T
E
L
K
O
M
NIKA
T
elec
o
mm
un
ica
t
io
n Co
m
pu
t
i
ng
E
lect
ro
nics
a
nd
Co
ntr
o
l
Vo
l.
24
,
No
.
2
,
A
p
r
il
20
26
,
p
p
.
574
~
587
I
SS
N:
1
6
9
3
-
6
9
3
0
,
DOI
: 1
0
.
1
2
9
2
8
/
T
E
L
KOM
NI
K
A
.
v
24
i
2
.
27599
574
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//jo
u
r
n
a
l.u
a
d
.
a
c.
id
/in
d
ex
.
p
h
p
/TELK
OM
N
I
K
A
A
co
m
prehe
nsiv
e
a
na
ly
sis
of f
ea
tur
e select
io
n
a
nd
X
AI for
m
a
chine
learning
cla
ss
ifiers t
o
r
ecog
niz
e
gua
v
a
disea
se
Su
j
o
n Cha
nd
ra
Su
t
ra
d
ha
r,
M
d.
M
ehedi H
a
s
a
n
D
e
p
a
r
t
me
n
t
o
f
G
e
n
e
r
a
l
Ed
u
c
a
t
i
o
n
,
F
a
c
u
l
t
y
o
f
D
i
g
i
t
a
l
T
r
a
n
sf
o
r
mat
i
o
n
E
n
g
i
n
e
e
r
i
n
g
,
U
n
i
v
e
r
si
t
y
o
f
F
r
o
n
t
i
e
r
T
e
c
h
n
o
l
o
g
y
,
G
a
z
i
p
u
r
,
B
a
n
g
l
a
d
e
sh
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Oct
11
,
2025
R
ev
i
s
ed
J
an
9
,
2
0
2
6
A
cc
ep
ted
J
an
30
,
2
0
2
6
Re
c
o
g
n
izin
g
a
n
d
c
las
sify
in
g
d
ise
a
se
s
in
g
u
a
v
a
is
c
r
u
c
ial
f
o
r
m
a
n
a
g
in
g
f
a
r
m
s
to
k
e
e
p
c
ro
p
s
h
e
a
lt
h
y
a
n
d
in
c
re
a
se
h
a
rv
e
st
q
u
a
li
ty
.
Cu
lt
iv
a
to
rs
f
a
c
e
th
e
m
o
st
se
v
e
r
e
c
h
a
ll
e
n
g
e
s
w
h
e
n
it
c
o
m
e
s
to
re
c
o
g
n
izin
g
a
n
d
d
iag
n
o
si
n
g
g
u
a
v
a
f
ru
it
a
n
d
lea
f
il
ln
e
ss
e
s,
a
tas
k
t
h
a
t
is
n
e
a
rl
y
i
m
p
o
ss
ib
le
to
p
e
rf
o
rm
m
a
n
u
a
ll
y
.
T
h
is
re
se
a
rc
h
f
o
c
u
se
s
o
n
d
e
v
e
lo
p
in
g
a
ro
b
u
st
d
ise
a
se
id
e
n
t
if
ica
t
io
n
m
o
d
e
l
u
sin
g
im
a
g
e
d
a
ta
c
o
ll
e
c
ted
lo
c
a
ll
y
f
ro
m
g
u
a
v
a
tre
e
s.
Af
ter
d
a
ta
c
o
ll
e
c
ti
o
n
,
v
a
rio
u
s
im
a
g
e
p
ro
c
e
ss
in
g
te
c
h
n
iq
u
e
s,
in
c
lu
d
i
n
g
sc
a
li
n
g
a
n
d
c
o
n
tras
t
e
n
h
a
n
c
e
m
e
n
t,
a
re
u
ti
li
z
e
d
to
m
a
k
e
th
e
d
a
ta
m
o
re
su
it
a
b
le
f
o
r
u
se
.
K
-
m
e
a
n
s
c
lu
ste
rin
g
is
e
m
p
lo
y
e
d
to
q
u
ick
ly
d
iv
id
e
t
h
e
im
a
g
e
s
in
to
g
ro
u
p
s,
f
o
ll
o
w
e
d
b
y
th
e
e
x
trac
ti
o
n
o
f
im
p
o
rtan
t
c
h
a
ra
c
teristics
.
T
w
o
se
p
a
ra
te
fe
a
tu
re
ra
n
k
in
g
a
p
p
ro
a
c
h
e
s,
a
n
a
ly
sis
o
f
v
a
ri
a
n
c
e
(A
N
OV
A
)
a
n
d
lea
st
a
b
so
lu
te
sh
rin
k
a
g
e
se
lec
ti
o
n
o
p
e
ra
to
r
(L
A
S
S
O)
,
a
re
u
se
d
to
se
lec
t
th
e
b
e
st
c
h
a
r
a
c
t
e
risti
c
s,
id
e
n
ti
f
y
in
g
th
e
1
0
m
o
st
i
m
p
o
rtan
t
a
tt
r
i
b
u
tes
.
T
h
e
a
d
a
p
ti
v
e
b
o
o
st
in
g
(A
d
a
Bo
o
st)
c
las
sif
ier
a
c
h
iev
e
s
th
e
h
ig
h
e
st
a
c
c
u
ra
c
y
a
m
o
n
g
six
c
la
ss
i
f
iers
f
o
r
th
e
to
p
se
v
e
n
c
h
a
ra
c
teristics
in
d
ica
ted
b
y
LA
S
S
O
a
m
o
n
g
th
e
sp
e
c
if
ied
f
e
a
tu
re
s.
T
o
e
n
h
a
n
c
e
th
e
m
o
d
e
l’
s
in
terp
re
tab
i
li
ty
,
tw
o
e
x
p
lan
a
ti
o
n
m
e
th
o
d
s,
lo
c
a
l
in
ter
p
re
tab
le
m
o
d
e
l
-
a
g
n
o
sti
c
e
x
p
lan
a
ti
o
n
s
(L
IM
E)
a
n
d
sh
a
p
l
e
y
a
d
d
it
iv
e
e
x
p
lan
a
ti
o
n
s
(S
HA
P
)
,
a
re
e
m
p
lo
y
e
d
to
il
lu
stra
te
h
o
w
th
e
c
las
si
f
i
e
r
re
a
c
h
e
s
it
s
c
o
n
c
lu
sio
n
s.
T
h
is
a
p
p
ro
a
c
h
n
o
t
o
n
ly
si
m
p
li
f
ies
d
ise
a
se
i
d
e
n
ti
f
ica
ti
o
n
b
u
t
a
lso
c
larif
ies
th
e
re
a
so
n
in
g
b
e
h
i
n
d
p
re
d
icti
o
n
s,
o
p
e
n
in
g
th
e
d
o
o
r
to
re
a
l
-
w
o
rld
a
p
p
li
c
a
ti
o
n
s
i
n
d
e
tec
ti
n
g
a
n
d
p
re
v
e
n
ti
n
g
d
a
n
g
e
ro
u
s
d
ise
a
se
s
.
K
ey
w
o
r
d
s
:
A
d
aB
o
o
s
t
E
x
p
lain
ab
le
ar
ti
f
icial
in
telli
g
e
n
ce
Gu
a
v
a
d
is
ea
s
e
Ma
ch
i
n
e
lear
n
i
n
g
R
ec
o
g
n
itio
n
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Md
.
Me
h
ed
i H
asan
Dep
ar
t
m
en
t o
f
Gen
er
al
E
d
u
ca
t
io
n
,
Facu
lt
y
o
f
D
ig
i
tal
T
r
an
s
f
o
r
m
atio
n
E
n
g
i
n
ee
r
i
n
g
Un
i
v
e
r
s
it
y
o
f
Fro
n
tier
T
ec
h
n
o
lo
g
y
Kaliak
o
ir
,
Gaz
ip
u
r
-
1750
,
B
an
g
lad
es
h
E
m
ail:
m
e
h
ed
i0
0
0
1
@
u
f
tb
.
ac
.
b
d
1.
I
NT
RO
D
UCT
I
O
N
Gu
a
v
a
(
P
s
id
iu
m
g
u
a
ja
va
)
i
s
o
n
e
o
f
t
h
e
m
o
s
t
w
id
el
y
cu
lti
v
ated
tr
o
p
ical
f
r
u
i
ts
,
co
n
tr
ib
u
ti
n
g
s
u
b
s
ta
n
tial
l
y
to
g
lo
b
al
ag
r
ic
u
ltu
r
al
o
u
tp
u
t
a
n
d
tr
ad
e.
I
n
2
0
2
4
,
th
e
an
n
u
al
g
lo
b
al
p
r
o
d
u
ctio
n
o
f
g
u
a
v
a
w
a
s
esti
m
ated
at
n
ea
r
l
y
5
5
m
illi
o
n
to
n
n
e
s
,
w
i
th
I
n
d
ia
alo
n
e
ac
co
u
n
ti
n
g
f
o
r
4
5
%
o
f
th
e
to
tal
[
1
]
.
I
n
m
o
s
t
ca
s
es
,
h
u
m
a
n
b
ein
g
s
h
av
e
b
ee
n
t
h
e
o
n
es
to
g
r
o
w
t
h
e
g
u
a
v
a
s
ee
d
li
n
g
.
T
h
er
e
ar
e
in
s
ta
n
ce
s
i
n
w
h
ich
th
e
s
ee
d
s
o
f
th
e
g
u
a
v
a
h
a
v
e
b
ee
n
d
is
p
er
s
ed
b
y
b
ir
d
s
an
d
o
th
er
cr
ea
tu
r
es
w
ith
f
o
u
r
f
ee
t
f
o
r
s
u
c
h
a
lo
n
g
p
er
io
d
o
f
ti
m
e
th
a
t
th
eir
o
r
ig
in
i
s
u
n
k
n
o
w
n
.
Nev
er
t
h
ele
s
s
,
it i
s
th
o
u
g
h
t to
b
e
a
r
eg
io
n
th
at
r
an
g
es
f
r
o
m
th
e
s
o
u
t
h
er
n
p
ar
t o
f
Me
x
ico
in
t
o
C
en
tr
al
Am
e
r
ica
o
r
v
ia
C
e
n
t
r
al
Am
er
ica.
I
t
is
n
o
t
u
n
co
m
m
o
n
to
co
m
e
ac
r
o
s
s
g
u
a
v
a
b
u
s
h
es
in
all
w
ar
m
r
eg
io
n
s
o
f
tr
o
p
ical
Am
er
ica,
a
s
w
el
l
as
in
t
h
e
W
est
I
n
d
ies
(
s
in
ce
1
5
2
6
)
,
th
e
B
ah
a
m
as,
B
er
m
u
d
a,
an
d
s
o
u
th
er
n
Flo
r
id
a.
I
t
is
s
aid
to
h
av
e
b
ee
n
f
ir
s
t
i
m
p
le
m
en
ted
in
th
e
y
e
ar
1
8
4
7
,
an
d
b
y
th
e
y
ea
r
1
8
8
6
,
it
h
ad
s
p
r
ea
d
th
r
o
u
g
h
o
u
t
m
o
r
e
th
a
n
h
al
f
o
f
th
e
s
tate
[
1
]
.
B
ey
o
n
d
its
ec
o
n
o
m
ic
i
m
p
o
r
ta
n
ce
,
g
u
a
v
a
h
o
ld
s
n
u
tr
itio
n
al
an
d
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
A
co
mp
r
eh
en
s
ive
a
n
a
lysi
s
o
f fe
a
tu
r
e
s
elec
tio
n
a
n
d
X
A
I
fo
r
ma
ch
in
e
lea
r
n
in
g
…
(
S
u
jo
n
C
h
a
n
d
r
a
S
u
tr
a
d
h
a
r
)
575
m
ed
icin
a
l
v
al
u
e,
m
ak
i
n
g
it
a
v
ital
cr
o
p
f
o
r
b
o
th
f
ar
m
er
s
an
d
co
n
s
u
m
er
s
.
Ho
w
e
v
er
,
g
u
av
a
p
r
o
d
u
ctio
n
is
h
ig
h
l
y
v
u
l
n
er
ab
le
to
p
lan
t
d
is
ea
s
e
s
,
w
h
ic
h
t
h
r
ea
ten
cr
o
p
y
ield
,
r
ed
u
ce
q
u
alit
y
,
an
d
u
lti
m
ate
l
y
i
m
p
ac
t
f
ar
m
er
s
’
liv
eli
h
o
o
d
s
an
d
th
e
ag
r
icu
ltu
r
al
ec
o
n
o
m
y
.
T
im
el
y
an
d
p
r
ec
is
e
d
etec
tio
n
o
f
g
u
a
v
a
d
is
ea
s
es
is
cr
u
cial
to
m
i
n
i
m
ize
cr
o
p
lo
s
s
es a
n
d
en
s
u
r
e
s
u
s
tai
n
ab
le
p
r
o
d
u
ctio
n
.
