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.
1
5
,
No
.
3
,
J
u
n
e
2
0
2
6
,
p
p
.
2935
~
2
9
4
5
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
15
.i
3
.
p
p
2
9
3
5
-
2
9
4
5
2935
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Enha
ncing
phishi
ng
website
det
ecti
o
n:
a
com
pa
ra
tive study
of
SM
O
TE
To
mek
-
X
G
B a
nd SM
O
TE
ENN
-
X
G
B
K
a
m
a
l
O
ma
ri
1
,
Ay
o
ub
O
uk
ha
t
a
r
2
1
La
b
o
r
a
t
o
r
y
o
f
C
o
m
p
u
t
e
r
S
y
s
t
e
m
s a
n
d
V
i
s
i
o
n
(
La
b
S
I
V
)
,
D
e
p
a
r
t
m
e
n
t
o
f
C
o
mp
u
t
e
r
S
c
i
e
n
c
e
,
P
o
l
y
d
i
sc
i
p
l
i
n
a
r
y
F
a
c
u
l
t
y
o
f
O
u
a
r
z
a
z
a
t
e
,
U
n
i
v
e
r
si
t
y
I
b
n
Zo
h
r
,
O
u
a
r
z
a
z
a
t
e
,
M
o
r
o
c
c
o
2
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
,
H
i
g
h
e
r
S
c
h
o
o
l
o
f
Te
c
h
n
o
l
o
g
y
O
u
a
r
z
a
z
a
t
e
,
U
n
i
v
e
r
si
t
y
I
b
n
Zo
h
r
,
O
u
a
r
z
a
z
a
t
e
,
M
o
r
o
c
c
o
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Sep
8
,
2
0
2
4
R
ev
is
ed
Ma
r
6
,
2
0
2
6
Acc
ep
ted
Ap
r
2
2
,
2
0
2
6
In
t
h
e
e
v
o
lv
in
g
lan
d
sc
a
p
e
o
f
c
y
b
e
rse
c
u
rit
y
,
p
h
is
h
in
g
we
b
sites
c
o
n
ti
n
u
e
to
b
e
a
p
e
rsiste
n
t
t
h
re
a
t,
c
h
a
ll
e
n
g
i
n
g
d
e
tec
ti
o
n
m
e
th
o
d
s
d
u
e
to
th
e
si
g
n
ifi
c
a
n
t
c
las
s
imb
a
lan
c
e
b
e
twe
e
n
p
h
is
h
in
g
a
n
d
leg
it
ima
te
we
b
sites
.
Th
is
stu
d
y
e
v
a
lu
a
tes
th
e
e
ffe
c
ti
v
e
n
e
ss
o
f
two
a
d
v
a
n
c
e
d
h
y
b
ri
d
-
re
sa
m
p
li
n
g
tec
h
n
iq
u
e
s
S
M
OTET
o
m
e
k
a
n
d
S
M
OTEE
N
N
in
teg
ra
ted
wi
th
t
h
e
e
x
trem
e
g
ra
d
ien
t
b
o
o
st
in
g
(
XG
Bo
o
st
)
c
las
sifier
to
e
n
h
a
n
c
e
p
h
ish
i
n
g
we
b
site
d
e
tec
ti
o
n
.
S
M
OTET
o
m
e
k
c
o
m
b
in
e
s
th
e
sy
n
th
e
ti
c
m
in
o
rit
y
o
v
e
r
-
sa
m
p
li
n
g
tec
h
n
iq
u
e
(S
M
OTE)
with
To
m
e
k
li
n
ks
,
c
r
e
a
ti
n
g
sy
n
th
e
ti
c
e
x
a
m
p
les
a
n
d
e
li
m
in
a
ti
n
g
o
v
e
rlap
p
in
g
i
n
sta
n
c
e
s
to
a
d
d
re
ss
d
a
tas
e
t
imb
a
lan
c
e
.
