I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
14
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
5
,
p
p
.
416
~
427
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
14
.i
1
.
p
p
4
1
6
-
4
2
7
416
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Weig
hted
nea
rest
neighbo
rs a
nd ra
dius
o
v
ersa
mpling
for
imba
la
nced data
cla
ss
ificatio
n
G
ede
Ang
g
a
P
ra
dip
t
a
1
, P
utu
Desia
na
Wula
nin
g
Ay
u
1
,
M
a
de
L
ia
nd
a
na
2
,
Da
nd
y
P
ra
ma
na
H
o
s
t
ia
di
1
1
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
mat
i
o
n
S
y
s
t
e
ms
,
I
n
st
i
t
u
t
Te
k
n
o
l
o
g
i
d
a
n
B
i
sn
i
s S
TI
K
O
M
B
a
l
i
,
D
e
n
p
a
sar,
I
n
d
o
n
e
s
i
a
2
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
mat
i
c
s
a
n
d
C
o
m
p
u
t
e
r
,
I
n
st
i
t
u
t
T
e
k
n
o
l
o
g
i
d
a
n
B
i
sn
i
s S
TI
K
O
M
B
a
l
i
,
D
e
n
p
a
s
a
r
,
I
n
d
o
n
e
s
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Au
g
19
,
2
0
2
3
R
ev
is
ed
Mar
7
,
2
0
2
4
Acc
ep
ted
Au
g
30
,
2
0
2
4
Th
e
c
h
a
ll
e
n
g
e
s
a
ss
o
c
iate
d
with
h
ig
h
-
d
ime
n
si
o
n
a
l
a
n
d
imb
a
lan
c
e
d
d
a
tas
e
ts
we
re
o
b
se
rv
e
d
t
o
o
ften
lea
d
t
o
a
d
e
g
ra
d
a
ti
o
n
i
n
th
e
p
e
rfo
rm
a
n
c
e
o
f
c
las
sic
a
l
m
a
c
h
in
e
lea
rn
in
g
a
lg
o
r
it
h
m
s.
I
n
th
e
c
a
se
o
f
h
i
g
h
d
ime
n
sio
n
a
l
d
a
t
a
,
n
o
t
a
l
l
fe
a
tu
re
s
c
o
n
tri
b
u
te
sig
n
ifi
c
a
n
tl
y
a
n
d
a
re
c
o
n
sid
e
re
d
re
lev
a
n
t
t
o
th
e
p
e
rfo
rm
a
n
c
e
o
f
t
h
e
m
o
d
e
l.
T
h
e
re
fo
re
,
t
h
is
stu
d
y
in
tro
d
u
c
e
d
a
n
o
v
e
l
m
e
th
o
d
c
a
ll
e
d
fe
a
tu
re
we
ig
h
ted
v
a
rian
c
e
a
n
a
ly
sis
-
n
e
a
re
st
n
e
ig
h
b
o
rs
(
WF
VA
NN
)
wh
ich
wa
s
d
e
v
e
lo
p
e
d
o
n
th
e
f
o
u
n
d
a
ti
o
n
o
f
k
-
n
e
a
re
st
n
e
i
g
h
b
o
rs
(
KN
N).
Th
e
p
ro
c
e
ss
in
v
o
lv
e
d
m
o
d
i
fy
i
n
g
t
h
e
c
a
lcu
latio
n
o
f
t
h
e
Eu
c
li
d
e
a
n
d
istan
c
e
b
y
fu
ll
y
c
o
n
sid
e
ri
n
g
t
h
e
re
lev
a
n
c
e
a
n
d
c
o
n
tri
b
u
t
io
n
lev
e
ls
o
f
fe
a
tu
re
s
b
a
se
d
o
n
t
h
e
ir
F
-
v
a
lu
e
.
W
F
VA
NN
a
t
th
e
a
lg
o
r
it
h
m
ic
lev
e
l
p
r
o
c
e
ss
in
g
a
n
d
ra
d
iu
s
-
sy
n
th
e
t
ic
m
in
o
rit
y
o
v
e
rsa
m
p
li
n
g
tec
h
n
iq
u
e
(R
-
S
M
OTE)
a
t
th
e
d
a
ta
lev
e
l
p
ro
c
e
ss
in
g
u
se
d
a
s
t
h
e
o
v
e
rsa
m
p
li
n
g
m
e
th
o
d
late
r
b
e
c
a
m
e
th
e
p
ro
p
o
se
d
m
o
d
e
l
to
so
l
v
e
th
e
a
fo
re
m
e
n
ti
o
n
e
d
iss
u
e
s.
M
o
re
o
v
e
r,
e
x
ten
si
v
e
e
x
p
e
rime
n
ts
we
re
c
o
n
d
u
c
ted
o
n
tw
o
d
ist
in
c
t
t
y
p
e
s
o
f
d
a
ta
i
n
c
lu
d
i
n
g
t
h
e
h
i
g
h
-
d
ime
n
sio
n
a
l
a
n
d
i
m
b
a
l
a
n
c
e
d
b
y
c
o
m
p
a
rin
g
W
F
VA
NN
with
th
e
sta
te
-
of
-
a
rt
KN
N
-
b
a
se
d
a
n
d
sy
n
t
h
e
ti
c
m
in
o
rit
y
o
v
e
rsa
m
p
li
n
g
tec
h
n
i
q
u
e
(
S
M
OTE
)
-
b
a
se
d
m
e
th
o
d
s.
T
h
e
re
su
lt
s
sh
o
we
d
th
a
t
t
h
e
p
ro
p
o
se
d
m
e
th
o
d
h
a
d
th
e
h
ig
h
e
st
a
c
c
u
ra
c
y
,
p
re
c
isio
n
,
re
c
a
ll
,
a
n
d
F
1
-
m
e
a
su
re
v
a
lu
e
s
a
c
ro
ss
th
e
m
a
jo
rit
y
o
f
tes
t
d
a
tas
e
ts
a
n
d
o
u
t
p
e
rfo
rm
e
d
th
e
o
t
h
e
r
m
e
th
o
d
s.
K
ey
w
o
r
d
s
:
Hig
h
d
im
en
s
io
n
al
d
ata
I
m
b
alan
ce
d
d
ata
K
-
n
ea
r
est n
eig
h
b
o
r
s
Sy
n
th
etic
m
in
o
r
ity
o
v
er
s
am
p
lin
g
tech
n
iq
u
e
W
eig
h
ted
f
ea
tu
r
e
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
:
Ged
e
An
g
g
a
Pra
d
ip
ta
Dep
ar
tm
en
t o
f
Ma
g
is
ter
I
n
f
o
r
m
atio
n
Sy
s
tem
s
,
I
n
s
titu
t T
ek
n
o
lo
g
i d
a
n
B
is
n
is
STI
KOM
B
ali,
I
n
d
o
n
esi
a
E
m
ail:
an
g
g
a
_
p
r
ad
ip
ta@
s
tik
o
m
-
b
ali.
ac
.
id
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
d
a
t
a
s
e
t
w
i
t
h
s
i
g
n
i
f
i
c
an
t
d
i
s
p
a
r
i
ty
d
u
r
in
g
th
e
p
r
o
c
e
s
s
o
f
d
i
s
t
r
ib
u
t
i
n
g
d
a
t
a
in
t
o
d
i
f
f
er
e
n
t
c
l
a
s
s
e
s
i
s
t
e
c
h
n
ic
a
l
l
y
s
t
a
te
d
to
b
e
i
m
b
a
l
a
n
c
e
d
.
T
h
i
s
s
i
g
n
if
i
c
an
t
d
is
p
a
r
i
t
y
,
o
f
t
e
n
e
x
tr
e
m
e
,
ex
i
s
t
s
a
m
o
n
g
c
l
a
s
s
e
s
o
r
l
a
b
e
l
s
o
f
d
a
t
a
c
a
s
e
s
,
t
h
e
r
eb
y
m
a
k
in
g
i
m
b
a
l
a
n
ce
d
d
a
ta
s
e
t
s
c
o
m
m
o
n
i
n
r
e
a
l
-
wo
r
ld
ap
p
l
i
ca
t
i
o
n
s
a
n
d
co
n
c
r
e
t
e
f
i
e
l
d
s
s
u
ch
a
s
m
e
d
ic
a
l
d
i
ag
n
o
s
i
s
[
1
]
-
[
3
]
,
f
au
l
t
d
i
ag
n
o
s
i
s
[
4
]
,
[
5
]
,
an
o
m
a
l
y
d
e
t
e
c
t
io
n
[
6
]
‒
[
8
]
,
in
t
r
u
s
i
o
n
d
e
t
e
c
t
io
n
[
9
]
‒
[
1
1
]
,
a
n
d
s
e
v
e
r
al
o
t
h
er
s
.
M
ea
n
wh
i
l
e
,
th
e
m
i
n
o
r
i
t
y
c
la
s
s
h
o
ld
s
a
h
i
g
h
e
r
l
e
v
e
l
o
f
im
p
o
r
t
an
c
e
an
d
i
n
t
e
r
e
s
t
d
u
r
in
g
th
e
r
ec
o
g
n
i
ti
o
n
p
r
o
c
e
s
s
b
y
m
a
c
h
in
e
l
ea
r
n
in
g
m
o
d
e
l
s
.
F
o
r
in
s
t
a
n
ce
,
i
d
en
t
i
f
y
i
n
g
a
n
d
r
e
c
o
g
n
i
z
in
g
p
a
t
t
er
n
s
o
f
r
a
r
e
d
i
s
e
a
s
e
s
in
m
e
d
ic
a
l
d
i
ag
n
o
s
i
s
i
s
cr
u
c
i
a
l
b
u
t
th
e
a
c
t
u
a
l
d
a
ta
co
u
n
t
f
o
r
n
o
r
m
a
l
c
o
n
d
i
t
i
o
n
s
f
a
r
o
u
t
w
e
ig
h
s
t
h
o
s
e
l
i
n
k
ed
to
t
h
e
d
i
s
e
a
s
e
s
.
A
c
r
i
t
i
c
a
l
ch
a
l
l
en
g
e
en
co
u
n
t
e
r
ed
w
h
e
n
d
e
a
l
i
n
g
w
i
t
h
c
l
a
s
s
i
m
b
a
la
n
c
e
d
u
r
in
g
l
e
ar
n
i
n
g
i
s
t
h
e
f
a
i
l
u
r
e
o
f
m
o
s
t
s
t
a
n
d
ar
d
m
a
ch
i
n
e
l
ea
r
n
in
g
a
lg
o
r
i
t
h
m
s
to
p
r
o
je
c
t
a
c
c
u
r
a
t
e
b
o
u
n
d
ar
y
l
i
n
e
s
f
o
r
ea
c
h
c
l
a
s
s
w
i
th
i
n
t
h
e
d
a
t
a
s
e
t
in
s
o
m
e
c
a
s
e
s
.
T
h
i
s
i
s
b
e
ca
u
s
e
m
ac
h
in
e
l
e
a
r
n
i
n
g
p
r
ed
o
m
in
an
t
l
y
l
e
ar
n
s
p
a
t
t
e
r
n
s
f
r
o
m
t
h
e
m
a
j
o
r
i
t
y
c
la
s
s
,
i
n
t
r
o
d
u
c
i
n
g
b
i
a
s
to
w
a
r
d
t
h
e
m
i
n
o
r
i
ty
c
l
a
s
s
a
n
d
le
a
d
in
g
t
o
c
l
a
s
s
o
v
e
r
l
ap
p
in
g
.
T
h
i
s
o
v
e
r
l
ap
p
in
g
,
a
l
s
o
r
e
f
er
r
ed
t
o
a
s
c
l
a
s
s
co
m
p
l
e
x
i
ty
o
r
s
e
p
ar
a
b
i
l
i
t
y
,
s
i
g
n
i
f
ie
s
t
h
e
d
e
g
r
e
e
o
f
s
e
p
ar
a
t
i
o
n
b
e
t
w
ee
n
c
l
a
s
s
e
s
i
n
t
h
e
d
a
ta
.
C
o
n
s
eq
u
en
t
l
y
,
s
t
a
n
d
ar
d
m
a
c
h
in
e
le
a
r
n
in
g
a
lg
o
r
i
t
h
m
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Weig
h
ted
n
ea
r
est n
eig
h
b
o
r
s
a
n
d
r
a
d
iu
s
o
ve
r
s
a
mp
lin
g
f
o
r
imb
a
la
n
ce
d
…
(
Ged
e
A
n
g
g
a
P
r
a
d
ip
ta
)
417
s
t
r
u
g
g
l
e
t
o
d
e
f
in
e
a
n
d
d
et
e
r
m
i
n
e
d
i
s
cr
i
m
in
a
t
i
v
e
r
u
l
e
s
f
o
r
c
l
a
s
s
s
e
p
a
r
a
t
i
o
n
.
T
h
i
s
o
v
e
r
l
ap
p
i
n
g
f
ea
t
u
r
e
s
p
a
c
e
l
e
a
d
s
t
o
t
h
e
l
o
s
s
o
f
i
n
tr
i
n
s
i
c
p
r
o
p
e
r
t
i
e
s
w
i
th
i
n
th
e
d
a
t
a,
r
en
d
e
r
in
g
i
t
r
ed
u
n
d
an
t
o
r
ir
r
e
le
v
a
n
t
i
n
t
h
e
p
r
o
c
e
s
s
o
f
r
e
c
o
g
n
i
z
in
g
g
o
o
d
d
e
c
i
s
i
o
n
b
o
u
n
d
a
r
i
e
s
b
e
t
w
e
en
c
la
s
s
e
s
.