I
t
is
p
o
s
s
ib
le
to
co
n
s
id
er
ar
tif
icia
l
in
te
lli
g
en
ce
(
A
I
)
an
d
d
ata
-
d
r
iv
en
m
et
h
o
d
o
lo
g
ies
to
b
e
th
e
co
n
te
m
p
o
r
ar
y
eq
u
i
v
alen
ts
o
f
th
e
ex
p
er
i
m
e
n
tal
co
m
p
o
n
e
n
t
o
f
th
e
s
cien
tific
m
et
h
o
d
,
w
h
ich
e
n
tail
s
th
e
m
et
h
o
d
ical
co
llectio
n
o
f
d
ata,
th
e
ex
a
m
i
n
atio
n
o
f
p
at
ter
n
s
,
co
r
r
elatio
n
s
,
an
d
th
e
est
ab
lis
h
m
e
n
t
o
f
lin
k
s
b
et
w
ee
n
e
v
en
t
s
t
h
at
h
a
v
e
b
ee
n
o
b
s
er
v
ed
[
2
]
.
AI
s
y
s
te
m
s
r
e
ca
p
tu
r
e
th
i
s
p
r
o
ce
s
s
b
y
r
ec
o
g
n
is
i
n
g
p
atter
n
s
a
n
d
g
en
er
ati
n
g
p
r
ed
ictio
n
s
o
n
l
y
b
ased
o
n
d
ata
th
at
h
as
b
ee
n
s
ee
n
,
w
h
ic
h
is
ac
co
m
p
lis
h
ed
v
ia
th
e
u
s
e
o
f
m
ac
h
i
n
e
le
ar
n
in
g
(
M
L
)
alg
o
r
ith
m
s
.
T
h
e
ass
o
ciatio
n
p
r
o
ce
s
s
th
at
i
s
in
tr
i
n
s
ic
to
AI
is
o
f
te
n
lack
i
n
g
in
tr
a
n
s
p
ar
en
c
y
,
w
h
ic
h
m
ak
e
s
it
d
if
f
ic
u
lt
f
o
r
h
u
m
a
n
s
to
ap
p
r
ec
iate
h
o
w
j
u
d
g
m
en
ts
ar
e
ar
r
i
v
ed
at
b
y
th
e
s
e
s
y
s
te
m
s
.
T
h
e
b
lac
k
b
o
x
asp
ec
t
o
f
m
a
n
y
A
I
m
o
d
els
r
ai
s
e
s
q
u
e
s
tio
n
s
ab
o
u
t
ac
co
u
n
tab
il
it
y
,
j
u
s
tice,
a
n
d
tr
u
s
t.
As
A
I
b
ec
o
m
e
s
a
n
in
cr
ea
s
i
n
g
l
y
d
r
i
v
in
g
f
o
r
ce
in
d
ec
is
io
n
-
m
a
k
i
n
g
ac
r
o
s
s
a
v
ar
i
et
y
o
f
s
ec
to
r
s
,
th
e
n
ee
d
f
o
r
e
x
p
lan
atio
n
s
t
h
at
ar
e
b
o
th
v
is
ib
le
a
n
d
u
n
d
er
s
tan
d
ab
l
e
f
o
r
th
ese
co
n
n
ec
tio
n
s
b
ec
o
m
es o
f
th
e
u
t
m
o
s
t i
m
p
o
r
tan
ce.
T
h
is
r
esear
ch
ai
m
s
to
d
ev
elo
p
an
d
ass
es
s
an
a
u
to
m
ated
s
y
s
t
e
m
to
aid
f
ar
m
er
s
in
id
e
n
ti
f
y
i
n
g
g
u
av
a
d
is
ea
s
es
b
y
p
r
o
m
p
t
d
etec
tio
n
an
d
an
al
y
s
i
s
,
th
u
s
a
v
er
tin
g
s
u
b
s
tan
tial
a
g
r
icu
ltu
r
al
lo
s
s
es.
A
s
y
s
te
m
t
h
at
is
ea
s
y
to
u
s
e
an
d
s
et
u
p
co
u
ld
im
p
r
o
v
e
g
u
a
v
a
p
r
o
d
u
ctio
n
,
w
h
ich
is
a
m
aj
o
r
co
n
tr
ib
u
to
r
t
o
g
lo
b
al
g
r
o
s
s
d
o
m
est
ic
p
r
o
d
u
ct
(
GDP
)
th
r
o
u
g
h
lar
g
e
-
s
ca
le
e
x
p
o
r
ts
,
w
h
ile
also
s
o
l
v
in
g
lo
n
g
s
tan
d
i
n
g
p
r
o
b
le
m
s
i
n
ag
r
ic
u
lt
u
r
e.
T
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
lo
w
er
s
t
h
e
ch
an
ce
o
f
w
id
esp
r
ea
d
h
ar
m
b
y
m
a
k
i
n
g
it
ea
s
ier
to
f
i
n
d
in
f
ec
tio
n
s
ea
r
l
y
o
n
.
I
t
also
g
i
v
es
ad
v
ice
o
n
s
a
f
e
g
r
o
w
i
n
g
p
r
ac
tices
an
d
,
i
n
t
h
e
e
n
d
,
h
elp
s
in
cr
ea
s
e
p
r
o
f
its
b
y
o
p
ti
m
izi
n
g
g
r
o
w
t
h
cir
cu
m
s
ta
n
ce
s
.
T
h
is
s
tu
d
y
is
m
o
s
tl
y
o
n
ex
p
la
in
ab
le
AI
(
XA
I
)
.
ML
h
a
s
e
m
er
g
ed
as
a
v
er
y
s
u
cc
e
s
s
f
u
l
m
e
th
o
d
o
lo
g
y
f
o
r
p
atter
n
r
ec
o
g
n
itio
n
i
n
e
x
ten
s
i
v
e
an
d
in
tr
icate
d
atasets
;
n
ev
er
t
h
eles
s
,
its
d
ef
icie
n
c
y
in
in
ter
p
r
etab
ilit
y
h
in
d
er
s
it
s
i
m
p
le
m
e
n
tatio
n
i
n
ess
e
n
tial
s
ec
to
r
s
s
u
ch
a
s
ag
r
ic
u
lt
u
r
e.
ML
h
a
s
alr
ea
d
y
s
h
o
w
n
p
r
o
m
is
i
n
g
r
esu
lt
s
in
d
iag
n
o
s
i
n
g
p
lan
t
d
is
ea
s
e
s
[
3
]
.
A
lg
o
r
it
h
m
s
m
ad
e
f
o
r
a
u
to
m
atic
d
i
s
ea
s
e
d
etec
tio
n
ca
n
g
i
v
e
u
s
e
f
u
l
h
i
n
t
s
t
h
at
h
elp
f
in
d
p
r
o
b
lem
s
ea
r
ly
o
n
,
m
ak
in
g
it
ea
s
ier
to
tr
ea
t
an
d
m
an
a
g
e
th
e
m
q
u
ic
k
l
y
.
Ho
w
ev
er
,
tr
a
d
itio
n
al
d
is
ea
s
e
d
etec
tio
n
s
til
l r
elies
h
ea
v
il
y
o
n
ex
p
er
t b
o
tan
i
s
ts
lo
o
k
i
n
g
at
p
lan
ts
,
w
h
ic
h
i
s
co
s
tl
y
,
ta
k
es
a
l
o
t o
f
ti
m
e,
an
d
i
s
n
’
t
al
w
a
y
s
e
f
f
ec
ti
v
e.
T
h
e
m
a
in
g
o
al
o
f
t
h
is
r
e
s
ea
r
ch
is
to
u
s
e
s
i
x
d
is
ti
n
ct
ML
m
o
d
els
in
a
n
i
m
a
g
e
-
b
ased
f
r
a
m
e
wo
r
k
to
tell
th
e
d
if
f
er
en
ce
b
et
w
ee
n
h
ea
l
th
y
a
n
d
d
ef
ec
ti
v
e
g
u
a
v
a
leav
e
s
an
d
f
r
u
it
s
.
T
o
ac
co
m
p
lis
h
th
is
,
d
if
f
er
en
t
i
m
a
g
e
p
r
o
ce
s
s
in
g
m
et
h
o
d
s
w
er
e
u
s
e
d
b
ef
o
r
e
d
is
ea
s
e
s
eg
m
e
n
tat
io
n
an
d
f
ea
tu
r
e
ex
tr
ac
tio
n
.
Als
o
,
tw
o
m
et
h
o
d
s
f
o
r
s
elec
ti
n
g
f
ea
t
u
r
es
w
er
e
u
s
ed
t
o
r
an
k
an
d
p
r
io
r
itize
th
e
m
,
wh
ich
m
ad
e
t
h
e
m
o
d
el
m
o
r
e
ac
cu
r
ate
an
d
e
f
f
ic
ien
t.
L
ast
l
y
,
X
A
I
m
e
th
o
d
s
w
er
e
u
s
ed
to
ex
p
lain
t
h
e
p
r
ed
ict
iv
e
r
esu
lt
s
,
w
h
ich
m
ad
e
t
h
e
m
o
d
el
’
s
d
ec
i
s
io
n
-
m
ak
i
n
g
p
r
o
ce
s
s
clea
r
an
d
en
s
u
r
ed
th
e
p
r
o
p
o
s
ed
s
y
s
te
m
w
as
r
eliab
le.
T
h
e
r
est
o
f
th
is
p
ap
er
is
s
tr
u
c
tu
r
ed
as:
s
ec
tio
n
2
r
ev
ie
w
s
r
elate
d
w
o
r
k
in
d
etec
tin
g
g
u
av
a
d
is
ea
s
es
a
n
d
ML
.
Sectio
n
3
d
escr
ib
es
th
e
p
r
o
p
o
s
ed
f
r
a
m
e
w
o
r
k
a
n
d
m
et
h
o
d
o
lo
g
ies.
Sectio
n
4
p
r
esen
ts
t
h
e
ex
p
er
i
m
en
tal
s
et
u
p
an
d
r
esu
lt
s
,
f
o
llo
w
ed
b
y
a
d
is
c
u
s
s
io
n
i
n
s
ec
tio
n
5
.
Fin
all
y
,
t
h
e
p
ap
er
co
n
clu
d
es
with
f
u
t
u
r
e
r
esear
ch
d
ir
ec
tio
n
s
.
2.
L
I
T
E
R
AT
U
RE
RE
VI
E
W
I
n
th
e
m
o
d
er
n
d
a
y
,
t
h
e
m
aj
o
r
it
y
o
f
r
esear
c
h
o
n
ML
a
n
d
d
ee
p
lear
n
i
n
g
is
m
o
s
tl
y
co
n
c
en
tr
ated
o
n
ag
r
icu
l
tu
r
al
p
r
o
b
le
m
s
s
i
n
ce
th
i
s
in
d
u
s
tr
y
m
a
k
es
a
s
i
g
n
i
f
ica
n
t
co
n
tr
ib
u
tio
n
to
th
e
ec
o
n
o
m
y
o
f
th
e
w
h
o
le
g
lo
b
e.
On
t
h
e
o
th
er
h
a
n
d
,
th
er
e
is
a
li
m
ited
a
m
o
u
n
t o
f
s
t
u
d
y
o
n
th
e
d
is
ea
s
e
id
en
t
if
ica
tio
n
o
f
f
r
u
it
s
lik
e
g
u
a
v
a,
m
an
g
o
,
j
ac
k
f
r
u
it
s
,
an
d
s
o
o
n
.
Fo
r
t
h
e
p
u
r
p
o
s
e
o
f
id
en
ti
f
y
i
n
g
g
u
a
v
a
l
ea
f
d
is
ea
s
e,
H
o
w
lad
er
et
a
l.
[
4
]
d
ev
elo
p
ed
a
d
ee
p
co
n
v
o
lu
tio
n
al
n
e
u
r
al
n
et
w
o
r
k
(
DC
NN)
-
m
o
d
el.
Fo
r
th
e
p
u
r
p
o
s
e
o
f
d
ev
elo
p
in
g
t
h
e
m
o
d
el,
2
7
0
5
p
h
o
to
s
illu
s
tr
atin
g
f
o
u
r
d
if
f
er
en
t
d
is
e
ases
w
er
e
u
s
ed
.
I
t
w
a
s
d
u
r
in
g
th
e
tr
ain
i
n
g
a
n
d
test
i
n
g
p
h
a
s
e
th
at
th
e
y
u
s
ed
2
5
ep
o
ch
s
th
at
t
h
e
y
a
ttain
ed
a
n
ac
cu
r
ac
y
o
f
9
8
.
7
4
% a
n
d
9
9
.
4
3
%
r
esp
ec
tiv
el
y
.
