
g
E
q
u
a
tio
n
(
2
)
,
th
e
attr
ac
tio
n
v
al
u
e
i
s
ca
lcu
lated
u
s
i
n
g
E
q
u
atio
n
(
1
)
,
th
e
n
e
w
p
o
s
itio
n
(
x
i
)
f
o
r
th
e
l
o
w
er
f
ir
ef
l
y
i
s
ca
lc
u
lated
u
s
i
n
g
E
q
u
atio
n
(
3
)
,
an
d
f
i
n
all
y
n
e
w
i
n
ten
s
it
y
w
i
ll
b
e
u
p
d
ated
th
r
o
u
g
h
ca
lcu
la
tin
g
f
itn
es
s
f
u
n
ctio
n
,
t
h
is
tas
k
r
elie
s
o
n
t
w
o
i
m
p
o
r
tan
t
f
ac
to
r
s
:
a.
Fire
f
l
y
in
ten
s
it
y
,
th
e
li
g
h
t in
te
n
s
it
y
p
r
o
d
u
ce
d
f
r
o
m
ea
ch
f
ir
e
f
l
y
in
s
p
ac
e.
b.
Fire
f
l
y
p
o
s
itio
n
,
t
h
e
p
o
s
itio
n
o
f
th
e
f
ir
e
f
l
y
i
n
s
p
ac
e,
it
k
ee
p
s
ch
an
g
i
n
g
ac
co
r
d
in
g
to
s
o
m
e
f
a
cto
r
s
T
h
is
p
r
o
ce
s
s
co
n
tin
u
e
s
f
o
r
a
n
u
m
b
er
o
f
iter
atio
n
s
o
r
g
e
n
er
atio
n
s
s
p
ec
if
ied
b
y
th
e
u
s
er
.
I
n
th
e
s
e
iter
atio
n
s
,
t
h
e
f
ir
ef
l
y
ar
e
al
w
a
y
s
i
n
m
o
v
i
n
g
s
tat
u
s
f
r
o
m
p
o
s
itio
n
to
p
o
s
itio
n
w
h
er
e
th
e
lo
w
er
l
i
g
h
t
in
ten
s
it
y
w
il
l
m
o
v
e
to
t
h
e
h
i
g
h
er
f
ir
e
f
l
y
.
Af
ter
t
h
at
th
e
h
ig
h
es
t
r
an
k
ed
f
ea
t
u
r
es
ar
e
i
n
tr
o
d
u
ce
d
to
th
e
clas
s
i
f
ier
in
cr
e
m
e
n
tall
y
s
tar
tin
g
f
r
o
m
th
e
h
ig
h
er
i
n
ten
s
it
y
(
m
o
s
t
i
n
f
o
r
m
ati
v
e)
u
n
til
r
ea
c
h
in
g
t
h
e
h
ig
h
est
p
o
s
s
ib
le
ac
cu
r
ac
y
.
3
.
1
.
6
.
Cla
s
s
if
ica
t
i
o
n pha
s
e
I
n
th
e
class
if
ica
tio
n
p
h
ase
t
h
e
f
ir
ef
lies
(
f
ea
t
u
r
es)
p
r
o
d
u
ce
d
f
r
o
m
t
h
e
p
r
ev
io
u
s
p
h
a
s
e
ar
e
ex
p
o
s
ed
to
a
class
i
f
ier
.
Di
f
f
er
en
t
m
ac
h
i
n
e
lear
n
i
n
g
tec
h
n
iq
u
es
ca
n
b
e
u
s
ed
a
s
cla
s
s
i
f
ier
s
;
i
n
t
h
i
s
r
esear
ch
K
-
n
ea
r
est
n
eig
h
b
o
r
(
KNN)
an
d
d
is
cr
i
m
i
n
an
t a
n
al
y
s
i
s
(
D
A
)
ar
e
u
s
ed
.
3
.
1
.
6
.
1
.
K
-
nea
re
s
t
neig
hb
o
r
cla
s
s
if
ie
r
K
-
n
ea
r
est
n
ei
g
h
b
o
r
(
K
-
NN)
ap
p
r
o
ac
h
is
co
n
s
id
er
ed
as
a
n
o
n
-
p
ar
a
m
etr
ic
lear
n
i
n
g
al
g
o
r
ith
m
.
No
n
p
ar
am
etr
ic
al
g
o
r
ith
m
m
ea
n
s
it
d
o
esn
’
t
n
ee
d
to
ass
u
m
e
a
n
y
d
ata
d
is
tr
ib
u
tio
n
[
6
]
.
T
h
e
K
-
NN
alg
o
r
ith
m
is
o
n
e
o
f
th
e
s
i
m
p
le
s
t
m
ac
h
i
n
e
lear
n
i
n
g
al
g
o
r
ith
m
s
a
n
d
it
is
co
n
s
id
er
ed
as
in
s
ta
n
ce
-
b
ased
lear
n
in
g
,
w
h
er
e
t
h
e
u
n
s
ee
n
d
ata
h
as
b
ee
n
c
lass
if
ied
b
ased
o
n
tr
ain
i
n
g
d
ataset
s
to
r
ed
b
ef
o
r
e.
T
h
e
alg
o
r
ith
m
r
elie
s
o
n
t
h
e
d
is
ta
n
ce
b
et
w
ee
n
th
e
tr
ai
n
i
n
g
d
ataset
a
n
d
t
h
e
u
n
s
ee
n
o
r
t
h
e
tes
tin
g
d
ata
s
et,
t
h
e
d
is
ta
n
ce
i
s
ca
lc
u
lated
b
y
a
k
i
n
d
o
f
s
i
m
ilar
it
y
m
ea
s
u
r
e,
s
u
c
h
as t
h
e
E
u
c
lid
ea
n
d
is
ta
n
ce
,
co
s
in
e
s
i
m
ilar
it
y
o
r
th
e
Ma
n
h
atta
n
d
is
ta
n
ce
.
3
.
1
.
6
.
2
.
Dis
cr
i
m
i
na
nt
a
na
ly
s
is
cla
s
s
if
ier
Dis
cr
i
m
in
a
n
t a
n
al
y
s
i
s
is
a
n
ap
p
r
o
ac
h
u
s
ed
f
o
r
cl
as
s
if
icatio
n
,
w
h
er
e
t
w
o
o
r
m
o
r
e
g
r
o
u
p
s
ar
e
k
n
o
w
n
a
s
a
p
r
io
r
i
an
d
o
n
e
o
r
m
o
r
e
n
e
w
o
b
s
er
v
atio
n
s
ar
e
cla
s
s
i
f
ied
in
t
o
o
n
e
o
f
th
e
k
n
o
w
n
g
r
o
u
p
s
b
a
s
ed
o
n
th
e
m
ea
s
u
r
ed
ch
ar
ac
ter
is
tic
s
.