Sev
er
al
s
o
lu
tio
n
s
ar
e
d
is
co
v
er
ed
to
h
av
e
b
ee
n
p
r
o
p
o
s
ed
to
ad
d
r
ess
th
is
i
s
s
u
e
o
v
er
tim
e
an
d
th
o
s
e
ap
p
lied
in
p
r
ev
i
o
u
s
s
tu
d
ies
ca
n
b
e
ca
te
g
o
r
ized
in
to
t
h
r
e
e
g
r
o
u
p
s
in
clu
d
i
n
g
d
ata
s
am
p
lin
g
,
al
g
o
r
ith
m
ic
m
o
d
if
icatio
n
,
an
d
co
s
t
-
s
en
s
itiv
e
lear
n
in
g
[
1
2
]
‒
[
1
4
]
.
T
h
is
is
b
ec
au
s
e
s
o
m
e
s
tu
d
ies
f
o
u
n
d
m
eth
o
d
s
b
y
p
r
ep
r
o
ce
s
s
in
g
th
e
d
ata,
p
ar
tic
u
lar
ly
b
y
r
esam
p
lin
g
th
e
m
in
o
r
ity
d
ata
to
alter
th
e
class
d
is
tr
ib
u
tio
n
an
d
tack
le
im
b
alan
ce
d
d
atasets
.
On
e
o
f
th
e
m
o
s
t
wid
ely
u
s
ed
m
eth
o
d
s
is
o
v
er
s
am
p
lin
g
a
n
d
th
is
i
n
v
o
lv
es
cr
ea
t
in
g
a
s
u
p
er
s
et
o
f
th
e
o
r
ig
in
al
d
ataset
b
y
r
ep
licatin
g
s
o
m
e
i
n
s
tan
ce
s
o
r
d
ev
elo
p
in
g
n
ew
i
n
s
tan
ce
s
f
r
o
m
e
x
is
tin
g
o
n
es.
Stu
d
ies o
n
d
ata
o
v
er
s
am
p
lin
g
wid
ely
u
s
ed
t
h
e
s
y
n
th
etic
m
in
o
r
ity
o
v
er
s
am
p
lin
g
tech
n
i
q
u
e
(
SMOT
E
)
[
1
5
]
.
T
h
e
m
ain
id
ea
g
e
n
er
ally
ass
o
ciate
d
with
th
is
m
eth
o
d
is
th
e
cr
e
atio
n
o
f
n
ew
ex
am
p
les
f
o
r
th
e
m
in
o
r
ity
class
b
y
in
ter
p
o
latin
g
s
ev
e
r
al
in
s
tan
ce
s
f
r
o
m
th
e
class
.
Ho
wev
er
,
SMOT
E
h
as
s
o
m
e
d
r
awb
ac
k
s
d
esp
ite
its
ab
ilit
y
to
im
p
r
o
v
e
th
e
d
is
tr
ib
u
tio
n
o
f
e
x
am
p
les
in
ea
ch
class
.
On
e
o
f
th
e
d
r
awb
ac
k
s
is
r
elate
d
to
b
lin
d
o
v
er
s
am
p
lin
g
wh
ich
in
v
o
lv
es
f
o
cu
s
in
g
o
n
ly
o
n
th
e
in
f
o
r
m
atio
n
f
r
o
m
its
n
ea
r
est
d
ata
o
r
n
ea
r
est
p
o
s
itiv
e
ex
a
m
p
le
with
o
u
t
co
n
s
id
er
in
g
t
h
e
s
p
atial
in
f
o
r
m
atio
n
o
f
t
h
e
n
eig
h
b
o
r
s
[
1
6
]
.
T
h
is
u
s
u
ally
r
esu
lts
in
s
ev
er
al
n
ewly
g
en
er
ated
d
ata
p
o
in
ts
f
allin
g
in
t
o
th
e
a
r
ea
s
o
f
t
h
e
n
eg
ativ
e
o
r
m
in
o
r
ity
class
,
lead
in
g
to
th
e
cr
ea
tio
n
o
f
n
o
is
y
d
ata
an
d
d
is
r
u
p
tio
n
in
th
e
in
te
r
-
class
ar
ea
s
with
in
th
e
d
ataset.
Fu
r
th
er
m
o
r
e,
t
h
e
o
v
er
la
p
p
in
g
f
ea
tu
r
e
s
p
ac
e
ca
u
s
es
th
e
f
ea
tu
r
es
to
lo
s
e
th
eir
in
tr
in
s
ic
p
r
o
p
er
ty
,
l
ea
d
in
g
to
r
ed
u
n
d
a
n
cy
o
r
ir
r
elev
an
ce
in
r
ec
o
g
n
izin
g
g
o
o
d
d
ec
is
io
n
b
o
u
n
d
a
r
ies
b
etwe
en
class
es.
On
e
ca
teg
o
r
y
o
f
s
o
lu
tio
n
s
d
ev
elo
p
ed
to
ad
d
r
ess
th
ese
ch
allen
g
es
f
o
cu
s
es
o
n
th
e
ty
p
e
o
f
in
t
er
p
o
latio
n
u
s
ed
an
d
th
e
d
ete
r
m
in
atio
n
o
f
th
e
r
eg
i
o
n
s
in
wh
ich
n
e
w
d
ata
ar
e
f
o
r
m
e
d
u
s
in
g
S
MO
T
E
m
eth
o
d
.
T
h
e
in
ter
p
o
latio
n
m
ec
h
an
is
m
ca
n
t
ak
e
v
ar
io
u
s
f
o
r
m
s
s
u
ch
as
th
e
r
an
g
e
-
r
estricte
d
wh
ich
in
v
o
lv
es
co
n
s
id
er
in
g
th
e
in
f
o
r
m
atio
n
o
f
b
o
t
h
th
e
n
ea
r
est
p
o
s
itiv
e
an
d
n
eg
ativ
e
n
eig
h
b
o
r
s
.
Mo
r
eo
v
er
,
s
o
m
e
s
tu
d
ies
u
s
ed
m
u
ltip
le
in
ter
p
o
latio
n
s
[
1
7
]
,
in
v
o
l
v
in
g
m
o
r
e
th
a
n
two
ex
a
m
p
les
o
r
f
o
l
lo
win
g
to
p
o
lo
g
ies
b
ased
o
n
g
e
o
m
etr
ic
s
h
ap
es
s
u
ch
as
ellip
s
es
[
1
8
]
,
Vo
r
o
n
o
i
d
iag
r
am
s
[
1
9
]
,
an
d
g
r
ap
h
s
[
2
0
]
.
S
ev
er
al
s
tu
d
ies
[
2
1
]
‒
[
2
3
]
also
ap
p
lied
clu
s
ter
in
g
-
b
ased
in
ter
p
o
latio
n
with
ea
ch
n
ew
ex
am
p
le
lim
ited
t
o
b
ei
n
g
f
o
r
m
ed
i
n
th
e
s
am
e
clu
s
ter
a
r
ea
f
r
o
m
th
e
s
am
p
le
p
o
in
t in
a
d
d
itio
n
to
th
e
co
m
b
i
n
atio
n
with
th
e
clu
s
ter
in
g
m
et
h
o
d
.
T
h
e
d
eter
m
in
atio
n
o
f
th
e
d
is
tan
ce
b
etwe
en
a
p
o
s
itiv
e
s
am
p
le
p
o
in
t
an
d
its
n
ea
r
est
n
eig
h
b
o
r
s
f
o
r
in
ter
p
o
latio
n
in
SMOT
E
m
eth
o
d
is
b
ased
o
n
th
e
ca
lcu
latio
n
o
f
E
u
clid
ea
n
d
is
tan
ce
.
Sev
er
al
SMOT
E
d
ev
elo
p
m
e
n
ts
also
co
n
ce
n
tr
at
e
o
n
ap
p
r
o
p
r
iate
d
ata
s
am
p
lin
g
to
r
ed
u
ce
th
e
o
cc
u
r
r
e
n
ce
o
f
o
v
er
lap
p
i
n
g
r
eg
io
n
s
an
d
p
r
e
v
en
t
th
e
g
en
er
atio
n
o
f
n
o
is
y
n
ew
d
ata.
T
o
id
en
tify
th
e
b
est
d
ata
s
am
p
les,
s
ev
er
al
m
eth
o
d
s
h
av
e
b
ee
n
u
s
ed
to
s
elec
t
ca
n
d
id
ate
s
am
p
les.
T
h
is
was
in
d
icate
d
in
p
r
ev
i
o
u
s
s
tu
d
ies
wh
er
e
s
am
p
les
we
r
e
ca
teg
o
r
ized
in
t
o
s
af
e
an
d
d
an
g
er
o
u
s
zo
n
es
[
2
4
]
s
elec
ted
b
o
r
d
er
r
eg
io
n
s
[
2
5
]
,
an
d
d
eter
m
in
e
d
d
if
f
icu
lty
weig
h
ts
f
o
r
ea
ch
in
s
tan
ce
[
2
6
]
,
[
2
7
]
.
T
h
e
two
ex
ce
p
tio
n
s
id
en
tifie
d
wer
e
th
e
g
e
n
er
ati
o
n
o
f
s
y
n
th
etic
e
x
am
p
les
af
te
r
a
lea
r
n
in
g
v
ec
t
o
r
q
u
an
tizatio
n
(
L
VQ)
o
p
tim
izat
io
n
p
r
o
ce
s
s
[
2
8
]
a
n
d
th
e
s
elec
tio
n
o
f
i
n
itial
p
o
in
ts
f
r
o
m
t
h
e
s
u
p
p
o
r
t
v
ec
to
r
s
o
b
tain
ed
b
y
an
SVM
[
2
9
]
.
I
n
g
en
er
al,
t
h
e
d
is
tan
ce
ca
lcu
latio
n
s
u
s
ed
in
b
o
th
tr
a
d
itio
n
al
SMOT
E
an
d
its
d
ev
elo
p
m
e
n
ts
wer
e
o
b
s
er
v
ed
t
o
r
ely
o
n
k
-
n
ea
r
est
n
eig
h
b
o
r
s
(
KNN)
alg
o
r
ith
m
u
s
ed
in
s
ele
ctin
g
d
ata
s
am
p
les.
Ho
wev
er
,
th
e
u
s
e
o
f
KNN
also
h
as
s
ev
er
al
s
h
o
r
tco
m
in
g
s
s
u
c
h
as
th
e
s
en
s
i
tiv
ity
to
th
e
n
eig
h
b
o
r
h
o
o
d
s
ize
k
[
3
0
]
an
d
th
e
d
is
tan
ce
f
u
n
ctio
n
a
p
p
l
ied
to
s
elec
t
KN
N.
T
h
e
id
en
tific
atio
n
o
f
th
e
m
o
s
t
s
u
itab
le
d
is
tan
ce
f
o
r
m
u
la
f
o
r
all
tr
ain
in
g
s
am
p
les
was
f
o
u
n
d
to
b
e
a
ch
allen
g
in
g
ex
e
r
cise.
KNN
also
h
as
h
i
g
h
co
m
p
le
x
ity
d
u
e
to
th
e
n
ee
d
to
s
ea
r
ch
f
o
r
n
ea
r
est n
eig
h
b
o
r
s
a
n
d
is
co
n
s
id
er
ed
less
ef
f
ec
tiv
e
f
o
r
im
b
ala
n
ce
d
class
d
atasets
.
Mo
s
t
SMOT
E
m
eth
o
d
s
wer
e
c
o
n
tin
u
o
u
s
ly
b
ein
g
d
ev
el
o
p
ed
b
ased
o
n
tr
a
d
itio
n
al
KNN
m
eth
o
d
,
th
e
r
eb
y
lead
in
g
to
th
e
p
er
s
is
ten
ce
o
f
th
e
lim
itatio
n
s
.
T
h
is
led
p
r
ev
io
u
s
s
tu
d
ies
to
p
r
o
p
o
s
e
im
p
r
o
v
em
en
ts
to
KNN
m
eth
o
d
,
p
ar
ticu
lar
ly
f
o
c
u
s
in
g
o
n
th
e
is
s
u
e
o
f
s
en
s
itiv
ity
to
t
h
e
k
v
alu
e
.
T
h
e
lo
ca
l
m
ea
n
f
a
cto
r
was
ap
p
lied
t
o
m
itig
ate
th
e
ef
f
ec
ts
o
f
k
s
en
s
itiv
ity
wh
ile
s
ev
er
al
o
th
er
m
eth
o
d
s
in
clu
d
in
g
k
-
h
ar
m
o
n
i
c
n
ea
r
est
n
eig
h
b
o
r
s
(
KHNN
)
[
3
1
]
,
lo
ca
l
m
ea
n
-
b
ased
k
-
n
ea
r
est
n
ei
g
h
b
o
r
s
(
L
MK
NN)
[
3
2
]
,
l
o
ca
l
m
ea
n
-
b
a
s
ed
p
s
eu
d
o
n
ea
r
est
n
eig
h
b
o
r
s
(
L
PMNN)
[
3
3
]
,
a
n
d
m
u
lti
-
lo
ca
l
m
ea
n
s
-
b
ased
n
ea
r
est
n
eig
h
b
o
r
s
(
ML
NN)
[
3
4
]
co
n
ce
n
tr
ated
o
n
r
ed
u
cin
g
o
u
tlier
s
in
th
e
v
icin
i
ty
o
f
th
e
s
am
p
le
p
o
i
n
ts
.