T
h
e
id
en
tif
icatio
n
o
f
p
lan
t
leaf
d
is
ea
s
e
w
as
ac
co
m
p
li
s
h
ed
b
y
Gee
th
ar
a
m
an
i
an
d
P
an
d
ia
n
[
5
]
v
ia
th
e
d
ev
elo
p
m
en
t
o
f
a
m
o
d
el
th
a
t
u
s
ed
a
n
i
n
e
-
la
y
er
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
e
t
w
o
r
k
(
C
NN)
ar
ch
itect
u
r
e.
B
o
th
th
e
P
lan
t
Villag
e
d
ata
s
et
an
d
th
e
Kag
g
le
d
ataset
w
er
e
u
s
ed
b
y
t
h
e
m
.
T
h
e
Kag
g
le
d
ataset
h
ad
5
5
4
4
8
p
h
o
to
s
o
f
1
3
d
if
f
er
e
n
t
p
lan
t
leav
e
s
t
h
at
w
e
r
e
ca
teg
o
r
ized
in
to
3
8
d
if
f
er
e
n
t
g
r
o
u
p
s
.
C
o
m
p
ar
i
n
g
th
e
s
u
g
g
e
s
ted
m
o
d
el
w
it
h
o
th
er
class
i
f
icatio
n
m
eth
o
d
s
s
u
ch
a
s
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
i
n
es
(
SVM)
,
lo
g
is
t
ic
r
eg
r
ess
io
n
(
L
R
)
,
d
ec
is
io
n
tr
ee
s
(
DT
)
,
an
d
k
-
n
ea
r
est
n
ei
g
h
b
o
r
s
(
KNN
)
class
if
ier
s
r
ev
ea
led
th
at
th
e
C
N
N
-
m
o
d
el
p
er
f
o
r
m
e
d
th
e
b
est,
w
i
th
an
i
m
p
r
ess
i
v
e
p
r
ed
ictio
n
ac
cu
r
ac
y
o
f
9
6
.
4
6
%.
T
u
r
k
o
g
lu
et
a
l.
[
6
]
p
r
esen
ted
a
m
u
lti
-
m
o
d
el
p
r
e
-
tr
ai
n
ed
C
NN
m
o
d
el
f
o
r
id
en
tify
i
n
g
A
p
p
le
illn
ess
an
d
p
ests
.
T
h
e
m
o
d
el
u
s
ed
th
e
A
le
x
Net,
Go
o
g
leNe
t,
an
d
D
en
s
eNe
t2
0
1
m
o
d
els,
an
d
it
was
tr
ain
ed
o
n
1
1
9
2
p
ictu
r
es
th
at
d
ep
icted
f
o
u
r
p
r
o
m
i
n
en
t
ap
p
le
d
is
ea
s
es.
W
it
h
a
s
co
r
e
o
f
9
6
.
1
0
%,
th
e
Den
s
eN
et2
0
1
ac
h
iev
ed
th
e
g
r
ea
test
ac
c
u
r
ac
y
s
co
r
e
a
m
o
n
g
th
e
m
o
d
els
t
h
at
w
er
e
ap
p
lied
.
Fo
r
th
e
p
u
r
p
o
s
e
o
f
ev
alu
a
ti
n
g
t
h
e
e
f
f
ec
tiv
e
n
e
s
s
o
f
d
ee
p
lear
n
in
g
ap
p
r
o
ac
h
es
in
d
etec
ti
n
g
s
w
ee
t
n
ess
a
n
d
q
u
alit
y
,
J
u
p
u
d
i
[
7
]
u
s
ed
an
i
m
ag
e
c
lass
if
ica
tio
n
s
y
s
te
m
o
n
a
n
o
r
an
g
e.
E
v
en
t
h
o
u
g
h
t
h
e
s
o
u
r
ce
o
f
th
e
d
ataset
w
a
s
n
o
t
d
is
clo
s
ed
,
t
h
e
o
b
j
ec
tiv
e
o
f
t
h
e
r
esear
c
h
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
24
,
No
.
2
,
A
p
r
il
20
26
:
5
7
4
-
587
576
p
r
o
j
ec
t
w
as
ap
p
lied
to
a
to
tal
o
f
f
iv
e
t
h
o
u
s
a
n
d
i
m
a
g
es.
T
h
e
m
o
d
el
w
as
tr
ain
ed
u
s
in
g
SV
M,
A
lex
Net,
s
tack
ed
au
to
e
n
co
d
er
(
S
A
E
)
,
an
d
k
er
n
el
s
p
ar
s
e
s
tac
k
ed
au
to
e
n
co
d
er
(
KSS
A
E
)
,
w
ith
KSS
A
E
r
ea
ch
in
g
t
h
e
h
i
g
h
e
s
t
p
o
s
s
ib
le
ac
cu
r
ac
y
o
f
9
2
.
1
%
th
r
o
u
g
h
o
u
t
th
e
co
u
r
s
e
o
f
th
e
tr
ain
in
g
p
r
o
ce
s
s
.
Den
s
eNe
t2
0
1
s
ee
m
s
to
h
av
e
th
e
m
o
s
t o
u
t
s
ta
n
d
in
g
p
er
f
o
r
m
a
n
ce
,
as seen
b
y
it
s
s
co
r
e
o
f
9
6
.
1
%
f
r
o
m
th
e
to
tal.
A
d
ee
p
r
esid
u
al
n
et
w
o
r
k
(
R
es
Net)
th
at
u
ti
lizes
a
co
n
tr
ast
en
h
an
ce
m
e
n
t
a
n
d
tr
an
s
f
er
lear
n
i
n
g
s
tr
ate
g
y
w
a
s
s
u
g
g
es
ted
b
y
T
r
an
g
et
a
l.
[
8
]
in
o
r
d
er
to
r
ec
o
g
n
ize
m
an
g
o
illn
e
s
s
.
W
it
h
an
ac
c
u
r
ac
y
r
ate
o
f
8
8
.
4
6
%,
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
w
a
s
ab
le
to
p
r
o
p
er
ly
d
etec
t
t
h
r
ee
p
r
ev
a
len
t
ai
l
m
e
n
t
s
b
ased
o
n
a
to
tal
o
f
3
9
4
i
m
a
g
es.
I
t
w
a
s
s
u
g
g
ested
b
y
N
ik
h
it
h
a
et
a
l.
[
9
]
th
at
th
e
I
n
ce
p
tio
n
V3
m
o
d
el
b
e
u
s
ed
f
o
r
th
e
p
r
o
ce
s
s
es o
f
illn
e
s
s
d
etec
tio
n
an
d
f
r
u
i
t
id
en
ti
f
icatio
n
.
B
an
a
n
as,
ap
p
les,
an
d
ch
er
r
ies
w
er
e
s
elec
t
ed
as
d
is
ea
s
e
d
etec
tio
n
tar
g
et
s
,
an
d
th
e
I
n
ce
p
tio
n
V3
m
o
d
el
w
as a
p
p
lied
o
n
l
y
to
th
ese
f
r
u
its
.
GitH
u
b
w
as t
h
e
s
o
u
r
ce
o
f
th
is
i
n
f
o
r
m
ati
o
n
as
w
ell.
A
D
C
NN
w
a
s
p
r
o
p
o
s
ed
b
y
Ma
et
a
l
.
[
1
0
]
f
o
r
th
e
p
u
r
p
o
s
e
o
f
s
y
m
p
to
m
-
w
is
e
d
ia
g
n
o
s
is
o
f
f
o
u
r
cu
cu
m
b
er
d
is
o
r
d
er
s
.
T
h
e
n
et
w
o
r
k
ac
h
iev
ed
a
r
ec
o
g
n
itio
n
r
at
e
o
f
9
3
.
4
%.
P
r
ep
r
o
ce
s
s
in
g
an
d
class
if
icat
io
n
ar
e
t
w
o
ex
a
m
p
les
o
f
w
e
ll
-
k
n
o
w
n
i
m
a
g
e
p
r
o
ce
s
s
in
g
p
r
o
c
ess
es
th
at
ar
e
u
s
ed
in
th
e
m
et
h
o
d
th
at
w
as
p
r
esen
ted
b
y
P
r
ak
ash
et
a
l
.
[
1
1
]
f
o
r
th
e
p
u
r
p
o
s
e
o
f
id
en
ti
f
y
in
g
ill
n
es
s
es
th
at
a
f
f
ec
t
lea
v
es.
An
e
v
alu
a
t
io
n
o
f
t
h
e
o
f
f
er
ed
m
et
h
o
d
is
ca
r
r
ied
o
u
t
o
n
a
co
l
lectio
n
o
f
s
i
x
t
y
i
m
a
g
es,
o
f
w
h
ich
t
h
ir
t
y
-
f
i
v
e
ar
e
ca
n
ce
r
o
u
s
a
n
d
t
w
e
n
t
y
-
f
i
v
e
ar
e
b
en
ig
n
,
w
it
h
a
n
ac
cu
r
ac
y
r
at
e
o
f
9
0
%.
T
h
e
K
-
m
ea
n
s
cl
u
s
ter
in
g
m
et
h
o
d
is
u
s
ed
to
s
e
g
m
e
n
t
th
e
d
i
s
ea
s
e
-
af
f
ec
ted
r
eg
io
n
,
an
d
th
en
th
e
g
r
a
y
-
le
v
el
co
-
o
cc
u
r
r
en
ce
m
atr
ix
(
GL
C
M
)
alg
o
r
ith
m
i
s
u
s
ed
to
ex
tr
ac
t
f
ea
tu
r
es
f
r
o
m
t
h
e
s
e
g
m
en
ted
ar
ea
.
A
f
t
e
r
th
e
f
ea
t
u
r
e
v
ec
to
r
h
a
s
b
ee
n
f
o
r
m
ed
,
it
is
n
e
x
t
ca
te
g
o
r
iz
ed
b
y
u
tili
zi
n
g
th
e
SVM
clas
s
i
f
ier
.
W
ith
th
e
u
s
e
o
f
6
8
8
i
m
ag
e
s
,
A
l
B
u
h
ai
s
i
[
1
2
]
w
a
s
ab
le
to
d
eter
m
i
n
e
t
h
e
k
i
n
d
o
f
p
i
n
ea
p
p
le
b
y
u
s
i
n
g
th
e
VGG1
6
m
o
d
el.
A
h
u
n
d
r
ed
p
er
ce
n
t
ac
cu
r
ac
y
w
a
s
ac
h
iev
ed
b
y
th
e
t
r
ain
ed
m
o
d
el,
an
d
it
is
q
u
ite
p
r
o
b
a
b
le
th
at
th
i
s
d
ataset
w
as
o
v
er
f
itti
n
g
; o
th
er
w
i
s
e,
th
e
ac
c
u
r
ac
y
w
o
u
ld
n
o
t h
a
v
e
b
ee
n
ac
h
iev
a
b
le.
A
d
ia
g
n
o
s
tic
tec
h
n
iq
u
e
t
h
at
u
t
ilizes
d
ee
p
lear
n
i
n
g
w
as
d
escr
ib
ed
b
y
E
lle
u
ch
e
t
a
l.
[
1
3
]
.
D
u
r
in
g
t
h
is
in
v
e
s
ti
g
atio
n
,
th
e
y
m
ad
e
u
s
e
o
f
th
eir
r
ec
en
tl
y
d
ev
elo
p
ed
d
a
taset,
w
h
ic
h
in
cl
u
d
ed
f
iv
e
d
if
f
er
en
t
k
in
d
s
o
f
p
lan
t
d
ata.
A
s
p
ar
t
o
f
t
h
e
tr
ai
n
in
g
p
r
o
ce
s
s
f
o
r
th
eir
m
o
d
el,
th
e
y
u
s
ed
tr
an
s
f
er
lear
n
i
n
g
ar
ch
itect
u
r
e
u
s
in
g
VGG
-
1
6
an
d
R
e
s
Net
.
t.
I
n
o
r
d
er
to
ev
alu
ate
t
h
e
v
alid
it
y
o
f
t
h
is
m
o
d
el,
th
e
y
co
m
p
ar
ed
th
e
s
u
g
g
e
s
ted
m
o
d
el
to
b
o
th
ac
tu
al
d
ata
an
d
d
ata
th
at
co
n
t
ain
s
a
u
g
m
e
n
tatio
n
s
.
W
ith
t
h
e
u
s
e
o
f
tr
a
n
s
f
er
lear
n
i
n
g
,
VG
G
-
1
6
p
r
o
g
r
ess
iv
el
y
g
av
e
r
es
u
lts
t
h
at
w
er
e
b
o
th
p
r
o
m
is
i
n
g
an
d
r
ea
lis
tic
i
n
ter
m
s
o
f
ac
cu
r
ac
y
,
w
i
th
9
9
.
0
2
%
an
d
9
8
.
3
5
%
r
esp
ec
tiv
el
y
.