S
M
OTEE
N
N,
o
n
th
e
o
th
e
r
h
a
n
d
,
m
e
rg
e
s S
M
OTE
with
e
d
it
e
d
n
e
a
re
st n
e
ig
h
b
o
rs (E
NN
)
t
o
imp
ro
v
e
c
las
s
b
a
lan
c
e
th
ro
u
g
h
sy
n
th
e
ti
c
s
a
m
p
le
g
e
n
e
ra
ti
o
n
a
n
d
n
o
ise
re
d
u
c
ti
o
n
.
Th
e
c
o
m
p
a
ra
ti
v
e
a
n
a
ly
sis
re
v
e
a
ls
t
h
a
t
b
o
th
m
e
th
o
d
s
sig
n
ifi
c
a
n
t
ly
e
n
h
a
n
c
e
c
las
sifica
ti
o
n
p
e
rfo
rm
a
n
c
e
,
S
M
OTET
o
m
e
k
-
XG
B
c
o
n
siste
n
tl
y
o
u
tp
e
rfo
rm
s
S
M
OTEE
NN
-
XG
B
a
c
ro
ss
k
e
y
e
v
a
lu
a
ti
o
n
m
e
tri
c
s,
in
c
lu
d
in
g
a
c
c
u
ra
c
y
,
F
1
-
sc
o
re
,
re
c
a
ll
,
a
n
d
re
c
e
iv
e
r
o
p
e
ra
ti
n
g
c
h
a
ra
c
teristic
-
a
re
a
u
n
d
e
r
th
e
c
u
rv
e
(ROC
-
AU
C)
,
u
n
d
e
rsc
o
rin
g
i
ts
su
p
e
ri
o
r
e
ffe
c
ti
v
e
n
e
ss
in
d
ist
in
g
u
ish
i
n
g
p
h
ish
in
g
sites
fro
m
leg
it
ima
te
o
n
e
s.
Th
is
stu
d
y
o
ffe
rs
p
ra
c
ti
c
a
l
i
n
s
ig
h
ts
in
to
th
e
a
p
p
li
c
a
ti
o
n
o
f
a
d
v
a
n
c
e
d
re
sa
m
p
li
n
g
m
e
th
o
d
s
f
o
r
imp
ro
v
i
n
g
m
a
c
h
in
e
lea
rn
in
g
m
o
d
e
l
p
e
rfo
rm
a
n
c
e
in
c
y
b
e
rse
c
u
rit
y
.
K
ey
w
o
r
d
s
:
C
las
s
im
b
alan
ce
Ph
is
h
in
g
web
s
ite
d
etec
tio
n
SMOT
E
E
NN
tech
n
iq
u
es
SMOT
E
T
o
m
ek
tech
n
iq
u
es
XGBo
o
s
t c
las
s
if
ier
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
:
Kam
al
Om
ar
i
L
ab
o
r
ato
r
y
o
f
C
o
m
p
u
ter
Sy
s
tem
s
an
d
Vis
io
n
(
L
ab
SIV
)
,
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scien
ce
Po
ly
d
is
cip
lin
ar
y
Facu
lty
o
f
O
u
ar
za
za
te,
Un
iv
er
s
ity
I
b
n
Z
o
h
r
Ou
ar
za
za
te,
Mo
r
o
cc
o
E
m
ail:
k
.
o
m
ar
i@
u
iz.
ac
.
m
a
1.
I
NT
RO
D
UCT
I
O
N
Ph
is
h
in
g
attac
k
s
p
o
s
e
a
s
ig
n
if
ican
t
an
d
g
r
o
win
g
th
r
ea
t
in
cy
b
er
s
ec
u
r
ity
,
with
cy
b
er
cr
im
in
als
em
p
lo
y
in
g
in
cr
ea
s
in
g
ly
s
o
p
h
is
ticated
tactics
to
d
ec
eiv
e
i
n
d
iv
id
u
als
in
t
o
r
e
v
ea
lin
g
s
en
s
itiv
e
in
f
o
r
m
atio
n
.