I
t
i
s
u
s
ed
to
p
r
ed
ict
th
e
m
e
m
b
er
s
h
ip
o
f
a
s
a
m
p
le
to
a
g
r
o
u
p
b
ased
o
n
a
s
et
o
f
i
n
d
ep
en
d
en
t
v
ar
iab
les.
T
h
e
p
r
o
ce
s
s
o
f
d
is
cr
i
m
i
n
a
n
t
a
n
al
y
s
is
r
el
ies
o
n
co
m
p
u
tin
g
t
h
e
r
elatio
n
s
h
i
p
o
f
v
ar
iab
les
b
y
m
i
n
i
m
izi
n
g
d
i
s
tan
ce
t
h
e
w
it
h
i
n
class
d
is
ta
n
ce
an
d
m
a
x
i
m
izi
n
g
t
h
e
b
et
w
ee
n
clas
s
d
is
tan
ce
s
i
m
u
lta
n
eo
u
s
l
y
,
to
ea
r
n
th
e
h
ig
h
es
t c
lass
d
i
s
cr
i
m
i
n
atio
n
r
a
te
[
7
]
.
T
h
e
f
r
a
m
e
w
o
r
k
in
cl
u
d
i
n
g
d
if
f
er
en
t
p
h
ases
f
o
r
s
e
lecti
n
g
th
e
h
ig
h
e
s
t
b
est
i
n
f
o
r
m
ati
v
e
s
u
b
s
e
t
o
f
f
ea
t
u
r
es u
s
in
g
f
ir
ef
l
y
i
s
ill
u
s
tr
a
ted
in
Fig
u
r
e
1
.
Fig
u
r
e
2
s
h
o
w
s
p
s
e
u
d
o
co
d
e
t
h
at
in
teg
r
ate
s
th
e
d
i
f
f
er
e
n
t ste
p
s
f
o
r
s
elec
tin
g
f
ea
t
u
r
es a
n
d
f
i
n
d
in
g
th
e
b
est in
f
o
r
m
ati
v
e
f
ea
tu
r
es
s
u
b
s
et
u
s
i
n
g
th
e
p
r
o
p
o
s
ed
f
ir
ef
l
y
f
r
a
m
e
w
o
r
k
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2088
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
4
,
A
u
g
u
s
t 2
0
1
8
:
2
3
3
8
–
2
3
5
0
2344
Fig
u
r
e
1
.
T
h
e
p
r
o
p
o
s
ed
f
ir
ef
l
y
f
r
a
m
e
w
o
r
k
f
o
r
p
ick
i
n
g
i
n
f
o
r
m
ativ
e
f
ea
tu
r
e
s
Fig
u
r
e
2
.
Sh
o
w
s
p
s
e
u
d
o
co
d
e
1
o
f
th
e
m
o
d
if
ied
f
ir
e
f
l
y
al
g
o
r
i
th
m
4.
RE
SU
L
T
S AN
D
AN
AL
Y
SI
S
T
h
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
h
a
s
b
ee
n
test
ed
u
s
i
n
g
f
o
u
r
p
u
b
lic
b
en
ch
m
ar
k
d
atasets
.
T
h
e
d
es
cr
ip
tio
n
o
f
th
ese
d
ataset
s
i
s
s
h
o
w
n
i
n
T
ab
le
2
.
Data
s
et
s
u
s
ed
ar
e
S
R
B
C
T
m
icr
o
ar
r
ay
ca
n
ce
r
d
ata
s
et
an
d
it
h
as
b
ee
n
o
b
tain
ed
f
r
o
m
t
h
e
GE
MS
w
e
b
s
ite
(
w
ww
.
g
e
m
s
-
s
y
s
te
m
.
o
r
g
)
.
T
h
e
th
r
ee
o
th
er
d
atasets
(
L
u
n
g
,
Hep
atiti
s
,
an
d
Der
m
ato
lo
g
y
)
h
a
v
e
b
ee
n
o
b
tain
ed
f
r
o
m
t
h
e
U
n
iv
er
s
it
y
o
f
C
ali
f
o
r
n
ia
at
I
r
v
in
e
(
UC
I
)
m
ac
h
i
n
e
lear
n
in
g
r
ep
o
s
ito
r
y
[
28
]
.
Data
s
ets ar
er
an
d
o
m
l
y
d
i
v
id
ed
in
to
7
5
% f
o
r
t
r
ain
in
g
a
n
d
2
5
% f
o
r
test
in
g
.
Af
ter
ap
p
l
y
in
g
th
e
p
r
ep
r
o
ce
s
s
in
g
a
n
d
r
an
k
i
n
g
p
h
ase
o
n
th
e
in
p
u
t
d
ataset,
a
p
o
o
l
co
n
tain
in
g
a
s
et
o
f
f
ea
t
u
r
es
f
o
r
ea
ch
d
ataset
r
an
k
ed
b
y
t
-
test
o
r
r
elief
f
is
co
n
s
tr
u
cted
f
ir
s
t,
an
d
th
e
n
e
x
p
o
s
ed
to
th
e
m
o
d
i
f
ied
f
ir
ef
l
y
alg
o
r
it
h
m
e
x
p
er
i
m
e
n
te
d
w
i
th
d
if
f
er
en
t
f
itn
e
s
s
f
u
n
cti
o
n
s
.
T
h
e
o
u
tp
u
t
o
f
t
h
e
f
ir
ef
l
y
p
r
o
ce
s
s
in
g
p
h
a
s
e
is
th
e
h
ig
h
es
t
r
an
k
ed
f
ea
t
u
r
es
s
u
b
s
et.
T
h
ese
f
ea
t
u
r
es
ar
e
p
ass
e
d
f
ea
tu
r
e
b
y
f
ea
t
u
r
e
to
t
h
e
cl
ass
i
f
ier
i
n
o
r
d
er
to
ev
alu
a
te
f
ea
t
u
r
es.