An
ap
p
r
o
ac
h
to
th
e
id
en
ti
f
icat
i
o
n
o
f
g
u
a
v
a
ill
n
es
s
th
at
i
s
b
ase
d
o
n
co
m
p
u
ter
v
i
s
io
n
w
a
s
d
ev
elo
p
ed
b
y
E
lleu
c
h
et
a
l.
[
1
3
]
.
T
h
is
ap
p
r
o
ac
h
m
a
k
es
u
s
e
o
f
t
h
r
ee
C
N
N
-
b
ased
m
o
d
el
s
w
it
h
d
i
s
ti
n
ct
o
p
ti
m
izer
s
.
O
n
t
h
e
o
th
er
h
an
d
,
th
e
y
d
o
n
o
t
s
p
ec
if
y
a
n
y
tr
u
s
t
w
o
r
t
h
y
o
n
li
n
e
s
o
u
r
ce
s
f
o
r
th
e
d
ata
th
at
w
a
s
ac
q
u
ir
ed
.
B
o
th
th
e
d
r
o
p
o
u
t
v
al
u
e
an
d
t
h
e
t
h
ir
d
o
p
ti
m
izer
s
h
o
w
ed
r
e
m
ar
k
ab
le
ac
cu
r
ac
y
w
h
e
n
th
e
d
r
o
p
o
u
t
was
5
0
%,
w
h
ic
h
w
as
9
6
.
1
%.
A
DC
NN
b
ased
tech
n
iq
u
e
w
a
s
p
r
esen
ted
b
y
Mo
s
ta
f
a
et
a
l.
[
1
4
]
f
o
r
th
e
p
u
r
p
o
s
e
o
f
d
etec
t
in
g
g
u
av
a
illn
e
s
s
.
T
h
is
ap
p
r
o
ac
h
u
s
ed
f
i
v
e
d
if
f
er
e
n
t
n
eu
r
al
n
et
w
o
r
k
n
e
t
w
o
r
k
ar
ch
itect
u
r
es.
T
h
e
d
atas
et
th
at
th
e
y
u
til
ized
w
a
s
o
n
e
th
at
w
as
ac
q
u
ir
ed
lo
ca
ll
y
in
P
ak
i
s
tan
.
Ha
v
i
n
g
ac
h
ie
v
ed
an
ac
cu
r
ac
y
r
ate
o
f
9
7
.
7
4
%,
th
e
class
i
f
ica
tio
n
r
esu
lt d
e
m
o
n
s
tr
ated
t
h
at
R
e
s
N
et
-
1
0
1
w
as t
h
e
m
o
d
el
t
h
at
w
as
th
e
m
o
s
t s
u
itab
le
f
o
r
th
eir
p
u
r
p
o
s
e.
An
ap
p
licatio
n
o
f
n
i
n
e
i
m
p
o
r
tan
t
cla
s
s
i
f
ier
s
w
a
s
p
r
o
p
o
s
ed
b
y
Hab
ib
et
a
l.
[
1
5
]
f
o
r
th
e
p
u
r
p
o
s
e
o
f
d
ev
elo
p
in
g
a
m
ac
h
in
e
v
is
io
n
-
b
ased
d
is
ea
s
e
d
etec
tio
n
s
y
s
te
m
f
o
r
th
e
p
u
r
p
o
s
e
o
f
id
en
ti
f
y
in
g
i
lln
e
s
s
i
n
t
h
r
ee
d
if
f
er
e
n
t
s
p
ec
ies
o
f
f
r
u
it,
n
a
m
el
y
g
u
a
v
a,
p
ap
a
y
a,
a
n
d
j
ac
k
f
r
u
it.
W
h
en
it
ca
m
e
to
r
ec
o
g
n
iz
in
g
i
lln
e
s
s
e
s
t
h
at
af
f
ec
t
g
u
av
a
a
n
d
j
ac
k
f
r
u
it,
th
e
r
an
d
o
m
f
o
r
est
(
R
F)
class
i
f
ier
ac
h
iev
ed
t
h
e
b
est
ac
cu
r
ac
y
,
with
a
r
ate
o
f
9
6
.
8
%
an
d
8
9
.
5
9
%,
r
esp
ec
tiv
el
y
.
Fr
o
m
th
e
ab
o
v
e
a
n
al
y
s
i
s
it
ca
n
b
e
s
aid
th
at
m
o
s
t
o
f
r
esear
c
h
o
n
g
u
av
a
d
is
ea
s
es
r
ec
o
g
n
itio
n
h
a
s
b
ee
n
p
er
f
o
r
m
ed
u
s
i
n
g
ML
o
r
d
ee
p
lear
n
in
g
class
i
s
if
ier
.
T
h
er
e
is
n
o
t
an
y
r
esear
ch
o
n
f
ea
tu
r
e
s
elec
tio
n
a
n
d
X
A
I
b
ased
i
m
p
l
e
m
en
tatio
n
w
h
ic
h
is
o
u
r
m
o
tiv
atio
n
to
w
o
r
k
w
i
th
.
3.
M
E
T
H
O
D
Dis
ea
s
e
s
th
at
a
f
f
ec
t
g
u
a
v
a
f
r
u
it
m
a
y
b
e
clas
s
i
f
ied
u
s
i
n
g
a
v
ar
iet
y
o
f
m
et
h
o
d
s
,
ea
ch
o
f
w
h
ic
h
is
f
u
r
t
h
er
s
u
b
d
iv
id
ed
i
n
to
a
g
r
ea
t
n
u
m
b
er
o
f
ca
te
g
o
r
ies.
T
h
e
f
ir
s
t
t
h
i
n
g
th
at
w
e
d
id
w
a
s
g
o
o
u
t
in
to
th
e
f
ie
ld
an
d
ca
p
tu
r
e
th
e
i
m
ag
e.
B
ef
o
r
e
b
eg
in
n
in
g
t
h
is
i
n
q
u
ir
y
,
th
e
s
ec
o
n
d
s
tep
is
to
d
o
s
o
m
e
p
r
eli
m
in
ar
y
p
r
o
ce
s
s
in
g
o
n
t
h
e
d
ata.
Fu
r
th
er
m
o
r
e,
b
y
u
s
in
g
G
L
C
M
an
d
s
t
atis
tical
f
ea
tu
r
e
ex
tr
ac
tio
n
,
w
e
w
er
e
ab
le
to
ex
tr
a
ct
th
ir
teen
d
is
ti
n
ct
ch
ar
ac
ter
is
tic
s
f
r
o
m
o
n
e
p
ictu
r
e.
T
h
er
e
ar
e
tw
o
d
is
ti
n
ct
f
ea
t
u
r
es
s
elec
tio
n
p
r
o
ce
d
u
r
es
th
at
a
r
e
u
s
ed
in
o
r
d
e
r
t
o
ascer
tain
t
h
e
m
u
t
u
al
s
co
r
e
o
f
ea
ch
f
ea
t
u
r
e.
T
h
ese
s
tr
ate
g
ies
in
cl
u
d
e
an
al
y
s
i
s
o
f
v
ar
ian
ce
(
A
N
OV
A
)
an
d
leas
t
ab
s
o
lu
te
s
h
r
in
k
a
g
e
s
elec
t
io
n
o
p
er
ato
r
(
L
ASSO)
.
Fo
llo
w
i
n
g
t
h
e
co
m
p
le
tio
n
o
f
th
e
p
r
o
ce
s
s
o
f
s
elec
t
in
g
f
ea
t
u
r
es,
th
e
to
p
ten
f
ea
tu
r
es
w
er
e
s
elec
ted
f
o
r
f
u
r
th
er
ex
a
m
i
n
atio
n
.
C
o
n
tr
ast
(
C
ON)
,
co
r
r
elatio
n
(
C
OR
)
,
s
k
e
w
n
es
s
(
SKEN
)
,
k
u
r
to
s
is
(
KT
S),
v
ar
ian
ce
(
VAR),
s
ta
n
d
ar
d
d
ev
iatio
n
(
ST
D)
,
en
tr
o
p
y
(
E
NT
)
,
en
er
g
y
(E
N
G)
,
m
ea
n
(
MN
)
,
an
d
h
o
m
o
g
e
n
eit
y
(
HGN)
ar
e
s
o
m
e
o
f
th
e
q
u
al
ities
t
h
at
a
r
e
in
c
lu
d
ed
in
th
is
ca
te
g
o
r
y
.
Af
ter
t
h
at,
w
e
d
i
v
id
ed
th
e
d
ataset
i
n
to
t
w
o
h
alv
e
s
,
ap
p
lied
alg
o
r
ith
m
s
,
an
d
u
s
ed
s
e
v
e
n
d
i
f
f
er
en
t
p
er
f
o
r
m
an
c
e
ass
es
s
m
en
t i
n
d
icato
r
s
to
ev
al
u
ate
ea
ch
tech
n
iq
u
e.
T
h
e
o
v
er
al
l
w
o
r
k
i
n
g
p
r
o
ce
d
u
r
e
is
p
r
esen
t
ed
in
Fig
u
r
e
1
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
A
co
mp
r
eh
en
s
ive
a
n
a
lysi
s
o
f fe
a
tu
r
e
s
elec
tio
n
a
n
d
X
A
I
fo
r
ma
ch
in
e
lea
r
n
in
g
…
(
S
u
jo
n
C
h
a
n
d
r
a
S
u
tr
a
d
h
a
r
)
577
Fig
u
r
e
1
.
W
o
r
k
in
g
p
r
o
ce
d
u
r
e
f
o
r
gua
v
a
d
is
ea
s
e
clas
s
if
icatio
n
3
.
1
.
Da
t
a
d
escript
io
n
I
n
th
is
r
esear
ch
,
th
e
S
a
m
s
u
n
g
S2
4
s
m
ar
tp
h
o
n
e
eq
u
ip
p
ed
w
i
t
h
a
m
ai
n
ca
m
er
a
th
at
h
a
s
a
r
eso
lu
tio
n
o
f
5
0
m
e
g
ap
ix
e
ls
,
a
s
to
r
ag
e
ca
p
ac
it
y
o
f
1
2
8
g
ig
ab
y
tes,
a
n
d
8
g
ig
ab
y
te
s
o
f
r
a
n
d
o
m
ac
ce
s
s
m
e
m
o
r
y
(
R
AM
)
w
as
u
s
ed
to
co
llect
th
e
i
m
a
g
es
as
w
ell
a
s
f
o
r
th
e
i
m
a
g
e
ac
q
u
i
s
itio
n
p
r
o
ce
s
s
.
2
3
9
d
i
s
ea
s
e
-
f
r
ee
i
m
ag
e
s
an
d
2
8
7
d
is
ea
s
e
-
a
f
f
ec
ted
i
m
a
g
e
s
ar
e
am
o
n
g
th
e
to
tal
o
f
5
2
6
g
u
a
v
a
i
m
a
g
es
t
h
at
h
a
v
e
b
ee
n
g
at
h
er
ed
f
r
o
m
t
h
e
g
u
a
v
a
f
ield
.
T
h
e
o
v
er
al
l d
ata
d
is
tr
ib
u
tio
n
is
d
ep
icted
in
T
ab
le
1
.
T
ab
le
1
.
Data
f
r
eq
u
en
c
y
f
o
r
ac
q
u
is
itio
n
o
f
i
m
ag
e
C
l
a
ss
B
i
n
a
r
y
N
u
m.
o
f
i
mag
e
s
S
a
mp
l
e
i
mag
e
D
i
se
a
se
-
a
f
f
e
c
t
e
d
1
2
8
7
D
i
se
a
se
-
f
r
e
e
0
2
3
9
T
o
t
a
l
20
5
2
6
1
5
.
3
3
.
2
.
I
m
a
g
e
prepro
ce
s
s
ing
I
m
ag
e
p
r
ep
r
o
ce
s
s
in
g
is
es
s
en
tial
f
o
r
ef
f
ec
ti
v
e
co
m
p
u
ter
v
is
io
n
tas
k
s
.
I
n
itiall
y
,
o
u
r
i
m
ag
es
w
er
e
ca
p
tu
r
ed
at
a
r
eso
lu
tio
n
o
f
2160×
2
1
6
0
p
ix
els
an
d
r
esized
to
3
0
0
×3
0
0
p
ix
els
u
s
i
n
g
th
e
b
ilin
ea
r
in
ter
p
o
latio
n
m
et
h
o
d
[
1
6
]
,
w
h
ic
h
ca
lcu
la
tes
p
ix
el
v
al
u
es
in
t
h
e
c
o
m
p
r
ess
e
d
im
a
g
e
b
y
i
n
ter
p
o
latin
g
d
is
ta
n
ce
s
b
et
w
ee
n
f
o
u
r
n
eig
h
b
o
r
in
g
p
ix
els
u
s
i
n
g
th
e
f
o
r
m
u
la
(
)
=
1
−
|
|
f
o
r
|
|
≤
1
an
d
0
o
th
er
w
is
e.