T
h
ese
p
h
is
h
in
g
web
s
ites
m
im
ic
leg
itima
te
o
n
lin
e
p
latf
o
r
m
s
,
p
u
ttin
g
b
o
th
in
d
i
v
id
u
als
an
d
o
r
g
an
izatio
n
s
at
r
is
k
b
y
tr
ick
i
n
g
u
s
er
s
in
to
d
i
v
u
lg
i
n
g
p
e
r
s
o
n
al
d
etails
o
r
cr
e
d
en
tials
[
1
]
.
Dete
ctin
g
t
h
ese
f
r
au
d
u
len
t
web
s
ites
is
cr
u
cial
f
o
r
cy
b
er
s
ec
u
r
ity
ex
p
er
ts
,
y
et
tr
ad
itio
n
al
d
etec
tio
n
m
eth
o
d
s
o
f
ten
f
all
s
h
o
r
t
d
u
e
to
th
e
p
r
o
n
o
u
n
ce
d
class
im
b
alan
ce
in
p
h
is
h
in
g
d
a
tasets
[
2
]
.
T
h
is
im
b
alan
ce
ar
is
es
b
ec
au
s
e
p
h
is
h
in
g
web
s
ites
ar
e
m
u
ch
less
co
m
m
o
n
th
a
n
leg
iti
m
ate
o
n
es,
lead
in
g
to
a
s
k
ewe
d
class
d
is
tr
ib
u
tio
n
th
at
h
am
p
er
s
th
e
ef
f
ec
t
iv
en
ess
o
f
class
if
icatio
n
alg
o
r
ith
m
s
[
3
]
.
Stan
d
ar
d
m
ac
h
in
e
lear
n
in
g
class
if
ier
s
ten
d
to
b
e
b
iased
to
war
d
th
e
m
ajo
r
ity
class
,
o
f
ten
r
esu
ltin
g
in
p
o
o
r
d
etec
tio
n
o
f
p
h
is
h
in
g
web
s
ites
[
2
]
.
T
h
is
is
s
u
e
h
ig
h
lig
h
ts
th
e
u
r
g
en
t
n
ee
d
f
o
r
a
d
v
an
ce
d
tec
h
n
iq
u
es
th
at
ca
n
a
d
d
r
ess
th
is
im
b
alan
ce
an
d
im
p
r
o
v
e
d
etec
t
io
n
ac
cu
r
ac
y
.
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.
1
5
,
No
.
3
,
J
u
n
e
2
0
2
6
:
2
9
3
5
-
2
9
4
5
2936
T
o
tack
le
th
ese
ch
allen
g
es,
r
esear
ch
er
s
h
av
e
d
e
v
elo
p
e
d
v
ar
io
u
s
r
esam
p
lin
g
tec
h
n
iq
u
es
aim
ed
at
ad
ju
s
tin
g
class
d
is
tr
ib
u
tio
n
an
d
en
h
an
cin
g
m
o
d
el
p
er
f
o
r
m
an
ce
.
Am
o
n
g
th
ese,
two
n
o
tab
le
h
y
b
r
id
-
r
esam
p
lin
g
m
eth
o
d
s
ar
e
SMOT
E
T
o
m
ek
an
d
SMOT
E
E
NN,
b
o
th
o
f
wh
ich
h
av
e
s
h
o
wn
p
r
o
m
is
e
in
b
alan
cin
g
class
d
is
tr
ib
u
tio
n
s
an
d
im
p
r
o
v
in
g
c
lass
if
icatio
n
r
esu
lts
ac
r
o
s
s
d
if
f
er
en
t
ap
p
licatio
n
s
[
4
]
.
SMOT
E
T
o
m
ek
c
o
m
b
in
es
th
e
s
y
n
th
etic
m
in
o
r
ity
o
v
er
-
s
am
p
lin
g
tech
n
iq
u
e
(
SMOT
E
)
with
T
o
m
ek
l
in
k
s
.
SMOT
E
g
en
er
ates
s
y
n
th
etic
s
am
p
les
f
o
r
th
e
m
in
o
r
ity
class
to
ad
d
r
ess
im
b
alan
ce
,
wh
ile
T
o
m
ek
l
in
k
s
r
em
o
v
e
am
b
ig
u
o
u
s
in
s
tan
ce
s
n
ea
r
class
b
o
u
n
d
ar
ies,
r
ef
i
n
in
g
th
e
s
ep
ar
atio
n
b
etwe
en
class
es
[
5
]
.