C
las
s
i
f
icati
o
n
r
ate
is
m
o
n
ito
r
ed
u
n
til
t
h
e
class
if
icatio
n
ac
cu
r
ac
y
h
ad
b
ee
n
im
p
r
o
v
ed
as
In
p
u
t
:
M
a
t
r
i
x
M
(
d
,
n
)
,
M
i
s
t
h
e
o
r
i
g
i
n
a
l
d
a
t
a
se
t
m
a
t
r
i
x
,
w
h
e
r
e
d
n
u
mb
e
r
o
f
sam
p
l
e
s &
n
n
u
mb
e
r
o
f
f
e
a
t
u
r
e
s
Ou
t
p
u
t
:
M
a
t
r
i
x
S
(
d
,
r
)
,
S
i
s a
r
e
d
u
c
e
d
m
a
t
r
i
x
d
a
t
a
se
t
,
w
h
e
r
e
d
n
u
mb
e
r
o
f
s
a
mp
l
e
s &
r
r
e
d
u
c
e
d
n
u
m
b
e
r
o
f
f
e
a
t
u
r
e
s
S
t
e
p
1
:
R
a
n
k
f
e
a
t
u
r
e
s u
si
n
g
t
-
t
e
st
/
r
e
l
i
e
f
f
f
o
r
t
h
e
o
r
i
g
i
n
a
l
d
a
t
a
se
t
M
.
S
t
e
p
2
:
I
n
i
t
i
a
l
i
z
e
p
a
r
a
me
t
e
r
s:
N
u
mb
e
r
o
f
i
t
e
r
a
t
i
o
n
s
t
L
i
g
h
t
a
b
so
r
p
t
i
o
n
c
o
e
f
f
i
c
i
e
n
t
A
t
t
r
a
c
t
i
o
n
c
o
e
f
f
i
c
i
e
n
t
R
a
n
d
o
m
i
z
a
t
i
o
n
p
a
r
a
me
t
e
r
α
(
n
p
o
p
)
n
u
m
b
e
r
o
f
f
i
r
e
f
l
i
e
s c
o
n
si
d
e
r
e
d
i
n
sp
a
c
e
S
t
e
p
3
:
S
i
m
u
l
a
t
e
f
e
a
t
u
r
e
s a
s f
i
r
e
f
l
i
e
s,
w
h
e
r
e
e
a
c
h
f
i
r
e
f
l
y
i
s re
p
r
e
se
n
t
e
d
b
y
a
v
e
c
t
o
r
o
f
v
a
l
u
e
s (o
r
i
g
i
n
a
l
v
a
l
u
e
s fo
r
e
a
c
h
f
e
a
t
u
r
e
)
S
t
e
p
4
:
A
ssi
g
n
p
o
si
t
i
o
n
v
a
l
u
e
s fo
r
e
a
c
h
f
i
r
e
f
l
y
,
b
y
c
a
l
c
u
l
a
t
i
n
g
c
e
l
l
C
h
i
-
s
q
u
a
r
e
f
o
r
e
a
c
h
v
a
l
u
e
e
x
i
st
e
d
i
n
e
a
c
h
f
e
a
t
u
r
e
i
n
t
h
e
d
a
t
a
se
t
.
S
t
e
p
5
:
I
n
i
t
i
a
l
i
z
e
i
n
t
e
n
si
t
y
v
a
l
u
e
f
o
r
e
a
c
h
f
i
r
e
f
l
y
,
t
h
r
o
u
g
h
c
a
l
c
u
l
a
t
i
n
g
o
n
e
o
f
t
h
e
f
o
l
l
o
w
i
n
g
f
i
t
n
e
ss fu
n
c
t
i
o
n
s:
R
o
se
n
b
r
o
c
k
S
p
h
e
r
e
A
c
k
l
e
y
R
a
st
r
i
g
i
n
S
c
h
w
e
f
e
l
S
a
l
o
mo
n
X
i
n
-
S
h
e
y
a
n
g
I
n
i
t
i
a
l
i
z
e
d
i
st
a
n
c
e
r
=
0
.
(
D
i
st
a
n
c
e
b
e
t
w
e
e
n
f
i
r
e
f
l
i
e
s)
.
S
t
e
p
6
:
W
h
i
l
e
(
l
o
o
p
<
n
u
m
b
e
r
o
f
i
t
e
r
a
t
i
o
n
s)
I
f
i
n
t
e
n
si
t
y
(
f
i
r
e
f
l
y
i
)
<
i
n
t
e
n
si
t
y
(
f
i
r
e
f
l
y
j
)
C
a
l
c
u
l
a
t
e
d
i
st
a
n
c
e
(
r
)
b
e
t
w
e
e
n
t
h
e
m.
C
a
l
c
u
l
a
t
e
a
t
t
r
a
c
t
i
o
n
b
a
se
d
o
n
d
i
st
a
n
c
e
.
C
a
l
c
u
l
a
t
e
t
h
e
mo
v
e
me
n
t
o
f
t
h
e
l
o
w
i
n
t
e
n
si
t
y
f
i
r
e
f
l
y
u
si
n
g
.
C
a
l
c
u
l
a
t
e
n
e
w
c
h
i
-
s
q
u
a
r
e
v
a
l
u
e
s a
s
n
e
w
p
o
si
t
i
o
n
s
f
o
r
t
h
e
mo
v
e
d
f
i
r
e
f
l
y
.
A
f
t
e
r
c
h
a
n
g
i
n
g
p
o
si
t
i
o
n
,
i
n
t
e
n
s
i
t
y
i
s r
e
c
a
l
c
u
l
a
t
e
d
f
o
r
t
h
e
mo
v
e
d
f
i
r
e
f
l
y
u
si
n
g
f
i
t
n
e
ss f
u
n
c
t
i
o
n
.
R
a
n
k
f
i
r
e
f
l
i
e
s a
c
c
o
r
d
i
n
g
t
o
t
h
e
i
n
t
e
n
s
i
t
y
.
C
o
n
si
d
e
r
t
h
e
n
e
x
t
f
i
r
e
f
l
i
e
s t
o
c
o
mp
a
r
e
.
En
d
w
h
i
l
e
S
t
e
p
7
:
R
e
p
e
a
t
R
e
p
e
a
t
(
s
t
e
p
1
t
o
s
t
e
p
6
)
w
i
t
h
d
i
f
f
e
r
e
n
t
c
o
m
b
i
n
a
t
i
o
n
u
n
t
i
l
a
c
c
u
r
a
c
y
i
s
a
c
c
e
p
t
a
b
l
e
.
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:
2
0
8
8
-
8708
F
ea
tu
r
e
S
elec
tio
n
A
p
p
r
o
a
ch
b
a
s
ed
o
n
F
ir
efly
A
lg
o
r
ith
m
a
n
d
C
h
i
-
s
q
u
a
r
e
(
E
ma
d
Mo
h
a
med
Ma
s
h
h
o
u
r
)
2345
p
o
s
s
ib
le
w
it
h
t
h
e
m
o
s
t i
n
f
o
r
m
ativ
e
f
i
r
e
f
lies
(
f
ea
t
u
r
es).