So
m
e
i
m
a
g
es
co
n
tai
n
ed
n
o
is
e,
w
h
ic
h
w
a
s
ad
d
r
ess
ed
u
s
in
g
a
Gau
s
s
ian
f
i
lter
to
en
h
an
ce
q
u
alit
y
v
ia
g
a
m
m
a
co
r
r
elatio
n
,
f
o
llo
w
ed
b
y
h
is
to
g
r
a
m
eq
u
aliza
tio
n
to
i
m
p
r
o
v
e
co
n
tr
ast
[
1
7
]
.
T
h
is
p
r
o
ce
s
s
n
o
r
m
alize
s
th
e
i
m
a
g
e
i
n
te
n
s
it
y
r
a
n
g
e
(
0
to
1
)
,
tr
an
s
f
o
r
m
i
n
g
t
h
e
i
n
te
n
s
it
y
d
is
t
r
ib
u
tio
n
d
en
s
it
y
f
u
n
c
tio
n
(
)
to
a
u
n
i
f
o
r
m
d
e
n
s
it
y
o
f
o
n
e
p
o
s
t
-
e
q
u
aliza
tio
n
.
3
.
3
.
Seg
m
ent
a
t
io
n
T
h
is
r
esear
ch
u
s
ed
th
e
K
-
m
ea
n
s
clu
s
ter
i
n
g
alg
o
r
it
h
m
,
s
u
p
p
l
e
m
en
ted
b
y
b
o
u
n
d
ar
y
an
d
s
p
o
t
d
etec
tio
n
tech
n
i
q
u
es,
to
s
eg
m
e
n
t
t
h
e
p
i
ctu
r
es
[
1
8
]
.
T
h
e
8
-
co
n
n
ec
ted
p
ix
el
ap
p
r
o
ac
h
is
u
tili
ze
d
f
o
r
b
o
u
n
d
ar
y
d
etec
tio
n
.
E
u
clid
ea
n
d
is
ta
n
ce
is
e
m
p
lo
y
ed
f
o
r
K
-
m
ea
n
s
clu
s
ter
i
n
g
in
th
is
i
n
s
ta
n
ce
.
T
h
e
K
-
m
ea
n
s
cl
u
s
ter
i
n
g
tec
h
n
iq
u
e
p
r
im
ar
il
y
a
s
s
i
g
n
s
ea
ch
p
i
x
el
in
t
h
e
i
m
ag
e
to
t
h
e
cl
u
s
ter
w
it
h
t
h
e
m
i
n
i
m
u
m
d
is
ta
n
ce
f
r
o
m
t
h
e
cl
u
s
ter
’
s
ce
n
tr
o
id
,
s
u
b
s
eq
u
e
n
tl
y
p
er
f
o
r
m
s
co
lo
r
s
eg
m
e
n
tatio
n
o
n
t
h
e
i
m
ag
e,
a
n
d
u
lti
m
atel
y
s
ele
c
ts
th
e
cl
u
s
ter
t
h
at
ex
clu
s
i
v
el
y
co
n
tai
n
s
r
eg
io
n
s
o
f
in
ter
est
(
R
OI
s
)
.
W
e
h
av
e
u
s
ed
th
e
K
-
m
ea
n
s
cl
u
s
ter
i
n
g
a
p
p
r
o
ac
h
to
s
eg
m
e
n
t
g
u
a
v
a
i
m
a
g
e
d
ata
in
to
s
m
aller
p
iece
s
i
n
t
h
is
p
ap
er
.
K
-
m
ea
n
s
is
a
n
u
n
s
u
p
er
v
is
ed
m
et
h
o
d
u
s
ed
to
f
i
n
d
s
ep
ar
ate
g
r
o
u
p
s
i
n
t
h
e
d
ata
b
ase
d
o
n
h
o
w
s
i
m
ilar
t
h
e
d
ata
i
s
.
T
h
is
i
s
o
n
e
o
f
th
e
m
o
s
t
u
s
ed
cl
u
s
ter
i
n
g
alg
o
r
ith
m
s
,
w
h
er
e
k
s
ta
n
d
s
f
o
r
th
e
n
u
m
b
er
o
f
cl
u
s
ter
s
.
W
e
h
av
e
ch
o
s
en
=
3
f
o
r
t
h
i
s
t
a
s
k
t
o
s
eg
m
en
t
an
im
a
g
e
,
w
h
ich
m
e
an
s
t
h
at
i
t
w
il
l
f
in
d
3
g
r
o
u
p
s
in
th
e
im
ag
e
.
T
h
e
K
-
m
ea
n
s
cl
u
s
t
e
r
in
g
a
lg
o
r
i
t
h
m
w
o
r
k
s
w
el
l
o
n
a
l
im
it
e
d
s
e
t
o
f
d
a
t
a
[
1
9
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
24
,
No
.
2
,
A
p
r
il
20
26
:
5
7
4
-
587
578
3
.
4
.
E
x
t
ra
ct
io
n o
f
r
elev
a
nt
f
ea
t
ures
I
m
ag
e
s
ar
e
r
ep
r
esen
ted
as
lar
g
e
p
ix
el
m
atr
ice
s
,
f
r
o
m
w
h
ich
d
i
m
en
s
io
n
al
it
y
ca
n
b
e
r
ed
u
ce
d
th
r
o
u
g
h
f
ea
t
u
r
e
ex
tr
ac
tio
n
[
2
0
]
.
I
n
th
i
s
r
esear
ch
,
t
h
ir
tee
n
f
ea
tu
r
e
s
w
er
e
e
x
tr
ac
ted
,
in
cl
u
d
i
n
g
G
L
C
M
m
ea
s
u
r
es
a
n
d
s
tatis
t
ical
d
escr
ip
to
r
s
:
C
ON,
C
OR
,
SKE
N,
KT
S,
VA
R
,
S
T
D
,
E
N
T
,
E
N
G,
MN
,
HGN,
r
o
o
t
m
ea
n
s
q
u
ar
e
(
R
MS)
,
s
m
o
o
th
n
es
s
(
SM)
,
an
d
in
v
er
s
e
d
if
f
er
en
ce
m
o
m
e
n
t
(
I
DM
)
.
Fo
r
t
w
o
p
ix
e
ls
at
(
,
)
s
ep
ar
ated
b
y
d
is
tan
ce
d
an
d
an
g
le
α
,
th
e
G
L
C
M
ca
n
b
e
d
ef
i
n
ed
as:
(
,
,
,
)
=
{
(
1
,
1
)
,
(
2
,
2
)
∈
×
∶
,
,
(
1
,
1
)
=
,
(
2
,
2
)
=
}
(
1
)
I
n
t
h
is
f
o
r
m
u
latio
n
,
(
,
)
r
ep
r
esen
ts
t
h
e
(
,
)
−
ℎ
ele
m
e
n
t
o
f
t
h
e
co
m
p
u
ted
GL
C
M,
is
t
h
e
n
u
m
b
er
o
f
g
r
a
y
le
v
el
s
,
an
d
µ
,
µ
,
,
ar
e
th
e
m
ea
n
s
a
n
d
s
tan
d
ar
d
d
ev
iatio
n
s
o
f
r
o
w
a
n
d
co
lu
m
n
s
u
m
s
.
T
h
e
m
ai
n
G
L
C
M
-
b
ased
f
ea
t
u
r
es a
r
e
d
ef
in
ed
as:
=
∑
∑
(
−
)
2
(
,
)
,
−
1
=
0
−
1
=
0
(
2
)
=
∑
∑
.
.
(
,
)
−
−
1
=
0
−
1
=
0
,
(
3
)
=
∑
∑
(
,
)
2
,
−
1
=
0
−
1
=
0
(
4
)
=
−
∑
∑
(
,
)
l
og
(
,
)
,
−
1
=
0
−
1
=
0
(
5
)
=
∑
∑
(
,
)
1
+
(
j
−
k
)
2
,
−
1
=
0
−
1
=
0
(
6
)
=
∑
∑
(
,
)
1
+
(
j
−
k
)
2
(
7
)
I
n
ad
d
itio
n
to
GL
C
M
f
ea
t
u
r
es,
s
tatis
t
ical
d
escr
ip
to
r
s
w
e
r
e
ex
tr
ac
ted
.
Fo
r
a
s
et
o
f
p
ix
els
w
i
th
in
te
n
s
it
y
,
MN
,
ST
D,
VA
R
,
K
T
S,
an
d
s
k
e
w
n
es
s
(
SKEN
)
ar
e
d
ef
in
ed
as:
=
1
∑
=
1
(
8
)
=
√
∑
(
−
)
2
=
1
(
9
)
=
1
∑
(
−
)
2
=
1
(
1
0
)
=
1
∑
(
−
)
4
=
1
(
1
∑
(
−
)
2
=
1
)
−
3
(
1
1
)
=
−
(
1
2
)
Her
e,
C
is
th
e
m
ea
n
in
te
n
s
it
y
,
G
th
e
p
ix
el
co
u
n
t a
t a
g
iv
e
n
i
n
ten
s
i
t
y
,
a
n
d
Q
th
e
n
o
r
m
aliza
ti
o
n
f
ac
to
r
.
T
h
ese
m
ea
s
u
r
es c
h
ar
ac
ter
ize
in
te
n
s
i
t
y
d
is
tr
ib
u
tio
n
s
i
n
b
o
th
d
ef
ec
ti
v
e
an
d
d
ef
ec
t
-
f
r
ee
r
eg
io
n
s
o
f
g
r
a
y
s
ca
le
i
m
a
g
es.
3
.
5
.
Descript
io
n o
f
f
e
a
t
ure
s
elec
t
io
n
t
ec
hn
iqu
e
s
Featu
r
e
s
elec
tio
n
w
as
ap
p
lied
to
r
etain
r
ele
v
an
t a
ttrib
u
tes a
n
d
r
e
m
o
v
e
r
ed
u
n
d
an
c
y
[
2
1
]
.
T
h
e
m
et
h
o
d
s
u
s
ed
f
o
r
r
an
k
in
g
f
ea
tu
r
e
s
w
er
e
A
NOV
A
a
n
d
L
ASS
O.
A
l
l
t
h
r
ee
m
etr
ics
s
u
c
h
as
R
MS,
s
m
o
o
t
h
n
es
s
,
an
d
I
DM
ar
e
am
p
lit
u
d
e
-
b
a
s
ed
m
ea
s
u
r
es
o
f
th
e
s
a
m
e
u
n
d
er
l
y
i
n
g
s
u
r
f
a
ce
s
ig
n
al,
o
f
ten
li
n
ea
r
l
y
r
elate
d
o
r
d
er
iv
ed
f
r
o
m
s
q
u
ar
ed
d
ev
iatio
n
s
.
Un
less
t
h
e
s
u
r
f
ac
e
p
r
o
f
ile
v
ar
ies
w
id
el
y
i
n
f
r
eq
u
e
n
c
y
co
n
ten
t,
t
h
e
y
w
il
l
r
esp
o
n
d
alm
o
s
t
id
en
ticall
y
[
2
2
]
an
d
w
er
e
e
x
cl
u
d
ed
,
leav
in
g
th
e
to
p
ten
f
ea
t
u
r
es:
C
NT
,
C
R
L
,
SKEN
,
KT
S,
VAR,
ST
D,
E
N
T
,
E
G,
MN
,
an
d
HGN.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
A
co
mp
r
eh
en
s
ive
a
n
a
lysi
s
o
f fe
a
tu
r
e
s
elec
tio
n
a
n
d
X
A
I
fo
r
ma
ch
in
e
lea
r
n
in
g
…
(
S
u
jo
n
C
h
a
n
d
r
a
S
u
tr
a
d
h
a
r
)
579
3
.
5
.
1
.
ANO
VA
A
N
O
VA
is
a
s
t
a
ti
s
t
i
c
a
l
m
eth
o
d
t
h
a
t
i
s
u
s
e
d
t
o
in
v
es
t
ig
at
e
w
h
e
th
e
r
t
h
e
r
e
is
an
y
e
q
u
al
v
a
r
i
a
ti
o
n
am
o
n
g
g
r
o
u
p
s
o
f
c
a
t
eg
o
r
i
ca
l
v
a
r
i
a
b
le
s
t
h
a
t
p
e
r
t
a
in
t
o
n
u
m
e
r
ic
a
l
r
es
p
o
n
s
e
.
A
NO
VA
c
a
n
b
e
r
e
p
r
e
s
en
te
d
b
y
(
1
3
)
:
=
/
(
−
1
)
/
(
−
)
(
1
3
)
A
N
OV
A
tes
ts
w
h
et
h
er
g
r
o
u
p
m
ea
n
s
d
if
f
er
s
i
g
n
i
f
ican
t
l
y
(
1
3
)
,
w
h
er
e
an
d
ar
e
th
e
b
et
w
ee
n
-
g
r
o
u
p
an
d
w
ith
i
n
-
g
r
o
u
p
v
ar
ian
ce
s
,
K
is
th
e
n
u
m
b
er
o
f
clas
s
es,
an
d
n
is
th
e
s
a
m
p
le
s
ize
[
2
3
]
.