Sev
er
al
im
p
r
o
v
em
en
ts
wer
e
also
m
ad
e
b
y
ass
ig
n
in
g
weig
h
ts
to
ea
ch
d
ata
p
o
in
t
wit
h
in
th
e
n
ei
g
h
b
o
r
h
o
o
d
s
u
c
h
as
p
s
eu
d
o
n
ea
r
est
n
eig
h
b
o
r
s
(
PN
N)
[
3
5
]
,
weig
h
ted
r
ep
r
esen
tatio
n
-
b
ased
k
-
n
ea
r
est
n
eig
h
b
o
r
s
(
W
R
KNN)
,
an
d
we
ig
h
ted
lo
ca
l
m
ea
n
r
e
p
r
esen
tatio
n
-
b
ased
k
-
n
ea
r
est
n
eig
h
b
o
r
s
(
W
L
MRKNN)
[
3
0
]
.
T
h
e
d
ev
elo
p
m
en
t
o
f
th
ese
wei
g
h
tin
g
m
eth
o
d
s
was
b
ased
o
n
t
h
e
o
b
s
er
v
atio
n
th
at
ea
ch
n
ea
r
est
n
eig
h
b
o
r
o
f
ten
c
o
n
tr
ib
u
tes
d
if
f
er
e
n
tly
to
th
e
class
if
icatio
n
o
u
tco
m
e
in
r
ea
l
-
wo
r
ld
p
r
o
b
lem
d
ata.
KNN
m
eth
o
d
was
o
b
s
er
v
ed
t
o
h
av
e
b
ee
n
d
ev
elo
p
e
d
b
ased
o
n
th
e
d
is
tan
c
e
a
n
d
p
o
s
itio
n
o
f
n
ea
r
est
n
eig
h
b
o
r
s
wh
ile
co
n
s
id
er
in
g
th
e
weig
h
ts
o
f
ea
ch
d
ata
p
o
in
t.
T
h
e
p
u
r
p
o
s
e
was
to
ac
co
u
n
t
f
o
r
p
o
te
n
tial
o
u
tlier
s
in
th
e
s
u
r
r
o
u
n
d
in
g
ar
ea
with
o
u
t
co
n
s
i
d
er
in
g
th
e
co
n
tr
ib
u
tio
n
o
f
ea
ch
f
ea
tu
r
e
in
d
eter
m
in
i
n
g
n
ea
r
est
n
eig
h
b
o
r
s
d
is
tan
ce
f
o
r
ea
c
h
d
ata
p
o
in
t.
T
h
e
b
asic
an
d
m
o
s
t
co
m
m
o
n
d
is
tan
ce
m
ea
s
u
r
em
en
t
in
KNN
was
f
o
u
n
d
to
b
e
t
y
p
ically
p
er
f
o
r
m
ed
u
s
in
g
E
u
clid
ea
n
d
is
tan
ce
ca
lcu
latio
n
.
Ho
wev
er
,
th
er
e
was
th
e
p
o
s
s
ib
ilit
y
o
f
ea
ch
f
ea
tu
r
e
h
a
v
in
g
v
ar
y
in
g
c
o
n
tr
ib
u
tio
n
s
to
th
e
class
if
icatio
n
o
u
tco
m
e.
T
h
e
p
r
o
x
im
ity
o
f
ea
ch
d
ata
p
o
in
t
was
lik
ely
to
b
e
in
f
lu
en
ce
d
b
y
s
ev
e
r
al
f
ea
tu
r
e
s
co
n
tr
ib
u
tin
g
to
an
d
co
r
r
ela
tin
g
with
th
e
class
lab
el.
T
h
er
ef
o
r
e,
th
is
s
tu
d
y
p
r
o
p
o
s
ed
a
n
ew
m
eth
o
d
b
as
ed
o
n
KNN
ca
lled
th
e
f
ea
tu
r
e
weig
h
ted
v
ar
ian
ce
an
al
y
s
is
-
n
ea
r
est
n
eig
h
b
o
r
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
5
:
416
-
4
2
7
418
(
W
FVAN
N)
.
T
h
e
p
r
o
ce
s
s
in
v
o
lv
ed
m
o
d
if
y
in
g
th
e
d
is
tan
ce
ca
lcu
latio
n
o
f
d
ata
p
o
in
ts
b
y
ad
d
in
g
a
weig
h
tin
g
f
ea
tu
r
e
to
th
e
e
x
is
tin
g
d
ata
f
ea
tu
r
es.
T
h
e
weig
h
t
v
alu
es
d
e
p
e
n
d
ed
o
n
th
e
co
r
r
elatio
n
a
n
d
c
o
n
tr
ib
u
tio
n
lev
el
o
f
ea
ch
f
ea
tu
r
e.
Mo
r
eo
v
er
,
t
h
e
F
-
v
alu
e
o
b
tain
e
d
th
r
o
u
g
h
an
a
ly
s
is
o
f
v
ar
ian
ce
(
ANOV
A)
m
eth
o
d
was
u
s
ed
to
co
m
p
u
te
th
e
f
ea
t
u
r
e
c
o
n
tr
ib
u
tio
n
v
al
u
es
an
d
f
u
r
th
er
c
o
m
b
i
n
ed
with
r
ad
iu
s
-
s
y
n
th
etic
m
in
o
r
ity
o
v
er
s
am
p
lin
g
tech
n
iq
u
e
(
R
-
SMOT
E
)
m
o
d
i
f
icatio
n
,
an
o
v
er
s
am
p
lin
g
m
eth
o
d
,
to
s
o
lv
e
th
e
im
b
alan
ce
d
d
ata
p
r
o
b
l
em
.
W
FVA
NN
ap
p
lied
at
th
e
alg
o
r
ith
m
ic
lev
el
an
d
R
-
SMOT
E
a
t
d
ata
lev
el
p
r
o
ce
s
s
in
g
wer
e
later
d
esig
n
ed
as
th
e
p
r
o
p
o
s
ed
m
o
d
el
to
s
o
lv
e
th
e
c
h
allen
g
es
ass
o
ciate
d
with
b
u
il
d
in
g
a
r
o
b
u
s
t
m
ac
h
in
e
lear
n
in
g
m
o
d
e
in
two
d
ata
co
n
d
itio
n
s
,
in
clu
d
in
g
h
ig
h
d
im
en
s
io
n
al
an
d
im
b
alan
ce
d
.
2.
M
E
T
H
O
D
2
.
1
.
Weig
hte
d f
ea
t
ure
v
a
ria
nce
a
na
ly
s
is
-
nea
re
s
t
neig
hb
o
rs
T
h
e
a
lg
o
r
i
t
h
m
m
o
d
i
f
ie
d
to
ca
l
c
u
l
a
t
e
th
e
E
u
c
l
id
e
a
n
d
i
s
t
an
c
e
i
n
K
N
N
w
a
s
o
b
s
er
v
ed
t
o
r
e
l
y
o
n
th
e
F
-
v
a
l
u
e
o
b
ta
i
n
ed
f
r
o
m
th
e
f
e
a
t
u
r
e
s
e
l
ec
t
i
o
n
p
r
o
c
e
s
s
u
s
i
n
g
A
N
O
V
A
.
T
h
e
b
i
g
g
e
s
t
c
h
al
l
e
n
g
e
in
m
a
ch
i
n
e
l
e
a
r
n
in
g
w
a
s
th
e
s
e
l
e
c
t
i
o
n
o
f
t
h
e
b
e
s
t
f
e
a
tu
r
e
s
to
t
r
a
in
th
e
m
o
d
e
l
.
T
h
er
e
f
o
r
e,
th
i
s
s
t
u
d
y
a
i
m
e
d
t
o
s
e
l
e
c
t
t
h
e
f
e
a
t
u
r
e
s
c
o
n
s
i
d
e
r
e
d
h
ig
h
ly
d
ep
e
n
d
e
n
t
o
n
th
e
r
e
s
p
o
n
s
e
v
a
r
ia
b
l
e
.
T
h
i
s
w
a
s
b
e
c
au
s
e
t
h
e
v
a
r
i
a
n
ce
o
f
a
f
e
a
t
u
r
e
u
s
u
a
l
l
y
d
e
t
e
r
m
i
n
e
s
t
h
e
l
e
v
e
l
o
f
i
t
s
im
p
a
c
t o
n
th
e
r
e
s
p
o
n
s
e
v
a
r
i
a
b
l
e
b
a
s
e
d
o
n
th
e
cr
i
t
e
r
i
o
n
t
h
a
t
a
l
o
w
v
a
r
i
an
c
e
i
n
d
i
c
a
te
s
a
l
a
ck
o
f
i
m
p
a
c
t
an
d
v
i
c
e
v
er
s
a
.
A
N
O
V
A
wa
s
d
e
f
i
n
ed
a
s
a
s
t
a
t
i
s
t
i
c
a
l
m
e
th
o
d
n
o
r
m
a
l
l
y
u
s
ed
t
o
c
h
e
ck
t
h
e
m
e
an
s
o
f
t
w
o
o
r
m
o
r
e
g
r
o
u
p
s
t
h
a
t
a
r
e
s
i
g
n
if
i
c
an
t
l
y
d
if
f
er
e
n
t
f
r
o
m
e
a
c
h
o
th
e
r
.
S
i
m
i
l
a
r
l
y
,
i
n
K
N
N
m
e
t
h
o
d
,
th
e
ca
l
c
u
l
a
t
io
n
o
f
p
r
o
x
i
m
i
ty
b
e
t
w
e
e
n
s
am
p
l
e
p
o
in
t
s
a
n
d
th
e
i
r
n
e
a
r
e
s
t
n
e
ig
h
b
o
r
s
s
h
o
u
ld
b
e
h
e
av
i
l
y
i
n
f
l
u
e
n
c
ed
b
y
f
e
a
tu
r
e
s
s
e
l
e
c
te
d
b
a
s
ed
o
n
h
i
g
h
r
e
l
ev
a
n
ce
to
th
e
d
ep
en
d
en
t
v
ar
i
a
b
l
e.
A
h
i
g
h
e
r
F
-
v
a
l
u
e
f
o
r
a
f
e
a
t
u
r
e
u
s
u
a
l
ly
s
ig
n
if
i
e
s
a
g
r
ea
t
e
r
we
i
g
h
t
a
s
s
i
g
n
e
d
i
n
d
e
t
e
r
m
i
n
i
n
g
th
e
E
u
c
l
i
d
e
an
d
i
s
t
a
n
ce
.
I
t
w
a
s
a
l
s
o
n
o
t
e
d
t
h
a
t
n
o
t
a
l
l
f
ea
t
u
r
e
s
p
o
s
s
e
s
s
s
i
g
n
i
f
ic
a
n
ce
i
n
s
h
ap
i
n
g
t
h
e
d
ec
i
s
i
o
n
b
o
u
n
d
a
r
y
b
e
t
w
e
en
c
l
as
s
e
s
i
n
t
h
e
c
a
s
e
o
f
h
i
g
h
d
im
e
n
s
io
n
a
l
d
a
t
a.
S
o
m
e
i
r
r
e
le
v
an
t
f
e
a
t
u
r
e
s
u
s
ed
in
d
a
ta
p
a
t
t
e
r
n
d
e
t
e
r
m
i
n
a
t
io
n
c
o
u
ld
r
e
d
u
c
e
th
e
p
e
r
f
o
r
m
a
n
c
e
o
f
t
h
e
m
a
ch
i
n
e
le
a
r
n
in
g
m
o
d
e
l
d
ev
e
lo
p
ed
.
T
h
e
r
e
f
o
r
e,
th
e
p
r
o
p
o
s
e
d
W
F
V
AN
N
m
e
t
h
o
d
a
s
s
i
g
n
e
d
w
e
i
g
h
t
s
to
e
a
ch
f
e
a
t
u
r
e
b
a
s
e
d
o
n
th
e
i
r
r
e
s
p
e
c
t
i
v
e
v
ar
i
an
c
e
an
a
l
y
s
e
s
a
s
in
d
i
ca
t
e
d
i
n
th
e
f
l
o
w
p
r
o
c
e
s
s
p
r
e
s
e
n
t
ed
i
n
F
i
g
u
r
e
1
.
Fig
u
r
e
1
.
T
h
e
f
lo
w
o
f
W
FVANN
p
r
o
ce
s
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Weig
h
ted
n
ea
r
est n
eig
h
b
o
r
s
a
n
d
r
a
d
iu
s
o
ve
r
s
a
mp
lin
g
f
o
r
imb
a
la
n
ce
d
…
(
Ged
e
A
n
g
g
a
P
r
a
d
ip
ta
)
419
T
h
e
weig
h
ts
wer
e
o
b
tain
ed
f
r
o
m
th
e
F
-
v
alu
e
d
eter
m
in
ed
u
s
in
g
ANOV
A
in
th
e
f
ea
tu
r
e
s
elec
tio
n
m
eth
o
d
.
T
h
e
f
ea
tu
r
es with
a
h
i
g
h
F
-
v
alu
e
wer
e
ass
ig
n
ed
p
r
o
p
o
r
tio
n
ally
h
ig
h
er
weig
h
ts
wh
ile
th
o
s
e
with
a
lo
w
o
r
ev
e
n
ze
r
o
F
-
v
alu
e
h
ad
n
o
i
m
p
ac
t
o
n
th
e
d
is
tan
ce
ca
lcu
lat
i
o
n
in
KNN.