T
w
o
r
an
k
i
n
g
m
eth
o
d
s
h
a
v
e
b
ee
n
te
s
ted
w
h
ic
h
ar
e
t
-
te
s
t
an
d
r
elief
f
,
s
e
v
en
d
i
f
f
er
e
n
t
f
it
n
es
s
f
u
n
ctio
n
s
h
a
v
e
b
ee
n
tr
ied
w
it
h
th
e
f
ir
e
f
l
y
p
r
o
ce
s
s
i
n
g
to
r
ep
r
esen
t
in
te
n
s
it
y
,
an
d
t
w
o
d
i
f
f
er
e
n
t
clas
s
i
f
ier
s
K
-
NN
a
n
d
d
is
cr
i
m
i
n
a
n
t
an
al
y
s
is
(
DA
)
a
r
e
u
s
ed
to
ev
al
u
ate
s
el
ec
ted
f
ea
tu
r
es.
T
h
e
n
ex
t
s
ec
tio
n
s
w
ill
d
e
m
o
n
s
tr
ate
r
esu
lts
p
r
o
d
u
ce
d
f
r
o
m
ap
p
l
y
i
n
g
f
ir
e
f
l
y
f
r
a
m
e
w
o
r
k
w
it
h
d
if
f
er
en
t
co
m
b
in
at
io
n
s
o
f
r
an
k
i
n
g
m
e
th
o
d
s
,
f
it
n
es
s
f
u
n
ct
io
n
s
,
an
d
cla
s
s
i
f
ier
s
o
n
d
atasets
.
O
u
r
e
x
p
er
i
m
e
n
t
h
as
b
ee
n
ap
p
lied
o
n
f
o
u
r
d
if
f
er
e
n
t
d
atase
ts
,
s
u
c
h
a
s
s
m
all
r
o
u
n
d
b
l
u
e
ce
l
l
t
u
m
o
r
s
(
S
R
B
C
T
)
w
h
ich
co
n
tai
n
s
4
d
if
f
er
en
t
t
u
m
o
r
s
.
L
u
n
g
d
ataset
co
n
tai
n
s
th
r
ee
d
i
f
f
er
e
n
t
cla
s
s
lab
el
s
.
Hep
atitis
d
ataset
co
n
tai
n
s
t
w
o
clas
s
es
(
li
v
e,
d
ie)
.
Der
m
a
to
lo
g
y
d
ataset
is
a
k
i
n
d
o
f
d
ataset
t
h
a
t su
f
f
er
f
r
o
m
d
if
f
er
en
tial d
iag
n
o
s
is
o
f
er
y
th
e
m
ato
-
s
q
u
a
m
o
u
s
d
is
ea
s
es
.I
t c
o
n
ta
in
s
6
d
if
f
er
en
t c
la
s
s
lab
els.
T
ab
le
2.
Dif
f
er
en
t
Data
s
ets
u
s
ed
in
t
h
e
E
x
p
er
i
m
en
ts
D
a
t
a
se
t
D
a
t
a
se
t
t
y
p
e
#
c
l
a
sse
s
#
f
e
a
t
u
r
e
s
S
a
mp
l
e
s
R
e
so
u
r
c
e
T
h
e
smal
l
r
o
u
n
d
b
l
u
e
c
e
l
l
t
u
mo
r
s (S
R
B
C
T
)
M
i
c
r
o
a
r
r
a
y
4
2
3
0
8
83
G
EM
S
w
e
b
si
t
e
(
w
w
w
.
g
e
ms
-
s
y
st
e
m.o
r
g
)
L
u
n
g
M
e
d
i
c
a
l
3
56
32
h
t
t
p
s:
/
/
a
r
c
h
i
v
e
.
i
c
s.
u
c
i
.
e
d
u
/
ml
/
d
a
t
a
se
t
s
/
L
u
n
g
+
C
a
n
c
e
r
H
e
p
a
t
i
t
i
s
M
e
d
i
c
a
l
B
i
n
a
r
y
c
l
a
ss
19
1
5
5
h
t
t
p
s:
/
/
a
r
c
h
i
v
e
.
i
c
s.
u
c
i
.
e
d
u
/
ml
/
d
a
t
a
se
t
s
/
H
e
p
a
t
i
t
i
s
D
e
r
mat
o
l
o
g
y
M
e
d
i
c
a
l
6
35
3
6
6
h
t
t
p
s:
/
/
a
r
c
h
i
v
e
.
i
c
s.
u
c
i
.
e
d
u
/
ml
/
d
a
t
a
se
t
s
/
D
e
r
mat
o
l
o
g
y
4
.
1
.
Resul
t
s
f
o
r
SRB
CT
d
a
t
a
s
et
T
h
is
s
ec
tio
n
d
e
m
o
n
s
tr
ate
s
th
e
r
esu
lts
o
f
ap
p
l
y
in
g
t
h
e
p
r
o
p
o
s
ed
f
r
am
e
w
o
r
k
o
n
SR
B
C
T
d
ataset.
T
ab
le
3
s
h
o
w
s
a
co
m
p
ar
is
o
n
a
m
o
n
g
d
i
f
f
er
e
n
t
f
it
n
es
s
f
u
n
cti
o
n
s
w
it
h
b
o
th
clas
s
i
f
ier
s
K
-
N
N
an
d
D
A
,
a
n
d
w
it
h
b
o
th
r
an
k
in
g
m
et
h
o
d
t
-
test
a
n
d
r
elief
f
.
I
t
s
h
o
w
s
t
h
at
u
s
i
n
g
r
an
k
i
n
g
m
et
h
o
d
t
-
te
s
t
w
i
th
r
o
s
en
b
r
o
ck
f
u
n
ctio
n
ev
alu
a
ted
b
y
clas
s
i
f
ier
K
-
NN
i
s
th
e
b
es
t c
o
m
b
i
n
atio
n
f
o
r
i
m
p
r
o
v
in
g
clas
s
i
f
i
ca
tio
n
ac
c
u
r
ac
y
.
I
n
w
h
ic
h
it
u
s
es 4
g
en
e
s
o
n
l
y
to
clas
s
i
f
y
u
n
s
ee
n
d
ata,
w
it
h
clas
s
if
icatio
n
ac
cu
r
ac
y
r
ea
ch
ed
its
h
i
g
h
er
v
alu
e
1
0
0
%.