3
.
5
.
2
.
L
ASSO
L
A
S
SO
is
a
r
eg
r
ess
io
n
-
b
ase
d
f
ea
tu
r
e
s
elec
tio
n
tec
h
n
iq
u
e
th
at
p
en
alize
s
t
h
e
s
u
m
o
f
ab
s
o
lu
te
co
ef
f
icie
n
t
v
al
u
es to
p
r
ev
en
t o
v
er
f
it
tin
g
(
1
4
)
,
w
it
h
α
as th
e
r
e
g
u
lar
iza
tio
n
p
ar
a
m
eter
[
2
4
]
.
=
1
2
|
|
−
|
|
2
+
|
|
|
|
1
(
1
4
)
3
.
6
.
Cla
s
s
if
ie
r
t
ra
ini
ng
a
nd
t
esting
I
n
th
i
s
s
ec
t
io
n
,
t
h
r
ee
s
p
litt
i
n
g
r
atio
s
o
f
th
e
d
ata
s
et,
s
u
ch
as
tr
ain
(
6
0
%),
v
alid
atio
n
(
2
0
%
)
,
an
d
test
(
2
0
%),
ar
e
u
s
ed
.
T
r
ain
in
g
d
at
a
is
u
s
ed
f
o
r
tr
ain
i
n
g
a
ML
m
o
d
el,
w
h
er
e
test
d
ata
e
v
al
u
a
tes
th
e
e
f
f
icie
n
c
y
o
f
ea
ch
m
o
d
e
l.
C
o
m
p
ar
i
n
g
t
h
e
p
e
r
f
o
r
m
an
ce
s
o
f
v
ar
io
u
s
tr
ai
n
ed
m
o
d
el
s
is
m
ea
s
u
r
ed
w
i
th
t
h
e
v
alid
atio
n
d
ata.
3
.
7
.
H
y
perpa
ra
m
et
er
t
un
ing
A
t
y
p
e
o
f
p
ar
a
m
eter
w
h
o
s
e
v
alu
e
i
s
d
eter
m
i
n
ed
p
r
io
r
to
alg
o
r
ith
m
tr
ain
i
n
g
is
k
n
o
w
n
as
a
h
y
p
er
p
ar
a
m
eter
.
Fi
n
e
-
tu
n
i
n
g
a
p
ar
am
eter
i
s
a
w
ell
-
u
s
ed
te
ch
n
iq
u
e
f
o
r
u
p
g
r
ad
in
g
t
h
e
ac
cu
r
ac
y
o
f
th
e
ML
m
o
d
el.
I
n
t
h
is
r
esear
ch
,
T
ab
le
2
in
cl
u
d
es
all
th
e
e
m
p
lo
y
ed
a
lg
o
r
ith
m
’
s
h
y
p
er
p
ar
a
m
e
ter
v
al
u
es.
Her
e,
L
B
FG
S
m
ea
n
s
li
m
it
ed
-
m
e
m
o
r
y
B
r
o
y
d
en
–
Fletch
er
–
Go
ld
f
ar
b
-
Sh
a
n
n
o
,
g
in
i
r
ef
er
s
to
g
i
n
i
i
m
p
u
r
it
y
,
an
d
r
ad
ial
b
asis
f
u
n
ctio
n
(
R
B
F
)
.
T
ab
le
2
.
Hy
p
er
p
ar
am
e
ter
s
o
f
c
lass
i
f
ier
s
u
s
ed
f
o
r
g
u
a
v
a
d
is
ea
s
e
r
ec
o
g
n
itio
n
M
o
d
e
l
H
y
p
e
r
p
a
r
a
me
t
e
r
V
a
l
u
e
A
d
a
B
o
o
st
e
st
i
ma
t
o
r
s
1
0
0
M
i
n
s
a
mp
l
e
s l
e
a
f
1
C
r
i
t
e
r
i
o
n
G
i
n
i
M
i
n
s
a
mp
l
e
s sp
l
i
t
2
M
a
x
f
e
a
t
u
r
e
s
S
q
u
a
r
e
r
o
o
t
(s
q
r
t
)
DT
M
i
n
s
a
mp
l
e
s l
e
a
f
1
C
r
i
t
e
r
i
o
n
g
i
n
i
M
i
n
s
a
mp
l
e
s sp
l
i
t
2
K
N
N
n
e
i
g
h
b
o
r
s
5
W
e
i
g
h
t
U
n
i
f
o
r
m
A
l
g
o
r
i
t
h
m
A
u
t
o
L
e
a
f
si
z
e
30
P
2
M
e
t
r
i
c
M
i
n
k
o
w
sk
i
S
V
M
C
1
K
e
r
n
e
l
R
BF
G
a
mm
a
S
c
a
l
e
RF
e
st
i
ma
t
o
r
s
1
0
0
M
i
n
s
a
mp
l
e
s l
e
a
f
1
C
r
i
t
e
r
i
o
n
G
i
n
i
M
i
n
s
a
mp
l
e
s sp
l
i
t
2
M
a
x
f
e
a
t
u
r
e
s
A
u
t
o
LR
C
1
M
a
x
_
i
t
e
r
a
t
i
o
n
1
0
0
0
S
o
l
v
e
r
L
B
F
G
S
3
.
8
.
Descript
io
n o
f
u
t
ilized X
AI
t
ec
hn
i
qu
es
XA
I
is
cr
u
ci
a
l
f
o
r
elu
cid
ati
n
g
A
I
m
o
d
el
f
u
n
ctio
n
a
lit
y
,
a
n
ticip
ated
im
p
ac
t
s
,
an
d
p
o
ten
tial
b
iases
,
en
s
u
r
in
g
ac
cu
r
ac
y
,
f
a
ir
n
es
s
,
an
d
tr
an
s
p
ar
en
c
y
i
n
A
I
-
d
r
iv
en
d
ec
is
io
n
-
m
a
k
i
n
g
.
I
n
th
i
s
ar
ticle
,
tw
o
m
o
s
t
ef
f
ec
tiv
e
ex
p
lai
n
ab
le
tech
n
iq
u
es
L
I
ME
an
d
SHA
P
ar
e
u
s
ed
.
L
I
ME
an
d
SH
A
P
ar
e
m
o
d
el
-
ag
n
o
s
tic
XA
I
to
o
ls
th
at
d
i
f
f
er
e
n
tl
y
attr
ib
u
te
a
s
i
n
g
le
p
r
ed
ictio
n
to
it
s
i
n
p
u
t
f
ea
t
u
r
es.
W
h
ile
L
I
ME
b
u
ild
s
a
s
i
m
p
le,
in
ter
p
r
etab
le
s
u
r
r
o
g
ate
(
o
f
te
n
a
s
p
ar
s
e
l
in
ea
r
m
o
d
el)
ar
o
u
n
d
th
e
s
p
ec
i
f
ic
i
n
s
ta
n
ce
b
y
p
er
tu
r
b
i
n
g
its
f
ea
t
u
r
es
an
d
w
ei
g
h
ti
n
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
24
,
No
.
2
,
A
p
r
il
20
26
:
5
7
4
-
587
580
ne
ar
b
y
s
a
m
p
les,
y
ield
i
n
g
q
u
i
ck
,
h
u
m
a
n
-
r
ea
d
ab
le
lo
ca
l
im
p
o
r
tan
ce
s
co
r
es
th
at
ca
n
v
ar
y
w
it
h
p
er
tu
r
b
atio
n
ch
o
ices.
E
s
s
e
n
tiall
y
,
L
I
ME
p
er
tu
r
b
s
th
e
in
p
u
t
f
ea
t
u
r
es
ar
o
u
n
d
a
g
iv
e
n
s
a
m
p
le
an
d
lear
n
s
a
lo
ca
l
lin
ea
r
ap
p
r
o
x
im
a
tio
n
o
f
t
h
e
b
lack
-
b
o
x
m
o
d
el,
f
r
o
m
w
h
ic
h
it
d
er
iv
e
s
f
ea
t
u
r
e
i
m
p
o
r
tan
c
e
f
o
r
th
at
p
ar
ticu
lar
p
r
ed
ictio
n
[
2
5
]
.
SHA
P
u
s
es
g
a
m
e
-
t
h
eo
r
etic
Sh
ap
le
y
v
alu
e
s
to
d
is
tr
ib
u
te
th
e
p
r
ed
ictio
n
(
r
elativ
e
to
a
b
aselin
e)
f
air
l
y
ac
r
o
s
s
f
ea
tu
r
es,
o
f
f
er
in
g
co
n
s
is
ten
t
attr
ib
u
tio
n
s
th
at
s
u
m
to
th
e
p
r
ed
ictio
n
d
if
f
er
en
ce
an
d
s
ca
lin
g
ef
f
icien
tl
y
f
o
r
tr
ee
m
o
d
els
v
ia
T
r
ee
SH
A
P
,
th
o
u
g
h
it
d
ep
en
d
s
o
n
t
h
e
c
h
o
s
en
b
ac
k
g
r
o
u
n
d
d
ata
a
n
d
ca
n
b
e
co
m
p
u
tatio
n
a
ll
y
h
ea
v
ier
.
SH
A
P
co
n
s
id
er
s
al
l p
o
s
s
ib
le
co
m
b
i
n
atio
n
s
o
f
f
ea
t
u
r
es a
n
d
allo
ca
te
s
cr
ed
it in
a
m
a
n
n
er
t
h
at
is
f
air
a
n
d
m
at
h
e
m
a
ticall
y
r
ig
o
r
o
u
s
,
en
s
u
r
in
g
t
h
at
th
e
s
u
m
o
f
attr
ib
u
t
io
n
s
eq
u
als
t
h
e
d
i
f
f
er
e
n
ce
b
et
wee
n
t
h
e
p
r
ed
ictio
n
an
d
th
e
d
ataset
b
asel
in
e
[
2
5
]
.
I
n
s
h
o
r
t,
L
I
ME
is
s
u
itab
le
f
o
r
f
ast,
ap
p
r
o
x
i
m
ate
lo
ca
l in
s
i
g
h
t,
an
d
SH
A
P
is
m
o
r
e
s
u
itab
le
f
o
r
p
r
in
cip
led
,
ad
d
itiv
e
attr
ib
u
tio
n
s
o
f
b
o
th
lo
ca
l a
n
d
g
lo
b
al
an
al
y
s
i
s.
3
.
9
.
P
er
f
o
rm
a
nce
ev
a
lua
t
io
n
m
et
ric
s
P
er
f
o
r
m
a
n
ce
ev
al
u
atio
n
m
atr
i
ce
s
ar
e
ess
en
tia
l
f
o
r
ev
al
u
ati
n
g
t
h
e
p
er
f
o
r
m
a
n
ce
o
f
a
m
o
d
el.
I
n
th
is
r
esear
ch
,
7
p
er
f
o
r
m
an
ce
m
a
tr
ices
s
u
ch
a
s
ac
c
u
r
ac
y
(
AC
C
)
,
s
en
s
it
iv
i
t
y
(
SE
N)
,
s
p
ec
if
ic
it
y
(
SP
E
)
,
f
alse
p
o
s
iti
v
e
r
ate
(
FP
R
)
,
f
alse
n
e
g
ati
v
e
r
ate
(
FNR
)
,
F1
-
Sco
r
e,
an
d
P
r
ec
is
i
o
n
ar
e
u
s
ed
.
All
t
h
e
eq
u
at
io
n
s
f
r
o
m
(
1
5
)
to
(
2
1
)
ar
e
u
s
ed
.
=
(
+
+
+
+
)
×
100%
(
1
5
)
=
(
+
)
×
100%
(
1
6
)
=
(
+
)
×
100%
(
1
7
)
=
(
+
)
×
100%
(
1
8
)
=
(
+
)
×
100%
(
1
9
)
=
(
+
)
×
100%
(
2
0
)
1
−
=
(
2
×
×
+
)
×
100%
(
2
1
)
4.
RE
SU
L
T
S
AND
D
I
SCU
SS
I
O
N
T
h
e
g
ath
er
in
g
o
f
i
m
a
g
e
d
ata
w
as
th
e
f
ir
s
t
s
ta
g
e
in
th
e
p
r
o
ce
s
s
o
f
ac
h
iev
i
n
g
t
h
e
w
o
r
k
w
i
th
th
e
id
en
ti
f
icatio
n
o
f
g
u
a
v
a
ill
n
es
s
es.