Mo
r
eo
v
er
,
W
FVANN
alg
o
r
ith
m
was
f
o
r
m
alize
d
i
n
Alg
o
r
ith
m
1
an
d
th
e
weig
h
tin
g
f
o
r
ea
c
h
f
ea
t
u
r
e
in
th
e
d
ata
was
d
eter
m
in
ed
u
s
in
g
ANOV
A
b
y
ca
lcu
latin
g
th
e
F
-
v
alu
e
f
o
r
ea
c
h
f
ea
tu
r
e.
ANOV
A
r
an
k
ed
t
h
e
f
ea
tu
r
es
b
y
ca
lc
u
latin
g
th
e
v
ar
i
an
ce
r
atio
b
etwe
e
n
an
d
with
in
g
r
o
u
p
s
.
Fu
r
th
er
m
o
r
e,
th
e
F
-
v
alu
e
was
co
m
p
u
ted
b
y
f
in
d
in
g
t
h
e
r
atio
o
f
m
ea
n
s
q
u
ar
e
b
etwe
en
(
MSB
)
to
m
ea
n
s
q
u
a
r
e
with
in
(
MSW
)
.
I
n
Step
1
,
th
e
v
a
r
ian
ce
v
al
u
e
f
o
r
ea
c
h
g
r
o
u
p
o
r
lab
el
was
c
alcu
lated
.
Fo
r
ea
ch
f
ea
tu
r
e
i
with
in
a
lab
el,
its
av
er
ag
e
v
alu
e
(
̅
)
was
d
eter
m
in
ed
an
d
s
u
b
tr
ac
ted
f
r
o
m
th
e
to
tal
av
er
ag
e
v
alu
e
o
f
th
e
f
ea
tu
r
e
(
̅
)
.
T
h
e
s
u
b
tr
ac
tio
n
r
esu
lt
was
later
m
u
ltip
lied
b
y
th
e
n
u
m
b
er
o
f
lab
els
in
t
h
e
d
ata,
d
en
o
ted
as
k
.
T
h
e
r
esu
lt
was
s
u
b
s
eq
u
en
tly
d
i
v
id
ed
b
y
th
e
d
eg
r
ee
o
f
f
r
ee
d
o
m
f
o
r
MSB
,
r
ep
r
esen
ted
as
−
1
.
Step
2
in
v
o
lv
e
d
ca
lcu
latin
g
th
e
MSW
v
alu
e.
E
ac
h
d
ata
p
o
in
t
with
in
lab
el
i
(
)
was
s
u
b
tr
ac
ted
f
r
o
m
th
e
av
er
ag
e
v
alu
e
o
f
t
h
e
lab
el
(
̅
)
an
d
th
e
r
esu
lt
was
d
iv
id
ed
b
y
th
e
d
e
g
r
ee
o
f
f
r
ee
d
o
m
f
o
r
MSW
,
r
ep
r
esen
ted
as
−
.
Step
3
was
u
s
ed
to
d
iv
id
e
MSB
b
y
MSW
to
o
b
tain
th
e
F
-
v
alu
e
f
o
r
f
ea
tu
r
e
i.
I
n
s
tep
s
4
-
7
,
th
e
r
e
s
u
ltin
g
F
-
v
alu
e
was
n
o
r
m
alize
d
u
s
in
g
th
e
Min
-
Ma
x
n
o
r
m
aliza
tio
n
f
o
r
m
u
la.
T
h
e
m
eth
o
d
was
u
s
ed
to
s
ca
le
th
e
f
ea
tu
r
e
v
alu
es
to
a
r
an
g
e
b
etwe
en
0
an
d
1
an
d
th
i
s
was
ac
h
iev
ed
b
y
s
u
b
tr
ac
tin
g
th
e
m
in
im
u
m
v
alu
e
o
f
th
e
f
ea
t
u
r
e
f
r
o
m
ea
ch
v
alu
e
an
d
d
iv
id
i
n
g
b
y
th
e
r
a
n
g
e
o
f
th
e
f
ea
tu
r
e
weig
h
ted
.
Su
b
s
e
q
u
en
tly
,
th
e
d
is
t
an
ce
f
o
r
th
e
n
ew
d
ata
p
o
in
t
was
ca
lcu
lated
ag
ain
s
t
all
tr
ain
in
g
d
ata
p
o
in
ts
X
u
s
in
g
th
e
E
u
clid
e
an
d
is
tan
ce
f
o
r
m
u
la
b
y
ad
d
in
g
th
e
ef
f
ec
t
o
f
weig
h
t
o
n
ea
c
h
f
ea
tu
r
e.
Fin
ally
,
t
h
e
p
r
o
ce
s
s
co
n
tin
u
e
d
i
n
s
tep
s
8
-
1
1
b
y
s
o
r
tin
g
with
n
ea
r
est
n
eig
h
b
o
r
s
v
al
u
e
f
r
o
m
th
e
ca
lcu
latio
n
r
esu
lts
.
Alg
o
r
ith
m
1
: Weig
h
ted
f
ea
tu
r
e
b
y
v
a
r
ian
ce
an
aly
s
is
alg
o
r
ith
m
u
s
in
g
F
-
v
alu
e
I
n
p
u
t:
X
:
tr
ain
in
g
d
ata,
Y
:
lab
el
o
f
x
,
m
:
Nu
m
b
er
Nea
r
est
Neig
h
b
o
r
s
(
1
,
2
,
…
,
)
,
:
in
d
ex
j
in
g
r
o
u
p
class
I
,
:
in
s
tan
ce
s
in
d
ex
I
,
̅
:
m
ea
n
wit
h
in
-
g
r
o
u
p
class
,
̅
:
m
ea
n
all
d
at
a,
:
n
u
m
b
er
o
f
class
,
f
:
f
ea
tu
r
e
in
r
o
w
d
ata,
: n
u
m
b
er
o
f
class
,
n
: n
u
m
b
er
o
f
f
ea
tu
r
es,
K
=
co
u
n
t o
f
all
d
ata,
an
d
k
=
co
u
n
t o
f
g
r
o
u
p
s
.
Ou
tp
u
t: L
ab
els C
lass
o
f
x
Sam
p
les
1.
C
alcu
latin
g
m
ea
n
s
q
u
ar
e
b
etwe
en
(
MSB
)
:
(
∑
×
(
−
̅
)
2
)
/
(
−
1
)
2.
C
alcu
latin
g
m
ea
n
s
q
u
ar
e
with
i
n
(
MSW
)
:
(
∑
∑
(
−
̅
)
=
1
2
=
1
)
/
(
−
)
3.
C
alcu
latin
g
F
-
v
alu
e
=
(
∑
×
(
−
̅
)
2
)
/
(
−
1
)
×
(
∑
∑
(
−
̅
)
=
1
2
=
1
)
/
(
−
)
4.
=
Min
-
Ma
x
n
o
r
m
aliza
tio
n
5.
W
h
ile
(
≤
)
C
alcu
late
d
is
tan
ce
:
(
,
)
=
√
∑
(
−
)
2
=
1
×
6.
=
+
7.
E
n
d
8.
(
,
)
9.
Hea
d
(
m
)
10.
C
las
s
if
y
x
:
(
)
=
∑
∈
(
,
)
11.
E
n
d
2
.
2
.
Ra
diu
s
s
y
nthet
ic
m
ino
rit
y
o
v
er
s
a
m
pli
ng
t
ec
hn
iqu
e
a
lg
o
rit
hm
T
h
e
is
s
u
es
o
f
o
v
er
lap
p
in
g
,
n
o
i
s
e,
an
d
s
m
all
d
is
ju
n
ct
ca
s
e
s
w
er
e
d
is
co
v
er
ed
to
b
e
em
e
r
g
in
g
f
r
o
m
th
e
r
an
d
o
m
s
elec
tio
n
o
f
s
am
p
les
with
in
th
e
m
in
o
r
ity
class
d
ata.
T
h
e
n
o
is
e
with
in
t
h
e
m
in
o
r
ity
class
d
ata
co
u
ld
lead
to
th
e
cr
ea
tio
n
o
f
n
ew
n
o
is
y
d
a
ta,
r
esu
ltin
g
in
co
n
f
licts
b
etwe
en
th
e
r
eg
io
n
s
o
f
ea
c
h
class
.
T
h
is
ch
allen
g
e
c
o
u
ld
b
e
tack
led
u
s
in
g
R
-
SMOT
E
m
eth
o
d
b
y
f
ilter
in
g
th
e
s
am
p
l
e
d
ata
to
e
n
s
u
r
e
a
m
o
r
e
p
r
ec
i
s
e
s
am
p
le
s
elec
tio
n
p
r
o
ce
s
s
.
T
h
er
e
f
o
r
e,
th
e
p
r
o
p
o
s
ed
m
o
d
if
ie
d
SMOT
E
m
o
d
el
was
in
itiated
b
y
ca
teg
o
r
izin
g
m
in
o
r
ity
class
d
at
a
p
o
in
ts
in
to
th
r
ee
g
r
o
u
p
s
in
clu
d
in
g
s
af
e,
n
o
is
e,
o
r
s
m
all
d
is
j
u
n
ct
.
T
h
e
d
ata
s
elec
tio
n
o
r
f
ilter
in
g
p
r
o
ce
s
s
was
p
er
f
o
r
m
ed
u
s
in
g
KNN
m
eth
o
d
b
ased
o
n
th
e
p
o
s
itio
n
an
d
p
r
o
x
im
ity
o
f
th
e
d
ata
to
o
th
e
r
class
es.
E
ac
h
m
in
o
r
ity
d
ata
p
o
in
t
was
s
elec
ted
u
s
in
g
KNN
p
ar
am
eter
s
et
to
5
,
an
d
t
h
o
s
e
co
r
r
ec
tly
class
if
ied
wer
e
lab
eled
as
s
af
e
wh
ile
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
5
:
416
-
4
2
7
420
th
o
s
e
class
if
ied
to
b
e
a
m
ajo
r
ity
class
d
ata
p
o
in
t
wer
e
tag
g
ed
n
o
is
e
o
r
s
m
all
d
is
ju
n
ct
.
T
h
e
ca
teg
o
r
izatio
n
p
r
o
ce
s
s
was
f
o
llo
wed
b
y
th
e
g
en
er
ati
o
n
o
f
n
ew
s
y
n
th
etic
d
ata
wh
i
ch
was
lim
ited
to
th
e
s
af
e
ca
teg
o
r
y
.
Similar
to
SMOT
E
m
eth
o
d
,
s
y
n
th
etic
d
a
ta
wer
e
g
en
e
r
ated
b
y
id
e
n
tify
i
n
g
th
e
n
ea
r
est
m
in
o
r
ity
d
ata
p
o
in
ts
an
d
d
r
awin
g
in
ter
p
o
latio
n
lin
es
b
etwe
en
th
em
.
T
h
e
d
ete
r
m
in
atio
n
o
f
th
e
n
u
m
b
er
o
f
n
ea
r
est
d
ata
p
o
in
ts
in
SMOT
E
m
eth
o
d
was a
ls
o
b
ased
o
n
KNN
with
p
ar
am
eter
k
r
ep
r
esen
tin
g
th
e
n
u
m
b
er
o
f
n
ea
r
est d
ata
p
o
in
t
s
.
Ho
wev
er
,
th
e
u
s
e
o
f
th
is
k
p
ar
am
eter
as
d
escr
ib
e
d
ea
r
lier
p
o
s
ed
th
e
r
is
k
o
f
g
en
e
r
atin
g
s
y
n
th
etic
d
ata
th
at
c
o
u
ld
ca
u
s
e
o
v
er
lap
p
in
g
b
etwe
en
th
e
m
in
o
r
ity
an
d
m
aj
o
r
ity
class
es.
T
h
er
ef
o
r
e,
th
is
s
t
u
d
y
p
r
o
p
o
s
ed
th
e
u
s
e
o
f
a
r
ad
i
u
s
p
ar
am
eter
in
s
tead
wh
ich
was
d
eter
m
in
ed
b
y
f
in
d
in
g
th
e
d
is
tan
ce
to
th
e
n
ea
r
est
m
ajo
r
ity
d
ata
p
o
in
t
f
r
o
m
th
e
s
am
p
le
to
b
e
u
s
ed
as
r
ad
iu
s
v
alu
e.
All
th
e
n
ew
d
at
a
p
o
in
ts
wer
e
g
en
er
ated
with
in
th
is
r
ad
iu
s
b
o
u
n
d
ar
y
u
s
in
g
th
e
cir
cle
eq
u
atio
n
p
r
esen
ted
in
(
1
)
a
n
d
ex
e
m
p
lifi
ed
in
a
two
-
d
im
en
s
i
o
n
al
v
ec
t
o
r
.
‖
~
−
~
‖
≤
2
‖
~
−
~
‖
≤
2
(
1
)
∑
(
−
)
2
≤
2
=
1
(
2
)
2
=
∑
(
−
)
2
=
1
(
3
)
W
h
er
e
r
ep
r
esen
ts
th
e
ce
n
ter
p
o
in
t
o
f
t
h
e
cir
cle
(
m
in
o
r
ity
s
am
p
le
p
o
in
t)
,
with
(
1
,
2
,
3
,
…
)
an
d
t
(
1
,
2
,
3
,
…
)
b
ein
g
th
e
n
ea
r
est
m
aj
o
r
ity
p
o
in
ts
to
th
e
ce
n
ter
o
f
th
e
cir
cle,
i
s
a
n
ew
d
ata
p
o
in
t
b
elo
w
r
ad
i
u
s
with
(
1
,
2
,
3
,
…
)
wh
er
e
i
=1
…n
,
th
en
2
is
th
e
d
is
tan
ce
b
etwe
en
p
an
d
as in
(
3
)
.