W
h
ile
u
s
i
n
g
r
elief
f
w
i
th
K
-
N
N,
th
e
h
ig
h
e
s
t
clas
s
i
f
icatio
n
ac
c
u
r
ac
y
r
ea
ch
ed
9
5
%
b
y
7
g
e
n
es
u
s
in
g
Xi
n
-
s
h
e
y
a
n
g
f
it
n
es
s
f
u
n
ctio
n
.
T
ab
le
3
.
C
lass
if
icatio
n
A
cc
u
r
a
c
y
f
o
r
ea
ch
f
it
n
es
s
Fu
n
ctio
n
with
d
if
f
er
e
n
t Cl
a
s
s
i
f
ier
s
u
s
i
n
g
T
-
T
est an
d
R
elief
f
f
o
r
SR
B
C
T
d
ataset
in
(
%)
R
a
n
k
i
n
g
me
t
h
o
d
F
i
t
n
e
ss
F
u
n
c
t
i
o
n
A
c
k
l
e
y
R
o
se
n
b
r
o
c
k
S
p
h
e
r
e
X
i
n
-
s
h
e
y
a
n
g
R
a
st
r
i
g
i
n
S
a
l
o
mo
n
S
c
h
w
e
f
e
l
c
l
a
ssi
f
i
e
r
#
f
e
a
t
u
r
e
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
T
-
t
e
st
1
55
55
55
55
55
55
50
60
55
60
55
60
45
35
2
90
80
90
80
90
80
75
85
55
65
55
65
45
30
3
90
80
90
95
90
80
80
80
75
75
60
80
50
50
4
85
90
95
1
0
0
85
90
80
75
80
80
85
95
65
85
R
e
l
i
e
f
f
5
65
65
65
65
65
65
90
90
65
65
65
65
70
85
6
65
65
65
70
65
65
90
85
65
65
65
65
75
80
7
65
65
85
90
65
65
90
95
85
85
65
65
90
85
8
85
85
85
90
85
85
90
95
85
85
65
70
95
80
4
.
2
.
Resul
t
s
f
o
r
lun
g
d
a
t
a
s
et
I
n
th
i
s
s
ec
tio
n
t
h
e
r
es
u
lt
s
o
f
a
p
p
ly
i
n
g
t
h
e
p
r
o
p
o
s
ed
f
r
a
m
e
wo
r
k
o
n
lu
n
g
d
ataset
w
i
ll
b
e
d
em
o
n
s
tr
ated
.
A
s
et
o
f
d
if
f
er
e
n
t
co
m
b
in
at
io
n
s
o
f
tech
n
iq
u
es
w
a
s
ap
p
lied
.
As
s
h
o
w
n
in
T
ab
le
4
lu
n
g
d
at
aset
w
a
s
test
ed
w
ith
d
if
f
er
e
n
t
co
m
b
in
a
tio
n
s
o
f
f
it
n
es
s
f
u
n
ctio
n
s
,
r
an
k
in
g
m
e
th
o
d
s
an
d
class
if
ier
s
.
Us
in
g
r
an
k
i
n
g
m
et
h
o
d
t
-
tes
t
w
it
h
s
alo
m
o
n
f
u
n
ctio
n
an
d
ev
alu
ated
u
s
i
n
g
class
i
f
ier
K
-
N
N
y
ield
s
8
0
%
w
it
h
4
f
ea
tu
r
e
s
.
W
h
ils
t
u
s
i
n
g
D
A
class
i
f
ier
w
it
h
r
an
k
i
n
g
m
et
h
o
d
r
elief
f
,
a
n
d
s
p
h
er
e
f
u
n
ctio
n
with
4
f
ea
t
u
r
es i
n
cr
ea
s
ed
to
h
i
g
h
er
ac
cu
r
ac
y
9
0
%.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2088
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
4
,
A
u
g
u
s
t 2
0
1
8
:
2
3
3
8
–
2
3
5
0
2346
T
ab
le
4.
C
lass
if
ica
tio
n
A
cc
u
r
a
c
y
f
o
r
E
ac
h
Fi
tn
e
s
s
F
u
n
ctio
n
w
it
h
d
i
f
f
er
e
n
t Cl
a
s
s
i
f
ier
s
u
s
in
g
T
-
T
est
a
n
d
R
el
ief
f
f
o
r
L
u
n
g
Data
s
et
i
n
(
%)
R
a
n
k
i
n
g
me
t
h
o
d
F
i
t
n
e
ss
F
u
n
c
t
i
o
n
A
c
k
l
e
y
R
o
se
n
b
r
o
c
k
S
p
h
e
r
e
X
i
n
-
s
h
e
y
a
n
g
R
a
st
r
i
g
i
n
S
a
l
o
mo
n
S
c
h
w
e
f
e
l
c
l
a
ss
i
f
i
e
r
#
f
e
a
t
u
r
e
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
T
-
t
e
st
1
30
30
50
30
50
30
30
30
40
30
30
30
40
30
2
30
30
40
30
10
50
50
70
30
20
50
60
50
30
3
40
40
40
40
20
10
50
60
10
10
50
70
40
30
4
20
40
40
40
70
30
50
70
50
30
50
80
40
30
R
e
l
i
e
f
f
1
30
30
30
30
3
0
30
30
30
30
30
30
40
20
30
2
60
50
60
50
60
50
60
50
60
50
40
50
50
50
3
60
50
60
50
60
50
60
50
50
50
40
50
30
50
4
60
60
60
60
90
50
40
60
60
50
20
50
50
60
4
.
3
.
Resul
t
s
f
o
r
hepa
t
it
is
da
t
a
s
et
A
d
if
f
er
en
t
k
i
n
d
o
f
d
ataset
ca
lled
h
ep
atitis
is
tes
ted
w
ith
t
h
e
p
r
o
p
o
s
ed
f
r
am
e
w
o
r
k
.
I
t
is
co
n
s
id
er
ed
as
a
b
in
ar
y
clas
s
d
ataset
w
i
th
s
ta
tu
s
d
ie
o
r
liv
e.
T
ab
le
5
s
h
o
w
s
a
n
u
m
b
er
o
f
e
x
p
er
i
m
e
n
ts
ap
p
lied
o
n
th
i
s
d
ataset
w
it
h
d
i
f
f
er
en
t
co
m
b
i
n
atio
n
s
.