T
h
is
d
u
al
ap
p
r
o
ac
h
n
o
t
o
n
ly
s
tr
en
g
th
e
n
s
th
e
m
in
o
r
ity
class
r
ep
r
esen
tatio
n
b
u
t a
ls
o
r
ef
in
es th
e
d
ataset,
p
o
ten
tially
en
h
an
cin
g
class
if
ier
p
er
f
o
r
m
a
n
ce
.
C
o
n
v
er
s
ely
,
SMOT
E
E
NN
in
teg
r
ates
SMOT
E
with
ed
ited
n
ea
r
est
n
eig
h
b
o
r
s
(
E
NN)
.
W
h
ile
SMOT
E
g
en
er
ates
s
y
n
th
etic
in
s
tan
ce
s
o
f
th
e
m
in
o
r
ity
class
,
E
NN
r
e
d
u
ce
s
n
o
is
e
b
y
r
em
o
v
in
g
in
s
tan
ce
s
m
is
clas
s
if
ied
b
y
th
eir
n
ea
r
est
n
eig
h
b
o
r
s
[
5
]
.
T
h
is
m
eth
o
d
s
ee
k
s
to
b
alan
ce
o
v
er
s
am
p
lin
g
with
d
ata
clea
n
in
g
,
o
f
f
er
in
g
an
ef
f
ec
tiv
e
s
tr
ateg
y
f
o
r
d
ea
lin
g
with
im
b
alan
ce
d
d
ata.
Desp
ite
th
e
p
o
ten
tial
o
f
th
ese
tech
n
iq
u
es,
th
eir
ef
f
ec
tiv
en
ess
in
d
etec
tin
g
p
h
is
h
in
g
web
s
ites
r
em
ain
s
u
n
clea
r
.
C
o
m
p
ar
in
g
d
if
f
er
e
n
t
u
n
d
er
-
an
d
o
v
er
s
am
p
lin
g
m
eth
o
d
s
ca
n
p
r
o
v
id
e
p
r
ac
tic
al
in
s
ig
h
ts
f
o
r
d
ata
s
cien
tis
t
s
an
d
cy
b
er
s
ec
u
r
ity
ex
p
er
ts
,
b
r
i
d
g
in
g
th
e
g
a
p
b
etwe
en
th
eo
r
etica
l
r
esear
ch
an
d
r
ea
l
-
wo
r
ld
ap
p
licatio
n
s
.
T
h
is
s
tu
d
y
ad
d
r
ess
es
th
is
g
ap
b
y
co
n
d
u
c
tin
g
a
co
m
p
r
eh
en
s
iv
e
co
m
p
ar
ativ
e
an
aly
s
is
o
f
SMOT
E
T
o
m
ek
-
XGB
an
d
S
MO
T
E
E
NN
-
XGB
f
o
r
d
etec
tin
g
p
h
is
h
in
g
web
s
ites
.
E
x
tr
e
m
e
g
r
ad
ie
n
t
b
o
o
s
tin
g
(
XGBo
o
s
t
)
was
s
elec
ted
,
a
le
ad
in
g
g
r
a
d
ien
t
b
o
o
s
tin
g
alg
o
r
ith
m
,
f
o
r
its
p
r
o
v
e
n
ab
ilit
y
to
h
an
d
le
im
b
alan
ce
d
d
atasets
an
d
its
ex
ce
p
tio
n
al
p
er
f
o
r
m
an
ce
ac
r
o
s
s
v
ar
io
u
s
m
ac
h
in
e
-
lear
n
i
n
g
task
s
[
6
]
.
B
y
co
m
p
ar
in
g
th
e
p
er
f
o
r
m
an
ce
o
f
SMOT
E
T
o
m
ek
an
d
SMOT
E
E
NN
with
XGBo
o
s
t,
th
is
r
esear
ch
aim
s
to
id
en
tify
th
e
m
o
s
t
ef
f
ec
tiv
e
r
esam
p
lin
g
tec
h
n
iq
u
e
f
o
r
p
h
is
h
in
g
d
etec
tio
n
an
d
p
r
o
v
id
e
ac
tio
n
ab
le
in
s
ig
h
ts
f
o
r
f
u
tu
r
e
ap
p
licatio
n
s
in
cy
b
er
s
ec
u
r
ity
.