T
h
e
p
r
o
p
o
s
ed
m
o
d
el
is
p
r
esen
ted
in
Fig
u
r
e
2
.
Fig
u
r
e
2
.
R
-
SMOT
E
alg
o
r
ith
m
s
ch
em
e
T
h
e
d
is
tan
ce
o
f
ea
ch
m
in
o
r
ity
s
am
p
le
f
r
o
m
th
e
m
ajo
r
ity
class
was
al
s
o
ca
lcu
lated
u
s
in
g
th
e
E
u
clid
ea
n
d
is
tan
ce
m
eth
o
d
.
T
h
e
n
ea
r
est
m
ajo
r
ity
d
ata
p
o
in
t
was
th
e
o
n
e
with
th
e
m
in
im
u
m
v
al
u
e
t
o
all
d
is
tan
ce
s
f
r
o
m
m
in
o
r
ity
d
ata
p
o
in
ts
,
as sh
o
wn
in
(
4
)
.
=
∑
∑
√
(
−
)
2
=
1
=
1
(
4
)
W
h
e
r
e
,
r
e
p
r
e
s
e
n
ts
t
h
e
s
h
o
r
t
e
s
t
d
i
s
t
a
n
c
e
b
e
t
we
e
n
m
i
n
o
r
i
t
y
d
at
a
t
o
j
a
n
d
m
a
j
o
r
i
t
y
d
a
t
a
t
o
i
.
T
h
e
i
d
e
n
t
i
f
i
c
at
i
o
n
o
f
t
h
e
m
a
j
o
r
i
t
y
d
a
t
a
p
o
i
n
t
w
a
s
f
o
l
lo
w
e
d
b
y
t
h
e
s
y
n
t
h
e
s
i
s
o
f
n
e
w
d
a
t
a
t
h
r
o
u
g
h
i
n
t
e
r
p
o
l
a
t
i
o
n
b
e
t
w
ee
n
t
h
e
s
e
t
w
o
p
o
i
n
t
s
.
T
h
e
s
y
n
t
h
e
t
i
c
d
at
a
w
e
r
e
c
r
e
a
t
ed
a
l
o
n
g
t
w
o
d
i
r
e
c
t
i
o
n
s
o
f
l
i
n
es
,
i
n
c
l
u
d
i
n
g
a
n
d
−
b
a
s
e
d
o
n
t
h
e
(
5
)
a
n
d
(
6
)
:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Weig
h
ted
n
ea
r
est n
eig
h
b
o
r
s
a
n
d
r
a
d
iu
s
o
ve
r
s
a
mp
lin
g
f
o
r
imb
a
la
n
ce
d
…
(
Ged
e
A
n
g
g
a
P
r
a
d
ip
ta
)
421
=
+
(
(
0
,
1
)
×
(
−
)
)
(
5
)
=
+
(
(
0
,
1
)
×
(
−
)
)
(
6
)
T
h
e
ar
ea
to
p
r
o
d
u
ce
th
ese
n
ew
d
ata
was
lim
ited
to
r
ed
u
ce
th
e
o
cc
u
r
r
e
n
ce
o
f
o
v
er
lap
p
i
n
g
as
r
ec
o
r
d
ed
in
SMOT
E
m
eth
o
d
.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
is
s
ec
tio
n
was
u
s
ed
to
im
p
lem
en
t
an
d
test
W
FVA
NN
m
eth
o
d
u
s
in
g
s
ev
er
al
h
ig
h
-
d
i
m
en
s
io
n
al
d
atasets
.
A
co
m
p
ar
ativ
e
an
aly
s
is
was
al
s
o
co
n
d
u
cted
b
etwe
en
W
FVANN
an
d
two
p
r
ev
io
u
s
m
eth
o
d
s
in
clu
d
in
g
th
e
o
r
ig
i
n
al
KNN
an
d
L
MK
NN.
Su
b
s
eq
u
en
tly
,
a
s
ec
o
n
d
r
o
u
n
d
o
f
ex
p
er
im
en
ts
was a
p
p
lied
to
ass
ess
th
e
ef
f
ec
t
o
f
em
p
lo
y
in
g
W
FVANN
as
a
class
if
ier
o
n
o
v
er
s
am
p
led
d
ata
with
s
o
m
e
SMOT
E
d
ev
elo
p
m
en
t
m
eth
o
d
,
in
clu
d
in
g
B
o
r
d
er
lin
e
-
SMOT
E
,
ad
ap
tiv
e
s
y
n
th
etic
s
am
p
lin
g
(
ADAY
SN
)
,
s
af
e
lev
el
SMOT
E
,
an
d
SMOT
E
-
I
PF
.
3
.
1
.
E
x
perim
ent
a
l
f
ra
m
ewo
rk
a
nd
da
t
a
s
et
cha
ra
ct
er
is
t
ics
T
h
e
ex
p
e
r
im
en
t
test
was
d
iv
id
ed
in
to
two
p
ar
ts
in
clu
d
i
n
g
th
e
h
ig
h
-
d
im
en
s
io
n
al
an
d
im
b
alan
ce
d
d
atasets
.
Hig
h
-
d
im
en
s
io
n
al
d
atasets
wer
e
u
s
ed
to
ass
ess
t
h
e
ef
f
ec
tiv
e
n
ess
an
d
ap
p
licatio
n
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
,
W
FVANN,
in
weig
h
t
in
g
ea
ch
f
ea
tu
r
e
wh
ile
im
b
ala
n
ce
d
d
atasets
wer
e
em
p
lo
y
ed
to
ex
am
in
e
its
ef
f
ec
t
as
a
class
if
ier
wh
en
in
teg
r
ate
d
in
to
o
v
e
r
s
am
p
led
a
n
d
n
o
n
-
o
v
er
s
am
p
led
d
ata.
T
a
b
le
1
p
r
o
v
i
d
es
an
o
v
er
v
iew
o
f
r
elev
an
t
m
eta
d
ata
f
o
r
th
e
h
ig
h
-
d
im
en
s
io
n
al
d
atasets
,
in
clu
d
i
n
g
attr
ib
u
tes,
s
am
p
le
co
u
n
ts
,
c
lass
es,
an
d
d
is
ea
s
e
ty
p
es.
T
h
ese
m
icr
o
ar
r
ay
d
ata
s
ets
wer
e
s
o
u
r
ce
d
f
r
o
m
R
p
ac
k
ag
es
d
esig
n
e
d
to
ev
al
u
ate
m
ac
h
in
e
lear
n
i
n
g
alg
o
r
ith
m
s
an
d
m
o
d
els.
T
ab
le
1
.
Me
tad
ata
f
o
r
h
ig
h
-
d
im
en
s
io
n
al
d
atasets
No
D
a
t
a
s
e
t
#
S
a
m
p
l
e
s
#
A
t
t
r
C
l
a
s
ses
D
i
sea
s
e
1
A
l
o
n
62
2
,
0
0
0
2
C
o
l
o
n
c
a
n
c
e
r
2
B
o
r
o
v
e
c
k
i
31
2
2
,
2
8
3
2
H
u
n
t
i
n
g
t
o
n
's
D
i
se
a
se
3
C
h
i
a
r
e
t
t
i
1
1
1
1
2
,
6
2
5
2
Le
u
k
e
mi
a
4
C
h
i
n
1
1
8
2
2
,
2
1
5
2
B
r
e
a
s
t
c
a
n
c
e
r
5
C
h
o
w
d
a
r
y
1
0
4
2
2
,
2
8
3
2
B
r
e
a
s
t
c
a
n
c
e
r
6
C
h
r
i
st
e
n
s
e
n
2
1
7
1
,
4
1
3
3
-
7
G
o
l
u
b
72
7
,
1
2
9
3
Le
u
k
i
mi
a
8
G
o
r
d
o
n
1
8
1
1
2
,
5
3
3
2
Lu
n
g
C
a
n
c
e
r
9
G
r
a
v
i
e
r
1
6
8
2
,
9
0
5
2
B
r
e
a
s
t
c
a
n
c
e
r
10
K
h
a
n
63
2
,
3
0
8
4
S
R
B
C
T
A
to
tal
o
f
1
3
d
if
f
er
e
n
t
im
b
ala
n
ce
d
d
ataset
s
wer
e
o
b
tain
ed
f
r
o
m
d
if
f
er
en
t
a
p
p
licatio
n
ar
ea
s
o
n
b
in
ar
y
an
d
m
u
lticlas
s
class
if
icatio
n
p
r
o
b
lem
s
.
T
h
e
d
ataset
s
u
s
ed
h
ad
a
d
if
f
e
r
en
t
n
u
m
b
er
o
f
f
ea
t
u
r
es
an
d
a
d
if
f
e
r
en
t
im
b
alan
ce
r
atio
an
d
wer
e
o
b
tai
n
ed
f
r
o
m
UC
I
m
ac
h
in
e
lear
n
in
g
r
ep
o
s
ito
r
y
[
3
6
]
an
d
k
n
o
wled
g
e
ex
tr
ac
tio
n
b
ased
o
n
ev
o
lu
tio
n
ar
y
lear
n
i
n
g
(
KE
E
L
)
r
ep
o
s
ito
r
y
[
3
7
]
.
T
a
b
le
2
s
h
o
ws
th
eir
ch
ar
ac
ter
is
tics
with
a
f
o
cu
s
o
n
th
e
im
b
alan
ce
d
r
atio
v
alu
e
(
I
R
)
,
wh
ich
r
ep
r
esen
ts
th
e
v
alu
e
o
f
th
e
r
atio
b
etwe
en
n
eg
ativ
e
an
d
p
o
s
itiv
e
class
e
s
,
th
e
n
u
m
b
er
o
f
f
ea
t
u
r
es
(
#
Attr
)
in
ea
ch
d
ataset,
t
h
e
n
u
m
b
e
r
o
f
d
a
ta
o
r
in
s
tan
ce
s
(
#
s
am
p
les)
in
e
ac
h
d
ataset,
as
well
as
th
e
n
u
m
b
e
r
o
f
c
o
m
p
a
r
is
o
n
s
o
f
p
o
s
itiv
e
an
d
n
e
g
ativ
e
in
s
tan
ce
s
in
p
er
ce
n
t size.
T
ab
le
2
.
Me
tad
ata
f
o
r
im
b
alan
ce
d
d
atasets
No
N
a
me
IR
#
A
t
t
r
#
S
a
m
p
l
e
s
P
o
si
t
i
v
e
i
n
st
a
n
c
e
s
(
%)
N
e
g
a
t
i
v
e
i
n
st
a
n
c
e
s
(
%)
1
0
3
s
u
b
c
l
5
-
6
0
0
-
5
-
70
-
BI
5
2
6
0
0
1
6
.
6
7
8
3
.
3
0
2
0
4
c
l
o
v
e
r
5
z
-
600
-
5
-
70
-
BI
5
2
6
0
0
1
6
.
6
7
8
3
.
3
0
3
e
c
o
l
i
-
0
-
1
-
3
-
7
_
v
s_
2
-
6
3
9
.
1
4
7
2
8
1
2
.
4
9
9
7
.
5
1
4
g
l
a
ss1
1
.
8
2
9
2
1
4
3
5
.
4
6
6
4
.
5
4
5
n
e
w
t
h
y
r
o
i
d
4
.
8
4
5
2
1
5
1
7
.
1
2
8
2
.
8
8
6
p
a
w
0
2
a
-
6
0
0
-
5
-
70
-
BI
5
2
6
0
0
1
6
.
6
7
8
3
.
3
0
7
w
i
n
e
1
.
5
13
1
7
8
4
0
.
0
0
6
0
.
0
0
8
y
e
a
st
-
1
-
4
-
5
-
8
_
v
s
_
7
2
2
.
1
0
8
6
9
3
4
.
3
3
0
9
5
.
6
7
9
U
mb
i
l
i
c
a
l
C
o
r
d
1
8
.
8
7
5
1
5
1
5
.
3
0
0
9
4
.
7
0
10
B
r
e
a
s
t
2
.
3
6
9
2
8
6
2
9
.
1
2
7
0
.
3
8
11
H
a
b
e
r
m
a
n
2
.
7
8
3
3
0
6
2
6
.
3
9
7
3
.
6
1
12
P
i
ma
1
.
8
7
8
7
6
8
3
4
.
8
6
6
5
.
1
4
13
B
u
p
a
1
.
3
8
6
3
4
5
4
2
.
1
9
5
7
.
8
1
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
5
:
416
-
4
2
7
422
T
h
e
v
alid
atio
n
m
ec
h
an
is
m
u
s
ed
was
k
-
f
o
ld
c
r
o
s
s
-
v
alid
atio
n
with
a
to
tal
o
f
1
0
f
o
ld
s
.
Mo
r
eo
v
er
,
t
h
e
p
er
f
o
r
m
an
ce
o
f
th
e
class
if
icatio
n
m
o
d
el
was
test
ed
u
s
in
g
f
o
u
r
m
etr
ics
in
clu
d
in
g
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F
-
Me
asu
r
e.