Usi
n
g
K
-
N
N
clas
s
i
f
ier
w
it
h
d
if
f
er
en
t
f
itn
e
s
s
f
u
n
c
tio
n
s
an
d
T
-
t
est
r
an
k
i
n
g
m
et
h
o
d
s
,
th
e
b
est
f
itn
e
s
s
f
u
n
c
tio
n
f
o
r
th
i
s
co
m
b
i
n
atio
n
w
a
s
Xin
-
s
h
e
y
a
n
g
w
h
ic
h
g
iv
e
s
cla
s
s
i
f
icatio
n
ac
cu
r
ac
y
7
9
%
w
i
th
2
f
ea
tu
r
es,
w
h
ile
u
s
in
g
r
el
ief
f
w
it
h
t
h
e
K
-
NN,
v
er
y
p
o
o
r
r
esu
lt
s
w
a
s
p
r
o
d
u
ce
d
.
Usi
n
g
r
elie
f
f
r
an
k
i
n
g
m
et
h
o
d
an
d
D
A
ac
h
iev
ed
t
h
e
b
e
s
t
class
i
f
icat
io
n
ac
c
u
r
ac
y
with
b
o
th
f
it
n
es
s
f
u
n
ctio
n
s
r
ast
r
ig
in
a
n
d
Xi
n
-
S
h
e
y
an
g
w
it
h
8
5
% b
y
2
f
ea
t
u
r
es.
T
ab
le
5.
C
lass
if
icatio
n
A
cc
u
r
a
c
y
f
o
r
E
ac
h
Fi
tn
e
s
s
F
u
n
ctio
n
w
it
h
d
i
f
f
er
e
n
t Cl
a
s
s
i
f
ier
s
u
s
in
g
T
-
T
est
a
n
d
R
elief
f
f
o
r
Hep
atitis
Data
s
e
t
i
n
(
%)
R
a
n
k
i
n
g
me
t
h
o
d
F
i
t
n
e
ss
F
u
n
c
t
i
o
n
A
c
k
l
e
y
R
o
se
n
b
r
o
c
k
S
p
h
e
r
e
X
i
n
-
s
h
e
y
a
n
g
R
a
st
r
i
g
i
n
S
a
l
o
mo
n
S
c
h
w
e
f
e
l
c
l
a
ss
i
f
i
e
r
#
f
e
a
t
u
r
e
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
T
-
t
e
st
1
60
55
68
50
68
50
81
42
10
21
63
42
68
50
2
55
66
11
40
74
61
73
79
10
29
55
42
73
60
3
60
68
18
63
76
71
71
58
26
60
58
60
76
71
4
5
52
58
55
55
50
45
71
74
71
45
76
73
76
76
55
47
50
44
71
71
74
71
78
55
71
45
76
73
R
e
l
i
e
f
f
1
76
21
76
21
73
21
73
21
73
21
73
21
81
42
2
76
71
76
21
73
60
85
50
85
50
10
21
73
60
3
26
60
71
68
68
66
73
55
71
68
13
18
10
55
4
5
50
50
60
60
23
50
60
60
50
50
60
60
68
50
63
60
68
50
63
60
50
50
60
60
10
50
55
60
4
.
4
.
Resul
t
s
f
o
r
de
r
m
a
t
o
lo
g
y
I
n
th
i
s
s
ec
tio
n
d
er
m
a
to
lo
g
y
d
ataset
is
in
tr
o
d
u
ce
d
w
it
h
a
co
m
p
ar
at
iv
e
T
ab
le
6
s
h
o
w
i
n
g
th
e
r
es
u
lt
s
o
f
ap
p
ly
i
n
g
th
e
p
r
o
p
o
s
ed
f
r
a
m
e
w
o
r
k
.
Der
m
ato
lo
g
y
i
s
a
k
i
n
d
o
f
s
k
i
n
ca
n
ce
r
t
h
at
co
n
tai
n
s
s
ix
d
if
f
er
en
t
cla
s
s
e
s
.
Dif
f
er
en
t
co
m
b
i
n
atio
n
s
o
f
tec
h
n
iq
u
es
ar
e
test
ed
f
o
r
th
e
b
est
p
er
f
o
r
m
a
n
ce
.
T
h
e
b
est
class
if
icatio
n
ac
cu
r
ac
y
w
a
s
o
b
tain
ed
th
r
o
u
g
h
ap
p
ly
in
g
K
-
NN
clas
s
i
f
ier
w
it
h
t
-
test
a
n
d
u
s
i
n
g
A
ck
le
y
f
itn
es
s
f
u
n
ctio
n
,
t
h
e
class
i
f
icatio
n
ac
cu
r
ac
y
w
as
9
7
%
b
y
1
4
f
ea
tu
r
e
s
.
W
h
ile
u
s
i
n
g
r
elie
f
f
,
r
es
u
lts
w
er
e
d
is
ap
p
o
in
ted
b
ec
au
s
e
i
t
d
ec
r
ea
s
es
w
it
h
h
ig
h
er
p
er
ce
n
t
ag
e.
D
A
h
as
b
ee
n
u
s
ed
w
it
h
t
-
test
,
t
h
e
h
i
g
h
est
clas
s
if
icati
o
n
ac
cu
r
ac
y
9
1
%
w
a
s
o
b
tain
ed
w
ith
9
f
ea
tu
r
e
s
u
s
i
n
g
R
o
s
e
n
b
r
o
ck
f
itn
e
s
s
f
u
n
ctio
n
,
an
d
9
5
%
w
it
h
1
0
f
ea
tu
r
e
s
u
s
i
n
g
Sc
h
w
e
f
el
f
it
n
es
s
f
u
n
ctio
n
.
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:
2
0
8
8
-
8708
F
ea
tu
r
e
S
elec
tio
n
A
p
p
r
o
a
ch
b
a
s
ed
o
n
F
ir
efly
A
lg
o
r
ith
m
a
n
d
C
h
i
-
s
q
u
a
r
e
(
E
ma
d
Mo
h
a
med
Ma
s
h
h
o
u
r
)
2347
T
ab
le
6.