Up
o
n
th
e
en
d
o
f
th
e
p
ict
u
r
e
co
llectio
n
p
r
o
ce
s
s
,
th
e
i
m
a
g
es
t
h
at
h
a
v
e
b
ee
n
ac
q
u
ir
ed
ar
e
s
u
b
s
eq
u
e
n
tl
y
s
ca
l
ed
to
a
r
eso
lu
tio
n
o
f
3
0
0
×3
0
0
p
ix
els.
I
n
cr
ea
s
in
g
t
h
e
co
n
tr
as
t
o
f
th
e
p
ic
t
u
r
es t
h
at
ar
e
s
h
o
w
n
in
Fig
u
r
e
2
is
ac
co
m
p
lis
h
ed
b
y
th
e
u
s
e
o
f
th
e
m
ap
p
in
g
o
f
co
lo
u
r
in
ten
s
i
t
y
.
A
f
ter
th
en
,
th
e
co
lo
u
r
p
ictu
r
es
ar
e
d
iv
id
ed
u
p
in
to
a
n
u
m
b
er
o
f
d
if
f
er
e
n
t
clu
s
ter
s
.
T
h
e
K
-
m
ea
n
s
cl
u
s
ter
in
g
al
g
o
r
ith
m
,
w
h
ich
i
s
s
ee
n
in
F
ig
u
r
e
3
,
is
u
s
ed
i
n
o
r
d
er
to
ca
r
r
y
o
u
t
t
h
i
s
s
e
g
m
en
ta
tio
n
.
I
t
h
a
s
b
ee
n
s
h
o
w
n
th
at
K
-
m
ea
n
s
cl
u
s
ter
in
g
i
s
s
u
p
er
io
r
to
o
th
er
tech
n
iq
u
es
t
h
at
ar
e
u
s
ed
f
o
r
s
eg
m
en
tatio
n
.
I
n
o
r
d
e
r
to
g
et
th
e
f
ea
t
u
r
e
v
e
cto
r
s
th
at
ar
e
s
h
o
w
n
in
Fi
g
u
r
e
4
,
w
e
e
x
tr
ac
ted
t
h
e
m
f
r
o
m
ea
c
h
clu
s
ter
.
Fo
llo
w
i
n
g
th
e
co
n
clu
s
io
n
o
f
t
h
e
s
e
g
m
en
tatio
n
p
r
o
ce
s
s
,
w
e
r
etr
iev
ed
f
ea
tu
r
e
v
ec
to
r
s
f
r
o
m
ea
ch
s
eg
m
e
n
ted
i
m
ag
e.
T
h
ese
f
ea
t
u
r
e
v
ec
to
r
s
w
er
e
th
en
u
s
ed
f
o
r
th
e
tr
ain
in
g
o
f
th
e
cla
s
s
i
f
ier
s
.
ANOV
A
a
n
d
L
A
S
SO
ar
e
t
h
e
t
w
o
f
ea
t
u
r
e
-
se
lectio
n
ap
p
r
o
ac
h
es
t
h
at
ar
e
u
s
ed
o
n
ce
t
h
e
f
ea
tu
r
e
ex
tr
ac
tio
n
p
r
o
ce
s
s
h
as
b
ee
n
co
m
p
leted
.
T
h
ese
tech
n
iq
u
e
s
ar
e
u
tili
s
ed
to
estab
lis
h
t
h
e
r
ele
v
a
n
ce
o
f
f
ea
t
u
r
es
b
y
tak
i
n
g
i
n
to
co
n
s
id
er
atio
n
t
h
e
r
an
k
v
al
u
es o
f
t
h
e
f
ea
tu
r
e
s
.
Up
o
n
co
m
p
letio
n
o
f
th
e
f
ea
tu
r
e
s
elec
tio
n
,
w
e
d
is
co
v
er
ed
th
at
th
r
ee
ch
ar
ac
ter
is
tics
(
R
MS,
s
m
o
o
th
n
e
s
s
,
I
DM
)
h
ad
v
al
u
es
th
at
ar
e
al
m
o
s
t
id
en
tical
to
o
n
e
an
o
t
h
er
.
A
s
a
r
esu
lt,
w
e
e
x
clu
d
e
th
e
s
e
f
ea
t
u
r
es
b
ased
o
n
th
e
f
ea
tu
r
e
r
an
k
in
g
s
co
r
e
in
o
r
d
er
to
r
eso
lv
e
th
e
d
u
p
licatio
n
co
n
ce
r
n
s
,
an
d
w
e
u
lt
i
m
a
tel
y
c
h
o
o
s
e
th
e
to
p
1
0
f
ea
tu
r
es
to
co
n
d
u
ct
t
h
is
r
esear
ch
.
B
o
th
th
e
A
N
OV
A
a
n
d
L
ASSO
f
ea
t
u
r
e
s
elec
tio
n
s
ar
e
r
ep
r
esen
ted
b
y
th
eir
r
esp
ec
tiv
e
m
u
t
u
al
s
co
r
e
s
in
T
ab
les
3
an
d
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
A
co
mp
r
eh
en
s
ive
a
n
a
lysi
s
o
f fe
a
tu
r
e
s
elec
tio
n
a
n
d
X
A
I
fo
r
ma
ch
in
e
lea
r
n
in
g
…
(
S
u
jo
n
C
h
a
n
d
r
a
S
u
tr
a
d
h
a
r
)
581
Fig
u
r
e
2
.
C
o
n
tr
ast e
n
h
a
n
ce
m
e
n
t f
r
o
m
th
e
o
r
ig
in
al
i
m
a
g
e
Fig
u
r
e
3
.
E
n
h
a
n
ce
d
i
m
a
g
e
s
e
g
m
en
tatio
n
u
s
i
n
g
K
-
m
ea
n
s
cl
u
s
ter
in
g
Fig
u
r
e
4
.
Sa
m
p
le
o
f
ex
tr
ac
ted
f
ea
t
u
r
es
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
24
,
No
.
2
,
A
p
r
il
20
26
:
5
7
4
-
587
582
T
ab
le
3
.
Featu
r
e
r
an
k
i
n
g
s
an
d
s
co
r
es f
o
r
A
N
OV
A
f
ea
t
u
r
e
s
elec
tio
n
T
ab
le
4
.
Featu
r
e
r
an
k
i
n
g
s
an
d
s
co
r
es f
o
r
L
ASSO
f
ea
t
u
r
e
s
elec
tio
n
R
a
n
k
i
n
g
F
e
a
t
u
r
e
S
c
o
r
e
R
a
n
k
i
n
g
F
e
a
t
u
r
e
S
c
o
r
e
1
H
G
N
0
.
1
4
7
2
6
C
O
R
0
.
0
8
4
6
2
C
O
N
0
.
1
3
7
MN
0
.
0
8
4
2
3
V
A
R
0
.
1
2
3
7
8
S
K
EN
0
.
0
8
1
4
4
S
T
D
0
.
1
1
5
3
9
K
T
S
0
.
0
7
2
9
5
EN
T
0
.
0
9
10
EN
G
0
.
0
7
0
8
R
a
n
k
i
n
g
F
e
a
t
u
r
e
S
c
o
r
e
R
a
n
k
i
n
g
F
e
a
t
u
r
e
S
c
o
r
e
1
S
T
D
1
6
S
K
EN
0
.
3
2
8
1
2
H
G
N
0
.
9
3
6
0
7
EN
T
0
.
1
7
5
9
3
C
O
N
0
.
7
2
0
6
8
EN
G
0
.
1
1
0
1
4
MN
0
.
6
5
3
2
9
C
O
R
0
.
0
8
8
2
5
V
A
R
0
.
6
4
5
0
10
K
T
S
0
.
0
1
T
h
e
p
er
f
o
r
m
a
n
ce
o
f
s
i
x
clas
s
i
f
ier
s
h
as
b
ee
n
ca
lc
u
lated
u
s
i
n
g
t
h
e
co
n
f
u
s
io
n
m
atr
i
x
.
Ta
b
le
5
p
r
o
v
id
es
th
e
p
er
f
o
r
m
a
n
ce
m
etr
ic
s
f
o
r
s
i
x
class
if
ier
s
w
it
h
1
0
f
ea
tu
r
es.
Fro
m
T
ab
le
5
,
it
ca
n
b
e
claim
ed
th
at
th
e
h
i
g
h
est
ac
cu
r
ac
y
is
8
0
.
1
9
%,
w
h
ic
h
is
o
b
tain
ed
b
y
th
e
SVM
class
i
f
ier
.
On
th
e
o
th
er
h
a
n
d
,
th
e
lo
w
est
ac
c
u
r
ac
y
is
7
2
.
6
4
%
ac
h
iev
ed
b
y
th
e
A
d
aB
o
o
s
t
class
if
ier
.
I
t
is
clai
m
ed
f
r
o
m
m
an
y
r
e
s
ea
r
ch
ar
ticles
th
a
t
an
o
d
d
n
u
m
b
er
o
f
f
ea
t
u
r
e
s
elec
tio
n
w
o
r
k
s
w
el
l
i
n
p
r
ed
ictio
n
.
A
ls
o
,
f
r
o
m
th
e
c
o
r
r
elatio
n
h
ea
t
m
ap
s
h
o
w
n
i
n
Fig
u
r
e
5
(
a)
,
w
h
ich
ap
p
lies
th
e
ANOV
A
tech
n
iq
u
e,
an
d
Fig
u
r
e
5
(
b
)
,
w
h
ic
h
a
p
p
lies
th
e
L
ASS
O
tec
h
n
iq
u
e,
to
g
et
h
er
w
it
h
t
h
e
L
I
ME
an
d
SH
A
P
f
ea
t
u
r
e
ex
tr
ac
tio
n
tech
n
iq
u
e
s
h
o
w
n
in
Fi
g
u
r
e
6
(
a)
an
d
(
b
)
w
e
s
elec
t
th
e
to
p
s
ev
en
(
7
)
an
d
n
in
e
(
9
)
f
ea
t
u
r
es.
I
n
ad
d
itio
n
,
t
h
e
to
p
s
e
v
en
an
d
n
i
n
e
f
ea
tu
r
e
s
th
at
w
er
e
g
ai
n
ed
v
ia
t
h
e
u
s
e
o
f
A
N
OV
A
f
ea
t
u
r
e
s
elec
t
io
n
a
p
p
r
o
ac
h
es
ar
e
u
tili
z
ed
to
tr
ain
an
d
test
th
e
class
i
f
ier
th
at
is
s
h
o
w
n
in
T
ab
le
6
.
I
t
ca
n
b
e
s
ee
n
f
r
o
m
T
ab
le
6
th
at
th
e
RF
cla
s
s
i
f
ier
ac
h
iev
e
s
t
h
e
m
ax
i
m
u
m
ac
cu
r
ac
y
o
f
8
7
.
7
4
%
f
o
r
th
e
to
p
s
ev
en
c
h
ar
ac
ter
is
t
ics
u
n
d
er
co
n
s
id
er
atio
n
.
On
t
h
e
o
th
er
s
id
e,
th
e
DT
cl
ass
if
ier
ac
h
iev
e
s
t
h
e
lo
w
est
ac
c
u
r
ac
y
,
wh
ich
is
7
2
.
6
4
%.
A
t
s
o
m
e
p
o
in
t
in
ti
m
e,
th
e
ass
e
s
s
m
en
t
m
etr
ics
f
o
r
o
th
er
class
i
f
i
er
s
w
ill
b
e
s
u
f
f
icie
n
tl
y
ac
c
u
r
a
te
to
c
o
n
s
tit
u
te
th
e
o
u
tco
m
e
o
f
t
h
e
to
p
f
ea
t
u
r
e
s
e
t
.
An
o
th
er
f
ea
t
u
r
e
s
et
is
a
ls
o
m
an
ip
u
lated
b
y
ap
p
l
y
i
n
g
L
A
S
S
O
f
ea
t
u
r
e
s
elec
tio
n
tech
n
iq
u
es.
T
ab
le
7
d
ep
icts
th
e
o
v
er
all
ev
alu
at
io
n
m
e
tr
ics
f
o
r
s
ix
class
i
f
ier
s
.
Fro
m
T
ab
le
7
,
it
is
clai
m
ed
th
at
m
o
s
t
o
f
th
e
cla
s
s
i
f
ier
’
s
ac
c
u
r
ac
y
is
w
e
ll
en
o
u
g
h
co
m
p
ar
ed
to
th
e
A
NOV
A
f
ea
t
u
r
e
s
elec
ted
s
et.
T
h
e
h
ig
h
est
ac
cu
r
ac
y
is
f
o
u
n
d
f
o
r
th
e
A
d
aB
o
o
s
t
class
if
ier
,
w
h
ic
h
is
8
8
.
6
8
%.
W
h
er
ea
s
th
e
lo
w
est
ac
cu
r
ac
y
is
7
3
.
5
8
%,
th
at
o
b
tain
ed
b
y
KN
N.