I
n
m
ac
h
in
e
lear
n
in
g
class
if
icatio
n
ta
s
k
s
,
co
n
f
u
s
io
n
m
atr
ix
in
clu
d
in
g
tr
u
e
p
o
s
itiv
e
(
T
P),
tr
u
e
n
eg
ativ
e
(
T
N)
,
f
alse
p
o
s
itiv
e
(
FP
)
,
an
d
f
alse
n
eg
ativ
e
(
F
N)
wer
e
th
e
m
ain
p
ar
a
m
eter
s
f
r
o
m
wh
ich
o
th
er
p
er
f
o
r
m
an
ce
m
etr
ics
s
u
ch
as
p
r
ec
is
io
n
,
r
ec
al
l,
an
d
F1
s
co
r
es
wer
e
co
m
p
u
ted
.
T
h
e
ac
cu
r
ac
y
was
u
s
ed
to
m
ea
s
u
r
e
th
e
am
o
u
n
t
o
f
d
ata
co
r
r
ec
tly
cl
ass
if
ied
ac
co
r
d
in
g
to
th
e
g
r
o
u
n
d
-
tr
u
th
la
b
el
d
iv
id
e
d
b
y
th
e
to
t
al
d
ata
u
s
ed
f
o
r
th
e
test
.
Pre
cisi
o
n
was
th
e
r
ate
o
f
co
r
r
ec
t
p
r
e
d
ictio
n
s
am
o
n
g
all
s
am
p
les
p
r
ed
icted
to
b
elo
n
g
to
th
e
m
in
o
r
ity
class
an
d
in
d
icate
d
th
e
n
u
m
b
er
o
f
p
o
s
itiv
e
p
r
ed
ictio
n
s
co
n
s
id
er
e
d
to
b
e
co
r
r
ec
t.
Me
an
wh
ile,
r
ec
all
f
o
cu
s
ed
o
n
th
e
p
r
o
p
o
r
tio
n
o
f
m
in
o
r
ity
-
class
s
am
p
les lab
eled
as p
o
s
itiv
e.
3
.
2
.
P
er
f
o
r
m
a
nce
a
na
ly
s
is
o
f
hig
h dim
ens
io
na
l da
t
a
T
h
e
p
e
r
f
o
r
m
an
ce
o
f
W
FVANN
m
o
d
el
was
o
b
s
er
v
e
d
to
h
av
e
ex
ce
lled
ac
r
o
s
s
m
o
s
t
o
f
th
e
u
s
ed
h
ig
h
-
d
im
en
s
io
n
al
d
atasets
as
p
r
esen
ted
in
T
ab
le
3
.
I
n
ter
m
s
o
f
ac
cu
r
ac
y
,
th
e
m
o
d
el
ex
h
ib
ited
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
in
d
atasets
s
u
ch
as
B
o
r
o
v
ec
k
i,
C
h
in
,
C
h
o
wd
ar
y
,
C
h
r
is
ten
s
en
,
Go
lu
b
,
an
d
Kh
an
.
A
s
u
b
s
tan
tial
im
p
r
o
v
em
e
n
t
was
also
o
b
s
er
v
ed
in
th
e
B
o
r
o
v
ec
k
i
d
ataset
co
m
p
ar
e
d
to
th
e
o
th
er
th
r
ee
m
eth
o
d
s
as
in
d
icat
ed
b
y
th
e
d
if
f
er
e
n
ce
o
f
3
3
%
r
ec
o
r
d
ed
with
KNN
a
n
d
L
MK
NN
ac
h
iev
i
n
g
an
ac
c
u
r
ac
y
o
f
5
6
%
wh
ile
W
FVANN
h
ad
8
9
%.
T
h
e
r
ec
all
me
asu
r
em
en
t
s
h
o
wed
th
at
W
FVANN
h
ad
en
h
a
n
ce
d
v
alu
es
i
n
6
o
u
t
o
f
th
e
to
tal
1
0
d
atasets
with
a
r
em
a
r
k
ab
le
3
5
%
in
cr
ea
s
e
s
p
ec
if
ically
r
ec
o
r
d
ed
i
n
B
o
r
o
v
ec
k
i
d
ataset.
Fu
r
th
er
m
o
r
e,
th
e
ass
ess
m
en
t
o
f
t
h
e
p
r
ec
is
io
n
a
n
d
F1
-
m
ea
s
u
r
e
v
alu
es
in
d
icate
d
th
at
W
FVANN
m
o
d
el
s
h
o
wed
c
o
m
m
en
d
a
b
le
p
er
f
o
r
m
a
n
ce
wit
h
in
cr
ea
s
ed
v
alu
es
r
ec
o
r
d
e
d
ac
r
o
s
s
6
d
atasets
.
T
ab
le
3
.
T
h
e
h
i
g
h
est
p
r
ec
is
io
n
,
r
ec
all
,
an
d
F1
s
co
r
e
(
%)
p
r
o
d
u
ce
d
b
y
KNN,
L
MK
NN,
an
d
W
F
ANN
D
a
t
a
s
e
t
A
c
c
u
r
a
c
y
R
e
c
a
l
l
P
r
e
c
i
s
i
o
n
F
-
1
S
c
o
r
e
K
N
N
LM
K
W
F
V
A
K
N
N
LM
K
W
F
V
A
K
N
N
LM
K
W
F
V
A
K
N
N
LM
K
W
F
V
A
A
l
o
n
63
89
78
65
90
80
78
91
85
59
89
78
B
o
r
o
v
e
c
k
i
56
56
89
60
55
90
75
55
90
50
55
90
C
h
i
a
r
e
t
t
i
54
77
64
25
53
51
21
49
45
22
51
46
C
h
i
n
83
89
92
79
87
89
89
90
94
80
88
91
C
h
o
w
d
a
r
y
94
97
1
0
0
95
96
1
0
0
93
97
1
0
0
93
97
1
0
0
C
h
r
i
st
e
n
s
e
n
1
0
0
97
1
0
0
1
0
0
98
1
0
0
1
0
0
98
1
0
0
1
0
0
98
1
0
0
G
o
l
u
b
86
95
1
0
0
82
88
1
0
0
77
97
1
0
0
79
92
1
0
0
G
o
r
d
o
n
85
1
0
0
96
56
1
0
0
89
92
1
0
0
98
56
1
0
0
93
G
r
a
v
i
e
r
73
98
75
50
96
56
36
99
71
42
97
54
K
h
a
n
84
1
0
0
1
0
0
81
1
0
0
1
0
0
87
1
0
0
1
0
0
77
1
0
0
1
0
0
T
h
e
ef
f
ec
t
o
f
th
e
k
v
alu
e
is
p
r
esen
ted
in
Fig
u
r
e
3
with
th
e
p
r
o
p
o
s
ed
m
o
d
el
d
en
o
ted
b
y
s
q
u
ar
e
s
h
ap
e
an
d
its
s
tab
ilit
y
was
f
o
u
n
d
to
be
q
u
ite
s
tab
le
co
m
p
ar
ed
to
o
t
h
er
m
o
d
els.
I
n
ca
s
e
Alo
n
d
ataset
Fig
u
r
e
3
(
a)
,
T
h
e
ac
cu
r
ac
y
r
em
ain
s
s
o
m
ewh
at
co
n
s
is
ten
t
as
th
e
n
u
m
b
er
o
f
n
e
ig
h
b
o
r
s
(
k
)
in
c
r
ea
s
es.
T
h
er
e
is
s
o
m
e
f
lu
ctu
atio
n
b
etwe
en
d
if
f
er
e
n
t
v
alu
es
o
f
k
,
b
u
t
it
s
tab
ilizes
at
ar
o
u
n
d
k
=5
.
Fu
r
th
er
m
o
r
e
,
b
o
r
o
v
ec
k
i
Fig
u
r
e
3
(
b
)
,
T
h
e
ac
cu
r
ac
y
r
em
ain
s
s
tab
le
an
d
h
ig
h
ac
r
o
s
s
m
o
s
t v
alu
es o
f
k
.
T
h
e
h
ig
h
est
p
er
f
o
r
m
an
ce
is
o
b
s
er
v
ed
with
k
=2
an
d
co
n
tin
u
es
ac
r
o
s
s
d
if
f
er
e
n
t
k
v
alu
es.
I
n
th
e
c
h
iar
etti
d
ataset
Fig
u
r
e
3
(
c)
s
h
o
ws
a
m
o
r
e
f
lu
ctu
atin
g
b
e
h
av
io
r
with
KNN,
with
a
s
ig
n
if
ican
t
d
i
p
ar
o
u
n
d
k
=
3
b
u
t
s
tab
ilizin
g
at
h
ig
h
e
r
v
al
u
e
s
o
f
k
.
Oth
er
wis
e
ch
in
d
ataset
Fig
u
r
e
3
(
d
)
s
h
o
ws
ac
cu
r
ac
y
f
lu
ctu
ates
q
u
ite
s
ig
n
if
ican
tly
f
o
r
d
if
f
er
en
t
v
alu
es
o
f
k
.
I
t
s
h
o
ws
th
at
t
h
e
ch
o
ice
o
f
k
ca
n
d
r
asti
ca
lly
af
f
ec
t
p
er
f
o
r
m
an
ce
o
n
th
is
d
at
aset.
C
h
o
wd
ar
y
d
ataset
Fig
u
r
e
3
(
e)
,
th
e
ac
cu
r
ac
y
is
q
u
ite
s
tab
le
ac
r
o
s
s
d
if
f
er
en
t
v
alu
es
o
f
k
,
s
h
o
win
g
o
n
ly
a
s
lig
h
t
d
ip
at
s
o
m
ek
v
alu
es.
Ov
er
all,
p
r
o
p
o
s
ed
m
eth
o
d
p
er
f
o
r
m
s
co
n
s
is
ten
tl
y
well
o
n
th
is
d
ataset.
Su
b
s
eq
u
en
tly
in
C
h
r
is
ten
s
en
Fig
u
r
e
3
(
f
)
,
s
h
o
ws
h
ig
h
ac
cu
r
ac
y
u
n
til
ar
o
u
n
d
k
=6
,
wh
er
e
th
er
e
i
s
a
s
h
ar
p
d
ec
lin
e
in
p
er
f
o
r
m
a
n
ce
.
Af
ter
th
is
p
o
in
t,
th
e
ac
cu
r
ac
y
r
em
ain
s
lo
w.
Glo
lu
b
Fig
u
r
e
3
(
g
)
,
th
er
e
is
n
o
tab
le
f
lu
ctu
atio
n
i
n
ac
c
u
r
ac
y
f
o
r
d
if
f
er
en
t v
alu
es
o
f
k
,
in
d
icatin
g
t
h
at
th
is
d
ataset
is
s
en
s
itiv
e
to
t
h
e
ch
o
ice
o
f
k
.
T
h
e
p
er
f
o
r
m
an
ce
s
h
o
ws
an
u
p
war
d
tr
en
d
af
ter
k
=3
.
Fu
r
th
er
m
o
r
e,
Go
r
d
o
n
d
ataset
Fig
u
r
e
3
(
h
)
,
s
h
o
ws
p
er
f
o
r
m
a
n
ce
o
n
th
is
d
ataset
is
g
en
er
ally
s
ta
b
le,
th
o
u
g
h
th
er
e
is
a
s
lig
h
t
d
r
o
p
at
s
p
ec
if
ic
k
v
alu
es.
Ho
we
v
er
,
it
r
em
ain
s
h
ig
h
ac
r
o
s
s
th
e
b
o
ar
d
.
Gr
av
ie
r
d
ata
s
et
in
Fig
u
r
e
3
(
i)
s
h
o
ws
f
lu
ctu
atin
g
ac
cu
r
ac
y
,
with
a
n
o
tab
le
d
ip
at
k
=
4
,
b
u
t
th
e
p
er
f
o
r
m
an
ce
b
o
u
n
ce
s
b
ac
k
w
ith
h
ig
h
er
v
alu
es
o
f
k
.
T
h
en
th
e
last
o
n
e
is
k
h
a
n
d
ataset
in
Fig
u
r
e
3
(
j)
s
h
o
ws
ir
r
eg
u
lar
p
er
f
o
r
m
an
ce
,
with
lar
g
e
f
lu
ctu
atio
n
s
in
ac
c
u
r
ac
y
as
k
ch
an
g
es.
T
h
e
p
er
f
o
r
m
a
n
ce
is
in
co
n
s
is
ten
t
ac
r
o
s
s
d
if
f
er
en
t
v
al
u
es
o
f
k
.
Ho
we
v
e
r
,
th
e
o
p
tim
al
k
v
alu
e
was
d
if
f
er
en
t
f
r
o
m
t
h
o
s
e
u
s
ed
in
th
e
a
cc
u
r
ac
y
m
etr
ic
f
o
r
h
ig
h
-
d
im
e
n
s
io
n
al
d
at
asets
.
W
FVANN
s
h
o
wed
r
eliab
ilit
y
o
n
h
ig
h
-
d
im
e
n
s
io
n
al
d
atasets
with
9
o
u
t
o
f
1
0
d
atasets
d
is
co
v
er
ed
to
h
av
e
ac
h
iev
ed
a
cc
u
r
ac
y
,
p
r
ec
is
io
n
,
an
d
F1
-
m
e
asu
r
e
v
alu
es
th
at
s
u
r
p
ass
ed
(
o
r
eq
u
al
to
)
th
o
s
e
o
f
o
th
er
m
eth
o
d
s
,
ex
ce
p
t in
th
e
G
r
av
ier
d
ataset
wh
er
e
L
MK
NN
m
eth
o
d
h
ad
a
h
ig
h
er
ac
c
u
r
ac
y
v
alu
e
o
f
9
8
%.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Weig
h
ted
n
ea
r
est n
eig
h
b
o
r
s
a
n
d
r
a
d
iu
s
o
ve
r
s
a
mp
lin
g
f
o
r
imb
a
la
n
ce
d
…
(
Ged
e
A
n
g
g
a
P
r
a
d
ip
ta
)
423
(
a)
(
b
)
(
c)
(
d
)
(
e)
(f)
(
g
)
(
h
)
(
i)
(
j)
F
i
g
u
r
e
3
.