C
lass
if
icatio
n
A
cc
u
r
a
c
y
f
o
r
E
ac
h
Fi
tn
e
s
s
F
u
n
ctio
n
w
it
h
d
i
f
f
er
e
n
t Cl
a
s
s
i
f
ier
s
u
s
in
g
T
-
T
est
a
n
d
R
elief
f
o
r
Der
m
ato
lo
g
y
Data
s
et
i
n
(
%)
R
a
n
k
i
n
g
me
t
h
o
d
F
i
t
n
e
ss
F
u
n
c
t
i
o
n
A
c
k
l
e
y
R
o
se
n
b
r
o
c
k
S
p
h
e
r
e
X
i
n
-
s
h
e
y
a
n
g
R
a
st
r
i
g
i
n
S
a
l
o
mo
n
S
c
h
w
e
f
e
l
c
l
a
ssi
f
i
e
r
#
f
e
a
t
u
r
e
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
DA
K
-
NN
T
-
t
e
st
9
72
75
91
80
8
5
72
79
70
46
52
72
52
90
72
10
80
77
89
88
83
81
91
88
64
74
66
54
95
77
11
80
78
91
89
86
85
91
88
77
74
66
60
95
79
12
80
78
91
89
95
94
91
87
77
74
78
66
92
82
13
94
89
91
89
95
95
91
88
77
74
78
66
92
82
14
95
97
92
86
94
94
93
90
77
74
89
7
8
93
83
15
95
97
92
86
94
94
93
90
74
74
89
78
93
83
R
e
l
i
e
f
f
1
20
35
25
36
54
54
40
36
54
54
24
37
36
36
2
27
35
38
40
46
58
59
54
36
54
24
37
54
54
3
53
33
70
58
50
58
72
54
63
54
24
37
61
54
4
5
6
7
68
70
70
65
51
51
54
50
70
73
73
69
58
61
63
6
3
50
64
64
71
58
56
56
58
82
82
75
79
58
60
64
66
63
63
63
63
54
54
54
54
35
41
49
77
40
40
48
63
61
60
63
63
54
54
54
54
5.
E
XP
E
R
I
M
E
NT
DI
SCUS
SI
O
N
T
h
e
p
r
o
p
o
s
ed
h
y
b
r
id
f
r
a
m
e
w
o
r
k
d
escr
ib
es
h
o
w
f
ir
ef
l
y
al
g
o
r
ith
m
h
a
s
b
ee
n
u
s
ed
as
a
f
ea
tu
r
e
s
elec
tio
n
to
o
l,
th
e
alg
o
r
it
h
m
w
as
a
s
s
e
s
s
ed
u
s
i
n
g
w
ell
-
k
n
o
w
n
d
atase
t
s
,
an
d
clas
s
i
f
icatio
n
er
r
o
r
r
ates
p
r
o
d
u
ce
d
b
y
t
h
e
s
elec
ted
f
ea
tu
r
es
w
er
e
m
o
n
it
o
r
ed
.
I
t
is
f
o
u
n
d
t
h
at
u
tili
z
in
g
c
h
i
-
s
q
u
ar
e
f
o
r
s
i
m
u
lati
n
g
f
ir
ef
l
y
p
o
s
itio
n
an
d
d
if
f
er
e
n
t
f
itn
e
s
s
f
u
n
ctio
n
s
f
o
r
s
i
m
u
lati
n
g
f
ir
e
f
l
y
l
ig
h
t
i
n
te
n
s
it
y
h
as
i
m
p
r
o
v
ed
th
e
f
ir
e
f
l
y
p
er
f
o
r
m
a
n
ce
f
o
r
f
ea
t
u
r
e
s
elec
tio
n
a
n
d
g
i
v
es
p
r
o
m
i
s
i
n
g
r
esu
lts
.
T
h
e
cla
s
s
i
f
i
ca
tio
n
p
er
f
o
r
m
a
n
ce
r
ep
r
esen
t
s
h
o
w
o
u
r
m
o
d
el
s
u
cc
ee
d
s
i
n
r
ed
u
ci
n
g
n
u
m
b
er
o
f
f
ea
t
u
r
es
a
n
d
s
elec
ti
n
g
th
e
m
o
s
t
i
n
f
o
r
m
ati
v
e
f
ea
tu
r
e
s
f
o
r
class
i
f
i
ca
tio
n
.
I
n
ev
er
y
r
u
n
n
i
n
g
tr
ial,
d
if
f
er
e
n
t
f
it
n
es
s
f
u
n
ctio
n
s
h
a
v
e
b
ee
n
e
x
p
er
i
m
e
n
ted
in
o
r
d
er
to
g
ain
th
e
m
o
s
t
s
u
itab
le
f
it
n
es
s
f
u
n
ctio
n
to
r
ep
r
esen
t
i
n
ten
s
it
y
.
E
ac
h
f
it
n
e
s
s
f
u
n
ctio
n
h
a
s
b
ee
n
ap
p
lied
o
n
d
i
f
f
er
e
n
t
d
ataset
s
w
it
h
t
w
o
d
if
f
er
e
n
t
r
a
n
k
i
n
g
ap
p
r
o
ac
h
e
s
t
-
te
s
t
a
n
d
r
elie
f
f
.
T
h
e
h
i
g
h
e
s
t
r
an
k
ed
f
ea
tu
r
e
s
h
av
e
b
ee
n
g
i
v
en
to
d
if
f
er
e
n
t
class
i
f
ier
s
o
n
e
b
y
o
n
e
f
o
r
e
v
alu
atio
n
.
T
esti
n
g
i
s
d
o
n
e
b
y
s
elec
ti
n
g
f
ir
s
t
f
ea
tu
r
e
f
r
o
m
th
e
f
ir
e
f
l
y
p
o
o
l
o
f
f
ea
t
u
r
es
f
o
r
class
i
f
icatio
n
,
th
e
n
r
esu
lt
s
ar
e
ev
alu
ated
,
if
cla
s
s
if
ica
tio
n
p
er
ce
n
ta
g
e
n
o
t
ac
ce
p
ted
an
o
th
er
f
ea
tu
r
e
f
r
o
m
th
e
p
o
o
l
is
ad
d
ed
to
th
e
p
r
ev
io
u
s
o
n
e,
a
n
d
th
en
p
ass
b
o
th
o
f
th
e
m
to
class
i
f
ier
,
an
d
ch
ec
k
f
o
r
class
i
f
icatio
n
ac
cu
r
ac
y
p
er
ce
n
tag
e,
th
e
p
r
o
ce
s
s
is
r
ep
ea
te
d
u
n
ti
l
t
h
e
h
ig
h
est
p
o
s
s
ib
le
ac
cu
r
ac
y
h
as
b
ee
n
ac
h
iev
ed
w
i
th
th
e
lo
w
e
s
t
n
u
m
b
er
o
f
f
ea
tu
r
e
s
.