T
h
e
r
est
o
f
th
is
p
ap
er
is
o
r
g
an
ized
as
f
o
llo
ws
.
Sectio
n
2
r
ev
iews
ex
is
tin
g
liter
atu
r
e
o
n
p
h
is
h
in
g
web
s
ite
d
etec
tio
n
an
d
r
esa
m
p
lin
g
tec
h
n
iq
u
es.
Sectio
n
3
d
etails
th
e
m
eth
o
d
o
lo
g
i
es
u
s
ed
,
in
clu
d
in
g
ex
p
lan
atio
n
s
o
f
SMOT
E
T
o
m
ek
,
SMOT
E
E
NN,
th
e
XG
B
o
o
s
t
clas
s
if
ier
,
an
d
ex
p
er
im
en
t
al
d
esig
n
.
Sectio
n
4
p
r
esen
ts
th
e
p
r
o
p
o
s
ed
f
r
am
e
wo
r
k
,
in
clu
d
in
g
class
im
b
alan
ce
h
an
d
lin
g
an
d
th
e
e
v
alu
at
io
n
m
etr
ics
u
s
ed
to
ass
es
s
m
o
d
el
p
er
f
o
r
m
a
n
ce
.
Se
ctio
n
5
p
r
esen
ts
th
e
r
esu
lts
o
f
co
m
p
ar
ativ
e
an
al
y
s
is
,
h
ig
h
lig
h
tin
g
p
er
f
o
r
m
a
n
ce
m
etr
ics
f
o
r
b
o
th
SMOT
E
T
o
m
ek
-
XGB
an
d
SMOT
E
E
NN
-
X
GB
.
Fin
ally
,
s
ec
tio
n
6
d
is
cu
s
s
es
th
e
im
p
licatio
n
s
o
f
th
ese
f
in
d
i
n
g
s
f
o
r
p
h
is
h
in
g
web
s
ite
d
etec
tio
n
an
d
co
n
clu
d
es with
r
ec
o
m
m
en
d
atio
n
s
f
o
r
f
u
tu
r
e
r
esear
ch
.
2.
L
I
T
E
R
AT
U
RE
R
E
VI
E
W
2
.
1
.
P
his
hin
g
websi
t
e
det
ec
t
i
o
n
Ph
is
h
in
g
attac
k
s
ar
e
a
s
ig
n
if
ican
t
th
r
ea
t
in
cy
b
er
s
ec
u
r
ity
,
ex
p
lo
itin
g
u
s
er
s
’
tr
u
s
t
b
y
tr
ic
k
in
g
th
em
in
to
r
ev
ea
lin
g
s
en
s
itiv
e
in
f
o
r
m
atio
n
o
n
f
r
au
d
u
len
t
web
s
ites
d
esig
n
ed
to
m
im
ic
leg
itima
te
o
n
es.
Dete
ctin
g
th
ese
d
ec
ep
tiv
e
s
ites
is
ch
alle
n
g
in
g
d
u
e
to
t
h
e
ad
v
an
ce
d
te
ch
n
iq
u
es
em
p
l
o
y
ed
b
y
attac
k
er
s
an
d
th
e
in
h
er
en
t
class
im
b
alan
ce
in
p
h
is
h
in
g
d
atasets
.
T
r
ad
itio
n
ally
,
p
h
i
s
h
in
g
d
etec
tio
n
r
elied
o
n
r
u
le
-
b
ased
m
eth
o
d
s
,
wh
ich
id
en
tifie
d
k
n
o
wn
p
h
is
h
in
g
tactics
o
r
s
p
ec
if
ic
f
ea
tu
r
es
lik
e
d
o
m
ain
n
am
e
d
is
cr
e
p
an
cies
an
d
v
is
u
al
s
im
ilar
ities
.