A
c
c
u
r
a
c
y
p
e
r
f
o
r
m
a
n
ce
u
s
i
n
g
d
i
f
f
e
r
e
n
t
k
f
o
r
h
i
g
h
-
d
i
m
e
n
s
i
o
n
a
l
d
a
ta
s
et
s
i
n
d
a
t
a
s
e
t
:
(
a
)
a
l
o
n
,
(
b
)
b
o
r
o
v
e
c
k
i
,
(
c
)
c
h
i
a
r
e
t
t
i
,
(
d
)
c
h
i
n
,
(
e
)
c
h
o
w
d
a
r
y
,
(
f
)
c
h
r
i
s
t
e
n
s
e
n
,
(
g
)
g
o
l
u
b
,
(
h
)
g
o
r
d
o
n
,
(
i
)
g
r
a
v
i
e
r
,
a
n
d
(
j
)
k
h
a
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
14
,
No
.
1
,
Feb
r
u
ar
y
2
0
2
5
:
416
-
4
2
7
424
3
.
3
.
P
er
f
o
r
m
a
nce
a
na
ly
s
is
o
f
t
he
im
ba
la
nced
da
t
a
T
h
e
p
u
r
p
o
s
e
o
f
t
h
is
ex
p
er
im
e
n
t
was
to
u
n
d
er
s
tan
d
th
e
p
er
f
o
r
m
an
ce
o
f
W
FVANN
m
o
d
el
d
u
r
i
n
g
th
e
p
r
o
ce
s
s
o
f
h
an
d
lin
g
im
b
alan
c
ed
d
atasets
.
T
h
is
was
co
n
s
id
er
ed
n
ec
ess
ar
y
d
u
e
to
th
e
ab
ilit
y
o
f
th
e
im
b
alan
ce
d
d
ataset
co
n
d
itio
n
s
to
ca
u
s
e
a
d
ec
lin
e
in
th
e
ac
cu
r
ac
y
o
f
co
n
v
en
tio
n
al
m
ac
h
i
n
e
lear
n
in
g
m
eth
o
d
s
,
p
ar
ticu
lar
ly
o
n
m
in
o
r
ity
d
ata.
Ov
er
s
am
p
lin
g
m
eth
o
d
was
d
is
co
v
er
ed
to
h
av
e
b
ec
o
m
e
a
tr
en
d
i
n
g
s
o
lu
tio
n
to
tack
le
th
ese
co
n
d
itio
n
s
a
n
d
SMOT
E
was o
b
s
er
v
ed
t
o
h
av
e
b
ee
n
wid
ely
u
s
ed
an
d
d
ev
el
o
p
ed
in
p
r
e
v
io
u
s
s
tu
d
ies.
T
h
e
r
e
f
o
r
e
,
a
co
m
p
ar
is
o
n
was
m
ad
e
b
etwe
en
th
e
p
r
o
p
o
s
ed
R
-
SMOT
E
an
d
two
p
r
ev
io
u
s
s
tate
-
of
-
th
e
-
a
r
t
m
eth
o
d
s
in
clu
d
in
g
SMOT
E
an
d
B
o
r
d
er
lin
e
-
SMOT
E
.
T
h
e
p
er
f
o
r
m
an
ce
ev
alu
atio
n
was
co
n
d
u
cted
u
s
in
g
f
o
u
r
m
etr
ics
in
clu
d
in
g
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
m
ea
s
u
r
e
as in
d
icate
d
in
t
h
e
f
o
llo
win
g
T
a
b
le
4
wh
er
e
th
e
h
ig
h
est r
esu
lts
ar
e
h
ig
h
lig
h
ted
in
b
o
ld
f
o
n
t.
R
-
SMOT
E
was
o
b
s
er
v
ed
to
h
av
e
th
e
b
est
p
er
f
o
r
m
an
ce
b
y
ac
h
iev
in
g
th
e
h
ig
h
est
ac
cu
r
ac
y
v
alu
es
in
9
d
atasets
in
clu
d
in
g
0
4
clo
v
er
5
z
-
600
-
5
-
70
-
B
I
,
n
ew
-
th
y
r
o
id
,
win
e,
y
ea
s
t
-
1
-
4
-
5
-
8
_
v
s
_
7
,
u
m
b
ilical
co
r
d
,
b
r
ea
s
t,
h
ab
er
m
an
,
p
im
a,
a
n
d
b
u
p
a.
T
h
e
m
o
s
t
s
ig
n
if
ican
t
d
if
f
er
en
ce
in
ac
cu
r
ac
y
was
r
ec
o
r
d
e
d
in
th
e
B
u
p
a
d
ataset
wh
er
e
R
-
SMOT
E
h
ad
8
3
.
2
%
wh
ile
B
o
r
d
er
lin
e
a
n
d
SMOT
E
attain
ed
5
7
.
3
%
an
d
6
4
.
3
%,
r
esp
ec
tiv
ely
.
Mo
r
eo
v
e
r
,
f
r
o
m
th
e
an
aly
s
is
o
f
r
ec
all
v
alu
es,
R
-
SMOT
E
co
n
s
is
ten
tly
o
u
tp
er
f
o
r
m
ed
th
e
o
th
er
th
r
ee
m
eth
o
d
s
.
T
ab
le
4
.
R
esu
lt
o
f
co
m
b
in
atio
n
s
o
f
W
F
ANN
alg
o
r
ith
m
with
o
v
er
s
am
p
lin
g
SMOT
E
,
B
o
r
d
e
r
lin
e
-
SMOT
E
,
an
d
R
-
SMOT
E
D
a
t
a
s
e
t
A
c
c
u
r
a
c
y
R
e
c
a
l
l
P
r
e
c
i
s
i
o
n
F
-
1
S
c
o
r
e
S
M
T
B
D
R
R
S
M
T
S
M
T
B
D
R
R
S
M
T
S
M
T
B
D
R
R
S
M
T
S
M
T
B
D
R
R
S
M
T
0
3
s
u
b
c
l
5
-
6
0
0
-
5
-
70
-
BI
6
6
.
4
6
7
.
2
6
5
.
1
6
6
.
2
6
8
.
1
6
4
.
9
6
6
.
3
6
7
.
2
6
6
.
2
6
6
.
2
6
8
.
3
6
6
.
4
0
4
c
l
o
v
e
r
5
z
-
600
-
5
-
70
-
BI
7
6
.
1
7
3
.
2
8
2
.
5
7
6
.
1
7
3
.
5
8
2
.
5
7
5
.
6
7
1
.
2
8
2
.
5
7
6
.
1
7
1
.
3
8
3
.
2
Ec
o
l
i
-
0
-
1
-
3
-
7
_
v
s_
2
-
6
9
7
.
1
9
9
.
1
9
8
.
6
9
7
.
1
9
9
.
5
9
7
.
2
9
8
.
2
9
9
.
1
9
8
.
5
9
8
.
1
9
9
.
1
9
8
.
5
G
l
a
ss
1
7
9
.
1
8
8
.
3
7
4
.
2
7
9
.
2
8
7
.
3
7
4
.
3
7
9
.
1
8
7
.
3
7
3
.
3
8
0
.
1
8
6
.
8
7
4
.
1
N
e
w
t
h
y
r
o
i
d
9
7
.
5
9
9
.
5
9
9
.
5
9
7
.
5
9
9
.
5
9
9
.
5
9
7
.
2
9
9
.
3
9
9
.
2
9
7
.
1
9
8
.
9
9
8
.
9
P
a
w
0
2
a
-
600
-
5
-
70
-
BI
7
6
.
2
7
4
.
6
7
7
.
2
7
6
.
1
7
5
.
2
7
8
.
2
7
5
.
9
7
5
.
2
7
7
.
2
7
6
.
7
7
4
.
9
7
8
.
2
W
i
n
e
9
5
.
4
9
5
.
4
1
0
0
9
5
.
2
9
5
.
5
1
0
0
9
4
.
8
9
4
.
8
1
0
0
9
4
.
8
9
4
.
8
1
0
0
Y
e
a
st
-
1
-
4
-
5
-
8
_
v
s
_
7
8
9
.
2
9
2
.
1
9
6
.
1
8
9
.
2
9
2
.
5
9
5
.
8
8
9
.
2
9
1
.
5
9
6
.
1
8
8
.
9
9
2
.
5
9
5
.
1
U
mb
i
l
i
c
a
l
C
o
r
d
9
6
.
3
9
6
.
2
9
6
.
3
9
6
.
3
9
6
.
2
9
6
.
3
9
6
.
3
9
6
.
4
9
7
.
2
9
6
.
3
9
7
.
7
9
8
.
1
B
r
e
a
s
t
7
4
.
3
7
6
.
2
8
7
.
2
7
5
.
1
7
6
.
2
8
6
.
4
7
4
.
3
7
6
.
2
8
6
.
2
7
3
.
4
7
6
.
3
8
8
.
3
H
a
b
e
r
m
a
n
7
1
.
2
6
5
.
1
7
8
.
3
7
2
.
1
6
5
.
1
7
9
.
2
7
3
.
2
6
6
.
7
8
0
.
2
7
3
.
2
6
6
.
8
8
1
.
6
P
i
ma
7
9
.
2
7
5
.
6
8
7
.
3
7
9
.
1
7
5
.
3
8
6
.
9
7
9
.
2
7
6
.
3
8
7
.
1
8
0
.
1
7
7
.
8
8
8
.
7
B
u
p
a
6
7
.
3
6
4
.
3
8
3
.
2
6
7
.
3
6
4
.
2
8
2
.
1
6
7
.
2
6
4
.
2
8
2
.
3
6
6
.
8
6
3
.
4
8
1
.
8
S
M
T=
S
M
O
TE
;
B
D
R
=
B
o
r
d
e
r
l
i
n
e
-
S
M
O
TE;
R
S
M
T=
R
-
S
M
O
TE
T
h
e
co
m
b
in
atio
n
m
eth
o
d
o
f
W
FVA
NN
an
d
R
-
SMOT
E
w
as
o
b
s
er
v
ed
to
h
av
e
p
r
o
d
u
ce
d
th
e
h
ig
h
est
v
alu
es
in
1
0
o
u
t
o
f
1
3
d
atasets
in
clu
d
in
g
0
4
clo
v
er
5
z
-
6
0
0
-
5
-
70
-
B
I
,
Paw0
2
a
-
6
0
0
-
5
-
70
-
B
I
,
N
ew
-
th
y
r
o
id
,
W
in
e,
Yea
s
t
-
1
-
4
-
5
-
8
_
v
s
_
7
,
Um
b
ilical
C
o
r
d
,
B
r
ea
s
t,
Hab
er
m
an
,
Pima
,
an
d
B
u
p
a.
T
h
is
u
n
d
e
r
s
co
r
ed
th
e
alig
n
m
en
t
b
etwe
en
th
e
test
d
ata
f
ac
ts
an
d
th
e
p
r
ed
ictio
n
o
u
tco
m
es
o
f
W
FVANN
an
d
R
-
SM
OT
E
,
th
er
eb
y
p
r
o
d
u
cin
g
s
atis
f
y
in
g
p
er
f
o
r
m
an
ce
im
p
r
o
v
em
en
ts
.
Mo
r
eo
v
er
,
th
e
p
r
ec
is
io
n
r
esu
lts
also
m
ir
r
o
r
ed
th
is
t
r
en
d
,
with
1
0
o
u
t
o
f
1
3
d
atasets
attain
in
g
th
e
h
ig
h
est
v
alu
es
th
r
o
u
g
h
W
FVANN
R
-
SMOT
E
m
eth
o
d
.
T
h
e
m
o
s
t
s
ig
n
if
ican
t
in
cr
ea
s
e
in
p
r
ec
is
io
n
was
o
b
s
er
v
ed
in
th
e
B
u
p
a
d
ataset,
with
8
2
.
3
%
r
ec
o
r
d
e
d
f
o
r
R
-
SMOT
E
,
6
4
.
2
%
f
o
r
B
o
r
d
er
lin
e
-
SMOT
E
,
an
d
6
7
.
2
%
f
o
r
SMOT
E
.
T
h
e
r
esu
lts
s
h
o
wed
th
e
alig
n
m
en
t
b
etwe
en
t
h
e
p
r
ed
ic
tio
n
o
u
tco
m
e
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
an
d
th
e
ac
t
u
al
d
ata
an
d
wer
e
f
o
u
n
d
to
b
e
s
atis
f
ac
to
r
y
f
o
r
b
o
th
th
e
n
eg
ativ
e
a
n
d
p
o
s
itiv
e
class
es.
Similar
ly
,
th
e
r
esu
lts
wer
e
r
ef
l
ec
ted
in
th
e
F1
-
Me
asu
r
e
m
etr
i
c
with
th
e
p
r
o
p
o
s
ed
co
m
b
i
n
atio
n
m
o
d
el
d
is
co
v
e
r
ed
to
h
av
e
ac
h
ie
v
ed
th
e
h
i
g
h
est
v
alu
e
in
1
0
d
atasets
in
clu
d
in
g
0
4
clo
v
er
5
z
-
600
-
5
-
70
-
B
I
,
Pa
w0
2
a
-
6
0
0
-
5
-
70
-
B
I
,
New
-
th
y
r
o
id
,
W
in
e,
Yea
s
t
-
1
-
4
-
5
-
8
_
v
s
_
7
,
Um
b
ilical
C
o
r
d
,
B
r
ea
s
t,
Hab
er
m
a
n
,
Pima
,
an
d
B
u
p
a.