I
n
s
id
e
t
h
e
f
r
a
m
e
w
o
r
k
t
h
er
e
ar
e
a
s
et
o
f
p
ar
am
eter
s
m
u
s
t
b
e
in
itial
ized
an
d
tu
n
ed
f
o
r
r
u
n
n
i
n
g
t
h
e
m
o
d
i
f
ied
f
ir
e
f
l
y
al
g
o
r
it
h
m
,
t
h
e
p
ar
a
m
eter
s
to
b
e
co
n
s
id
er
ed
s
u
ch
as
li
g
h
t
ab
s
o
r
p
tio
n
co
ef
f
icie
n
t
γ
,
attr
ac
tio
n
co
ef
f
icie
n
t
β,
r
an
d
o
m
i
za
tio
n
p
ar
a
m
eter
α
,
n
u
m
b
er
o
f
iter
atio
n
s
t
a
n
d
n
u
m
b
er
o
f
f
ir
e
f
l
y
p
o
p
u
latio
n
(
n
p
o
p
)
.
Nu
m
b
er
o
f
iter
atio
n
s
in
s
id
e
th
e
f
ir
e
f
l
y
f
r
a
m
e
w
o
r
k
m
a
y
v
ar
y
,
co
n
s
id
er
in
g
th
e
co
m
p
u
tatio
n
ti
m
e
a
n
d
co
s
t.
Af
ter
r
u
n
n
i
n
g
5
0
0
tr
ails
w
e
co
n
clu
d
e
t
h
at
t
h
e
b
est
r
an
g
e
o
f
iter
atio
n
s
f
o
r
th
e
p
r
o
p
o
s
ed
f
ir
ef
l
y
m
o
d
el
m
a
y
f
all
b
et
w
ee
n
1
5
0
an
d
4
0
0
ite
r
atio
n
s
.
O
u
ts
id
e
th
is
r
a
n
g
e
m
a
y
lead
t
h
e
f
ir
e
f
l
y
m
o
d
el
to
p
ick
lo
w
in
f
o
r
m
a
t
iv
e
f
ea
t
u
r
es
t
h
at
m
a
y
lead
t
h
e
class
if
ier
to
p
o
o
r
p
e
r
f
o
r
m
an
ce
.
N
u
m
b
er
o
f
p
o
p
u
latio
n
ch
o
s
e
n
f
o
r
f
ir
ef
l
y
p
r
o
ce
s
s
in
g
r
elies
o
n
n
u
m
b
er
o
f
f
ea
t
u
r
es
p
ick
ed
f
r
o
m
t
h
e
r
an
k
i
n
g
p
h
a
s
e,
as
s
tated
b
ef
o
r
e
r
an
k
i
n
g
p
h
ase
p
r
o
d
u
ce
as
m
u
c
h
as
p
o
s
s
ib
le
th
e
m
o
s
t
d
escr
ip
tiv
e
f
ea
tu
r
es
r
ea
d
y
f
o
r
f
ir
ef
l
y
al
g
o
r
it
h
m
.
γ
,
β
an
d
α
ar
e
th
r
ee
d
i
f
f
er
en
t
p
ar
a
m
eter
s
w
h
ic
h
m
a
y
co
n
tr
o
l
th
e
b
eh
a
v
io
u
r
o
f
f
ir
ef
l
y
in
s
p
ac
e,
tu
n
i
n
g
t
h
es
e
p
ar
am
eter
s
n
ee
d
s
m
o
r
e
t
h
a
n
o
n
e
ex
p
er
i
m
en
t.
T
h
e
aim
o
f
t
h
is
r
esear
ch
i
s
to
f
o
cu
s
o
n
i
m
p
r
o
v
i
n
g
f
ea
t
u
r
e
s
elec
tio
n
p
r
o
ce
s
s
u
s
i
n
g
f
ir
e
f
l
y
a
lg
o
r
ith
m
s
,
an
d
ac
h
iev
i
n
g
h
i
g
h
est
clas
s
i
f
i
ca
tio
n
r
ates
w
it
h
lo
w
est
n
u
m
b
er
o
f
f
ea
t
u
r
es.
T
h
e
lo
w
er
n
u
m
b
er
o
f
f
ea
tu
r
es
ca
n
b
e
s
elec
ted
as
lo
n
g
as
it
k
ee
p
s
th
e
ac
cu
r
ac
y
h
ig
h
.
T
h
e
f
ea
t
u
r
es
ex
tr
ac
ted
f
r
o
m
t
h
e
o
r
ig
i
n
al
d
ataset,
m
a
y
s
er
v
e
as
a
f
ea
tu
r
e
\
g
e
n
e
m
ar
k
er
s
t
h
at
ca
n
r
ec
o
g
n
ize
a
n
d
d
if
f
er
en
t
ia
te
class
e
s
.
T
h
e
f
o
llo
w
i
n
g
tab
l
es
s
h
o
w
t
h
e
n
a
m
e
s
o
f
th
e
d
o
m
i
n
an
t
f
ea
tu
r
es
\
g
e
n
es
t
h
at
i
m
p
r
o
v
e
th
e
c
lass
if
icatio
n
ac
c
u
r
ac
y
.
Fo
r
S
R
B
C
T
d
ataset,
T
ab
le
7
r
ep
r
esen
ts
s
elec
ted
g
en
e
s
th
a
t
m
a
y
h
elp
in
s
u
cc
e
s
s
f
u
l
d
iag
n
o
s
is
,
f
o
r
l
u
n
g
d
ataset,
t
h
e
r
e
is
n
o
a
p
r
o
p
er
d
escr
ip
t
io
n
f
o
r
t
h
e
f
ea
t
u
r
es
s
elec
ted
,
T
ab
le
8
s
h
o
w
s
t
h
e
d
o
m
in
a
n
t
f
ea
t
u
r
es
f
o
r
t
h
e
h
ep
atitis
d
ata
s
et,
w
h
ile
T
ab
le
9
s
h
o
w
s
t
h
e
i
m
p
o
r
tan
t
f
ea
tu
r
es i
n
th
e
d
er
m
a
to
lo
g
y
d
at
aset.
T
ab
le
7
.
Selecte
d
G
en
e
D
escr
ip
tio
n
f
o
r
SR
B
C
T
G
e
n
e
D
e
scri
p
t
i
o
n
G
e
n
e
1
:
i
n
su
l
i
n
-
l
i
k
e
g
r
o
w
t
h
f
a
c
t
o
r
2
(
so
mat
o
me
d
i
n
A
)
G
e
n
e
2
:
mi
c
r
o
t
u
b
u
l
e
-
a
sso
c
i
a
t
e
d
p
r
o
t
e
i
n
1
B
G
e
n
e
3
:
h
i
g
h
-
mo
b
i
l
i
t
y
g
r
o
u
p
(
n
o
n
h
i
st
o
n
e
c
h
r
o
mo
so
mal
)
G
e
n
e
4
:
EST
s
Evaluation Warning : The document was created with Spire.PDF for Python.