C
o
m
p
ar
ed
to
th
e
f
ea
tu
r
e
s
et
o
u
tco
m
e
s
,
it
is
s
aid
t
h
at
t
h
e
b
est
p
er
f
o
r
m
a
n
ce
i
s
ac
h
ie
v
ed
b
y
th
e
A
d
aB
o
o
s
t
class
i
f
ier
b
y
u
til
izi
n
g
t
h
e
to
p
7
f
ea
tu
r
e
s
ets
o
b
tain
ed
b
y
L
A
S
SO
f
ea
t
u
r
e
tech
n
iq
u
es.
T
h
e
g
r
ea
test
ar
ea
u
n
d
er
th
e
cu
r
v
e
(
A
UC
)
v
alu
e
f
o
r
th
e
A
NO
V
A
w
as
0
.
9
3
f
o
r
A
d
aB
o
o
s
t
an
d
R
F,
w
h
i
ch
w
as
a
s
u
f
f
ic
ien
t
a
m
o
u
n
t.
A
d
d
itio
n
all
y
,
t
h
e
LR
,
SVM
,
an
d
KNN
clas
s
i
f
ier
s
wer
ea
b
le
to
r
ea
ch
th
e
g
r
ea
test
r
esu
lt
s
,
w
h
ich
w
er
e
0
.
8
6
,
0
.
8
5
,
an
d
0
.
8
1
r
esp
ec
tiv
el
y
.
L
a
s
t
b
u
t
n
o
t
lea
s
t,
th
e
lo
w
e
s
t
A
UC
v
al
u
e
th
at
ca
n
b
e
r
ea
ch
ed
u
s
i
n
g
DT
is
0
.
7
3
f
o
r
all
o
f
t
h
e
m
o
d
el
s
t
h
at
ar
e
ap
p
lied
.
I
n
th
e
L
ASSO
a
lg
o
r
ith
m
,
t
h
e
lo
g
i
s
tic
A
d
aB
o
o
s
t
class
i
f
ier
w
a
s
ab
le
to
o
b
tain
th
e
h
i
g
h
li
g
h
ted
A
UC
v
al
u
e
o
f
0
.
9
3
,
w
h
ich
is
i
n
d
ica
tiv
e
o
f
it
s
h
ig
h
le
v
el
o
f
d
is
cr
i
m
i
n
ati
n
g
.
A
n
u
m
b
er
o
f
o
th
er
cla
s
s
i
f
ier
s
,
i
n
cl
u
d
in
g
R
F,
LR
,
SVM
,
a
n
d
KNN
,
al
s
o
p
er
f
o
r
m
ed
w
ell,
w
it
h
AUC
v
alu
es
o
f
0
.
9
2
,
0
.
8
6
,
0
.
8
5
,
an
d
0
.
8
1
r
esp
ec
tiv
el
y
.
Am
o
n
g
all
o
f
th
e
m
o
d
els
th
at
wer
e
ev
alu
at
ed
,
th
e
DT
h
ad
th
e
lo
w
es
t
A
U
C
v
al
u
e,
w
h
ic
h
w
as 0
.
7
5
.
T
ab
le
5
.
P
er
f
o
r
m
a
n
ce
m
etr
ic
s
o
f
d
if
f
er
en
t a
l
g
o
r
ith
m
s
u
s
in
g
1
0
f
ea
tu
r
es
A
l
g
o
r
i
t
h
m
A
C
C
(
%)
S
EN
(
%)
S
P
E
(
%)
F
P
R
(
%)
F
N
R
(
%)
F1
-
sco
r
e
(
%)
P
r
e
c
i
si
o
n
(
%)
LR
7
8
.
3
0
9
0
.
3
8
6
6
.
6
7
3
3
.
3
3
9
.
6
2
8
0
.
3
4
7
2
.
3
1
K
N
N
7
4
.
53
8
0
.
7
7
6
8
.
5
2
3
1
.
4
8
1
9
.
2
3
7
5
.
6
7
7
1
.
1
9
DT
7
5
.
4
7
7
5
.
0
0
7
5
.
9
3
2
4
.
0
7
2
5
.
0
0
7
5
.
0
0
7
5
.
0
0
A
d
a
B
o
o
st
7
2
.
6
4
9
2
.
3
1
5
3
.
7
0
4
6
.
3
0
7
.
6
9
7
6
.
8
0
6
5
.
7
5
RF
7
9
.
2
5
8
8
.
4
6
7
0
.
3
7
2
9
.
6
3
1
1
.
5
4
8
0
.
7
4
7
4
.
1
9
S
V
M
8
0
.
1
9
9
6
.
1
5
6
4
.
8
1
3
5
.
1
9
3
.
8
5
8
2
.
6
0
7
2
.
5
0
(
a)
(
b
)
Fig
u
r
e
5
.
C
o
r
r
elatio
n
h
ea
t
m
ap
u
s
i
n
g
:
(
a)
A
NOV
A
an
d
(
b
)
L
ASSO
tec
h
n
iq
u
es
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
A
co
mp
r
eh
en
s
ive
a
n
a
lysi
s
o
f fe
a
tu
r
e
s
elec
tio
n
a
n
d
X
A
I
fo
r
ma
ch
in
e
lea
r
n
in
g
…
(
S
u
jo
n
C
h
a
n
d
r
a
S
u
tr
a
d
h
a
r
)
583
(
a)
(
b
)
Fig
u
r
e
6
.
Mo
d
el
ex
p
lain
ab
ilit
y
v
is
u
alize
d
u
s
in
g
:
(
a)
L
I
ME
a
n
d
(
b
)
SHA
P
tech
n
iq
u
e
s
T
ab
le
6
.
Pe
r
f
o
r
m
an
c
e
m
et
r
i
cs
f
o
r
d
i
f
f
e
r
en
t
al
g
o
r
i
th
m
s
u
s
in
g
7
a
n
d
9
f
e
a
tu
r
es
w
ith
A
NO
VA
s
e
l
e
c
t
i
o
n
te
ch
n
i
q
u
e
F
e
a
t
u
r
e
n
u
m
b
e
r
A
l
g
o
r
i
t
h
m
A
C
C
(
%)
S
EN
(
%)
S
P
E
(
%)
F
P
R
(
%)
F
N
R
(
%)
F1
-
sco
r
e
(
%)
P
r
e
c
i
si
o
n
(
%)
7
F
e
a
t
u
r
e
s
LR
7
6
.
4
2
8
0
.
9
5
6
9
.
7
6
3
0
.
2
3
1
9
.
0
5
8
0
.
3
1
7
9
.
6
9
K
N
N
7
3
.
5
8
7
6
.
1
9
6
9
.
7
7
3
0
.
2
3
2
3
.
8
1
7
4
.
4
2
7
8
.
6
9
DT
8
4
.
1
3
8
4
.
1
3
7
6
.
4
4
2
3
.
5
6
1
5
.
8
7
8
4
.
1
3
8
4
.
1
3
A
d
a
B
o
o
st
8
5
.
8
5
9
5
.
2
4
7
2
.
0
9
2
7
.
9
1
4
.
7
6
8
8
.
0
0
8
2
.
7
6
RF
8
7
.
7
4
9
5
.
2
4
7
9
.
0
7
2
0
.
9
3
4
.
7
6
9
1
.
6
7
8
8
.
4
6
S
V
M
8
0
.
1
9
8
7
.
3
0
7
2
.
0
9
2
7
.
9
1
1
2
.
6
9
8
5
.
9
0
8
5
.
2
9
9
F
e
a
t
u
r
e
s
LR
8
0
.
1
9
8
2
.
5
4
7
6
.
7
4
2
3
.
2
6
1
7
.
4
6
8
1
.
8
2
8
1
.
8
2
K
N
N
7
3
.
5
8
7
6
.
1
9
6
9
.
7
7
3
0
.
2
3
2
3
.
8
1
7
4
.
4
2
7
8
.
6
9
DT
7
2
.
6
4
7
6
.
1
9
6
7
.
4
4
3
2
.
5
6
2
3
.
8
1
7
1
.
7
0
7
7
.
4
2
A
d
a
B
o
o
st
8
5
.
8
5
8
8
.
8
9
8
1
.
3
9
1
8
.
6
1
1
1
.
1
1
8
8
.
8
9
8
8
.
8
9
RF
8
3
.
0
2
8
8
.
5
7
7
7
.
9
1
2
2
.
0
9
1
4
.
2
8
8
5
.
7
1
8
5
.
7
1
S
V
M
8
0
.
1
9
8
7
.
3
0
6
9
.
7
7
3
0
.
2
3
1
2
.
6
7
8
0
.
8
7
8
0
.
8
8
T
ab
le
7
.
P
er
f
o
r
m
a
n
ce
m
etr
ic
s
f
o
r
d
if
f
er
e
n
t a
l
g
o
r
ith
m
s
u
s
i
n
g
7
an
d
9
f
ea
tu
r
es
w
it
h
L
A
S
SO
s
elec
tio
n
tec
h
n
iq
u
e
F
e
a
t
u
r
e
n
u
m
b
e
r
A
l
g
o
r
i
t
h
m
A
C
C
(
%)
S
EN
(
%)
S
P
E
(
%)
F
P
R
(
%)
F
N
R
(
%)
F1
-
sco
r
e
(
%)
P
r
e
c
i
si
o
n
(
%)
7
F
e
a
t
u
r
e
s
LR
8
1
.
1
3
8
4
.
1
3
7
6
.
7
4
2
3
.
2
6
1
5
.
8
7
8
4
.
1
3
8
4
.
1
3
K
N
N
7
3
.
5
8
7
6
.
1
9
6
9
.
7
7
3
0
.
2
3
2
3
.
8
1
7
7
.
4
2
7
8
.
6
9
DT
7
5
.
4
7
79
.
3
7
6
9
.
7
7
3
0
.
3
4
2
0
.
6
3
7
9
.
3
7
7
9
.
3
7
A
d
a
B
o
o
st
8
8
.
6
8
9
2
.
0
6
8
3
.
7
2
1
6
.
2
8
7
.
9
3
9
0
.
6
3
8
9
.
2
3
RF
8
3
.
0
2
8
5
.
7
1
7
9
.
0
7
2
0
.
9
3
1
4
.
2
9
8
5
.
7
1
8
5
.
7
1
S
V
M
8
0
.
1
9
8
7
.
3
0
6
9
.
7
7
3
0
.
2
3
1
2
.
6
9
8
3
.
9
7
8
0
.
8
9
9
F
e
a
t
u
r
e
s
LR
7
9
.
2
5
8
2
.
5
4
7
4
.
4
2
2
5
.
5
8
1
7
.
4
6
8
2
.
5
4
8
2
.
5
4
K
N
N
7
3
.
5
8
7
6
.
1
9
6
9
.
7
7
3
0
.
2
3
2
3
.
8
1
7
7
.
4
2
7
8
.
6
9
DT
7
5
.
4
7
7
7
.
7
7
7
2
.
0
9
2
7
.
9
1
2
2
.
2
2
7
9
.
0
3
8
0
.
3
3
A
d
a
B
o
o
st
8
5
.
8
5
8
8
.
8
9
8
1
.
3
9
1
8
.
6
0
1
1
.
1
1
8
7
.
5
0
8
7
.
5
0
RF
8
4
.
9
1
8
7
.
3
0
8
1
.
4
0
1
8
.
6
0
1
2
.
7
0
8
7
.
3
0
8
7
.
3
0
S
V
M
8
0
.
1
9
8
7
.
3
0
6
9
.
7
7
3
0
.
2
3
1
2
.
7
0
8
3
.
9
7
8
0
.
8
8
T
h
e
r
ec
eiv
er
o
p
er
atin
g
ch
ar
a
cter
is
tic
(
R
O
C
)
cu
r
v
e
il
lu
s
tr
a
tes
th
e
m
o
s
t
s
i
g
n
i
f
ican
t
p
o
r
tio
n
o
f
th
e
m
o
d
el
i
n
ter
m
s
o
f
i
ts
p
er
f
o
r
m
an
ce
.
W
h
en
t
h
e
R
OC
c
u
r
v
e
is
ev
al
u
ated
,
it
is
d
eter
m
in
ed
th
at
th
e
L
ASSO
f
ea
t
u
r
e
s
elec
tio
n
p
er
f
o
r
m
s
b
etter
th
an
t
h
e
ANOV
A
f
ea
tu
r
e
s
elec
tio
n
.
T
h
e
o
v
er
all
d
etai
l
s
ar
e
p
r
esen
ted
in
Fig
u
r
e
s
7
an
d
8
.
Fig
u
r
e
7
.
R
OC
c
u
r
v
e
f
o
r
m
o
d
el
p
er
f
o
r
m
a
n
ce
v
is
u
aliza
tio
n
u
s
in
g
ANOV
A
f
ea
t
u
r
e
s
et
Evaluation Warning : The document was created with Spire.PDF for Python.