T
h
e
F1
-
m
ea
s
u
r
e
also
im
p
lied
th
e
s
im
u
ltan
eo
u
s
m
a
x
im
izatio
n
o
f
b
o
th
p
r
ec
is
io
n
a
n
d
r
ec
all
th
at
o
f
f
er
ed
a
tr
a
d
e
-
o
f
f
with
o
n
e
m
etr
ic
co
m
in
g
at
th
e
co
s
t
o
f
an
o
th
er
.
Mo
r
e
p
r
ec
is
io
n
in
v
o
lv
ed
a
h
a
r
s
h
er
cr
itic
o
r
c
lass
if
ier
th
at
d
o
u
b
ts
ev
en
th
e
ac
tu
al
p
o
s
itiv
e
s
am
p
les
f
r
o
m
th
e
d
ataset,
th
er
eb
y
r
ed
u
cin
g
th
e
r
ec
all
s
co
r
e.
Me
a
n
w
h
ile,
m
o
r
e
r
ec
all
en
tailed
lax
cr
itic
wh
ich
allo
wed
an
y
s
am
p
le
r
esem
b
lin
g
a
p
o
s
itiv
e
clas
s
to
p
ass
an
d
m
ad
e
b
o
r
d
er
-
ca
s
e
n
e
g
ativ
e
s
am
p
les
clas
s
if
ied
as
“p
o
s
itiv
e
”,
r
ed
u
cin
g
t
h
e
p
r
ec
is
io
n
.
T
h
e
co
m
b
in
atio
n
o
f
W
FVANN
an
d
R
-
SMOT
E
m
o
d
els
ef
f
ec
tiv
ely
b
alan
ce
d
an
d
m
ax
im
ized
p
r
ec
is
io
n
an
d
r
ec
all
v
alu
es
b
ased
o
n
th
e
ex
p
e
r
im
e
n
t
r
esu
lts
s
h
o
wn
in
T
ab
les
3
an
d
4
.
Fu
r
th
e
r
m
o
r
e
,
Fig
u
r
e
4
s
h
o
ws
th
e
c
o
m
p
ar
is
o
n
o
f
ac
c
u
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
al
l,
an
d
F1
-
m
ea
s
u
r
e
o
u
tco
m
es
f
o
r
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
n
th
e
U
m
b
i
l
i
c
a
l
c
o
r
d
,
i
n
d
i
c
a
t
i
n
g
t
h
e
m
o
s
t
s
i
g
n
if
i
c
a
n
t
p
e
r
f
o
r
m
a
n
c
e
e
n
h
a
n
c
e
m
e
n
t
c
o
m
p
a
r
e
d
t
o
o
t
h
e
r
d
a
t
a
a
n
d
m
e
t
h
o
d
s
.
T
h
is
p
lo
t
s
h
o
ws
th
e
ac
cu
r
ac
y
in
Fi
g
u
r
e
4
(
a)
,
r
ec
all
in
Fig
u
r
e
4
(
b
)
,
p
r
ec
is
io
n
in
Fig
u
r
e
4
(
c)
,
an
d
F1
-
m
ea
s
u
r
e
in
Fig
u
r
e
4
(
d
)
o
f
W
FVANN
m
o
d
el
u
s
in
g
d
if
f
e
r
en
t
o
v
er
s
am
p
lin
g
tech
n
iq
u
es
as
th
e
n
u
m
b
er
o
f
k
n
eig
h
b
o
r
s
c
h
an
g
es.
SMOT
E
g
en
e
r
ally
p
er
f
o
r
m
s
well
ac
r
o
s
s
d
if
f
er
en
t
v
alu
es
o
f
k
,
m
ain
ta
in
in
g
h
i
g
h
er
ac
cu
r
ac
y
co
m
p
a
r
ed
to
th
e
o
th
er
m
eth
o
d
s
.
B
o
r
d
er
lin
e
SMOT
E
ex
p
er
ien
ce
s
s
i
g
n
if
ica
n
t
d
ip
s
at
k
=6
a
n
d
k
=1
0
,
s
h
o
win
g
u
n
s
tab
le
b
eh
av
i
o
r
with
d
i
f
f
e
r
en
t
v
alu
es
o
f
k
.
R
-
SMOT
E
r
em
ain
s
q
u
ite
s
tab
le
b
u
t te
n
d
s
to
s
lig
h
tly
u
n
d
e
r
p
er
f
o
r
m
c
o
m
p
ar
ed
to
SMOT
E
at
h
ig
h
er
k
v
a
lu
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
Weig
h
ted
n
ea
r
est n
eig
h
b
o
r
s
a
n
d
r
a
d
iu
s
o
ve
r
s
a
mp
lin
g
f
o
r
imb
a
la
n
ce
d
…
(
Ged
e
A
n
g
g
a
P
r
a
d
ip
ta
)
425
(
a)
(
b
)
(
c)
(
d
)
Fig
u
r
e
4
.
C
o
m
p
a
r
is
o
n
o
f
th
e
r
esu
lts
o
f
W
FVAN
N
co
m
b
in
atio
n
with
th
e
o
v
er
s
am
p
lin
g
alg
o
r
ith
m
in
(
a)
ac
cu
r
ac
y
,
(
b
)
r
ec
all
,
(
c
)
p
r
e
cisi
o
n
,
an
d
(
d
)
F1
-
m
e
asu
r
e
o
n
th
e
u
m
b
ilical
co
r
d
d
ataset
4.
CO
NCLU
SI
O
N
I
n
co
n
clu
s
io
n
,
th
is
s
tu
d
y
in
tr
o
d
u
ce
d
an
in
n
o
v
ativ
e
KNN
v
a
r
ian
t
n
am
ed
W
FVANN.
T
h
e
ess
en
ce
o
f
th
is
m
eth
o
d
d
ep
en
d
ed
o
n
th
e
o
b
s
er
v
atio
n
o
f
th
e
r
elev
an
ce
an
d
co
n
tr
ib
u
tio
n
o
f
ea
ch
f
ea
tu
r
e
to
war
d
th
e
ca
lcu
latio
n
o
f
E
u
clid
ea
n
d
is
tan
ce
in
KNN
m
eth
o
d
.
T
h
e
f
ea
tu
r
e
r
e
lev
an
ce
a
n
d
co
n
tr
ib
u
tio
n
w
er
e
m
ea
s
u
r
ed
u
s
in
g
th
e
F
-
v
alu
e
an
d
weig
h
t
o
f
ea
ch
f
ea
tu
r
e,
wh
ich
d
ep
en
d
ed
o
n
th
e
m
a
g
n
itu
d
e
o
f
th
e
r
esu
ltin
g
F
-
v
alu
e.
T
h
e
ev
alu
atio
n
was
co
n
d
u
cte
d
u
s
in
g
1
0
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
(
1
0
-
FC
V)
with
ex
p
er
im
en
ts
ap
p
lied
to
th
e
two
ty
p
es
of
d
ata
in
clu
d
i
n
g
h
ig
h
-
d
im
e
n
s
io
n
al
an
d
im
b
alan
ce
d
.
T
h
e
e
x
p
er
im
en
ts
o
n
h
ig
h
-
d
im
en
s
io
n
al
d
atasets
s
h
o
wed
th
at
W
FVANN
o
u
tp
er
f
o
r
m
ed
o
th
er
m
eth
o
d
s
in
clu
d
i
n
g
KNN
an
d
L
MK
NN.
T
h
is
was
co
n
f
ir
m
ed
b
y
th
e
f
ac
t
t
h
at
W
FVA
NN
m
o
d
el
y
ield
ed
s
ati
s
f
ac
to
r
y
r
esu
lts
with
6
o
u
t o
f
1
0
d
atasets
ac
h
iev
in
g
th
e
h
ig
h
est v
alu
es c
o
m
p
ar
ed
to
o
th
er
m
eth
o
d
s
.
T
h
e
p
h
en
o
m
en
o
n
in
d
icate
d
th
e
ef
f
ec
tiv
en
ess
o
f
u
s
in
g
weig
h
ts
f
o
r
r
elev
an
t
f
ea
tu
r
es
in
d
eter
m
in
in
g
p
r
ed
ictio
n
o
u
tco
m
es.
T
h
e
r
esu
lts
al
s
o
s
h
o
wed
th
at
n
o
t
all
f
ea
tu
r
es
co
n
tr
ib
u
ted
v
a
lu
ab
le
in
f
o
r
m
atio
n
in
d
eter
m
in
in
g
d
ata
p
atter
n
s
i
n
h
ig
h
-
d
im
en
s
io
n
al
d
atasets
,
b
u
t
s
o
m
e
h
ad
th
e
ca
p
ac
ity
to
d
is
r
u
p
t
th
e
lear
n
in
g
p
r
o
ce
s
s
.
T
h
e
co
m
b
i
n
atio
n
o
f
R
-
SMOT
E
o
v
er
s
am
p
lin
g
m
et
h
o
d
at
th
e
d
ata
lev
el
an
d
W
FVANN
m
eth
o
d
at
th
e
alg
o
r
ith
m
ic
lev
el
was p
r
o
p
o
s
e
d
in
t
h
e
test
o
f
im
b
alan
ce
d
d
at
asets
an
d
in
d
icate
d
s
atis
f
y
in
g
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
m
ea
s
u
r
e
v
alu
es.
R
-
SMOT
E
m
eth
o
d
s
h
o
wed
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
m
etr
ics
co
m
p
ar
ed
to
SMOT
E
an
d
B
o
r
d
er
lin
e
-
SMOT
E
.
T
h
e
r
esu
lts
v
alid
ated
th
e
ef
f
ec
tiv
en
ess
o
f
co
n
s
tr
ain
in
g
th
e
ar
ea
in
R
-
SMOT
E
an
d
m
o
d
if
y
in
g
f
ea
tu
r
e
weig
h
ts
in
W
FV
ANN
to
en
h
an
ce
r
o
b
u
s
tn
ess
ag
ain
s
t
im
b
alan
ce
d
d
ata
co
n
d
itio
n
s
.
T
h
e
lim
itatio
n
s
o
f
th
is
s
tu
d
y
wer
e
also
ac
k
n
o
wled
g
ed
.
T
h
e
co
m
p
u
tatio
n
al
tim
e
was
r
elativ
ely
h
ig
h
d
u
e
to
th
e
ca
lcu
latio
n
o
f
f
ea
tu
r
e
weig
h
ts
f
o
r
ea
c
h
f
ea
tu
r
e,
p
ar
ticu
lar
ly
in
h
ig
h
-
d
im
en
s
io
n
al
d
ata.
T
h
er
ef
o
r
e,
s
p
ec
ial
atten
tio
n
was
r
eq
u
ir
ed
to
ad
d
r
ess
co
m
p
u
tatio
n
al
tim
e
co
n
s
tr
ain
ts
to
f
u
r
th
er
r
e
f
in
e
th
is
m
o
d
el.
I
n
th
e
f
u
tu
r
e,
f
ea
tu
r
e
weig
h
tin
g
d
e
v
elo
p
m
e
n
t
s
h
o
u
ld
b
e
co
m
b
in
ed
with
o
th
er
f
ea
tu
r
e
s
elec
tio
n
m
et
h
o
d
s
,
an
d
d
is
tan
ce
ca
lcu
latio
n
s
ex
p
lo
r
e
d
u
s
in
g
alter
n
ativ
e
m
eth
o
d
s
s
u
ch
as
Min
k
o
wsk
i
an
d
Ma
n
h
attan
d
is
tan
ce
s
.
Fu
tu
r
e
d
ev
elo
p
m
e
n
ts
s
h
o
u
ld
also
f
o
c
u
s
o
n
d
eter
m
in
in
g
t
h
e
m
o
s
t o
p
tim
al
k
v
alu
e
au
to
m
atica
lly
.
ACK
NO
WL
E
DG
E
M
E
NT
S
T
h
e
au
th
o
r
s
ar
e
g
r
ate
f
u
l
to
th
e
Dir
ec
to
r
ate
o
f
R
esear
ch
,
T
ec
h
n
o
lo
g
y
,
an
d
C
o
m
m
u
n
ity
Ser
v
i
ce
(
DPR
M
)
I
n
d
o
n
esia
f
o
r
f
u
n
d
i
n
g
th
is
s
tu
d
y
th
r
o
u
g
h
Pro
g
r
am
Fu
n
d
in
g
(
Fu
d
am
en
tal
R
eg
u
ler
)
f
o
r
th
e
2
0
2
3
Fis
ca
l
Yea
r
.
Nex
t,
th
an
k
y
o
u
t
o
th
e
r
esear
ch
ass
is
tan
ts
,
n
am
ely
Ar
y
a
Fais
al
Ak
b
ar
an
d
Hen
d
r
a
W
ijay
a
wh
o
h
av
e
h
elp
ed
ca
r
r
y
o
u
t th
is
r
esear
ch
in
t
h
e
i
n
tellig
en
t sy
s
tem
s
lab
o
r
ato
r
y
o
f
I
n
s
titu
t T
ek
n
o
lo
g
i
d
an
B
is
n
is
STI
KOM
B
al
i.
Evaluation Warning : The document was created with Spire.PDF for Python.