Ho
wev
er
,
as
att
ac
k
er
s
h
av
e
r
ef
in
ed
th
eir
s
tr
ateg
ies,
th
ese
s
tatic
r
u
le
-
b
ased
ap
p
r
o
ac
h
es
h
av
e
b
ec
o
m
e
less
ef
f
ec
tiv
e
[
7
]
.
I
n
r
esp
o
n
s
e
to
th
ese
e
v
o
lv
in
g
th
r
ea
ts
,
r
ec
en
t
e
f
f
o
r
ts
h
av
e
s
h
if
ted
to
war
d
s
m
ac
h
in
e
lea
r
n
in
g
-
b
ased
m
eth
o
d
s
f
o
r
p
h
is
h
in
g
d
etec
ti
o
n
.
Dec
is
io
n
tr
ee
(
DT
)
,
lo
g
is
tic
r
eg
r
ess
io
n
(
L
R
)
,
k
-
n
ea
r
est
n
eig
h
b
o
r
s
(
KNN)
,
r
an
d
o
m
f
o
r
est
(
R
F),
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
SVM)
,
n
ai
v
e
B
ay
es
(
NB
)
,
an
d
g
r
a
d
ien
t
b
o
o
s
tin
g
ar
e
all
ex
am
p
les
o
f
alg
o
r
ith
m
s
.
h
a
v
e
b
ee
n
em
p
lo
y
ed
to
class
if
y
web
s
ites
as
p
h
is
h
in
g
o
r
leg
itima
te
b
ased
o
n
v
ar
io
u
s
f
ea
tu
r
es
[
8
]
.
Desp
ite
th
e
ad
v
an
ce
m
en
ts
in
th
ese
m
eth
o
d
s
,
th
e
is
s
u
e
o
f
class
im
b
alan
ce
r
em
ain
s
a
cr
itical
ch
allen
g
e.
Ph
is
h
in
g
web
s
ites
g
en
er
ally
r
ep
r
esen
t
o
n
ly
a
s
m
all
f
r
ac
tio
n
o
f
th
e
d
ata
,
ca
u
s
in
g
class
if
ier
s
to
p
er
f
o
r
m
p
o
o
r
ly
o
n
th
ese
in
f
r
eq
u
en
t in
s
tan
ce
s
[
9
]
.
2
.
2
.
Resa
m
pli
ng
t
ec
hn
iqu
e
s
f
o
r
im
ba
la
nced
da
t
a
T
o
ad
d
r
ess
th
e
ch
allen
g
e
o
f
c
lass
im
b
alan
ce
,
s
ev
er
al
r
esam
p
lin
g
tech
n
iq
u
es
h
av
e
b
ee
n
d
ev
elo
p
ed
.
Am
o
n
g
t
h
ese,
SMOT
E
[
1
0
]
c
o
m
b
in
ed
with
T
o
m
ek
lin
k
s
[
1
1
]
,
k
n
o
wn
as
SMOT
E
T
o
m
ek
[
1
2
]
,
an
d
SMOT
E
p
air
ed
with
E
NN
[
5
]
,
f
o
r
m
i
n
g
SMOT
E
E
NN
[
1
3
]
,
ar
e
p
ar
t
icu
lar
ly
n
o
tewo
r
t
h
y
.
C
h
awla
et
a
l.
[
1
0
]
ca
m
e
u
p
with
SMOT
E
,
wh
ich
m
ak
es
f
ak
e
s
am
p
les
f
o
r
th
e
m
i
n
o
r
ity
c
lass
b
y
f
illi
n
g
in
th
e
g
ap
s
b
et
wee
n
r
ea
l
ex
am
p
les.
T
h
is
ap
p
r
o
ac
h
in
c
r
ea
s
es
th
e
r
ep
r
esen
tatio
n
o
f
th
e
m
in
o
r
ity
class
an
d
ad
d
r
ess
es
cla
s
s
im
b
alan
ce
.
Ho
wev
er
,
wh
ile
SMOT
E
ef
f
ec
tiv
ely
b
o
o
s
ts
m
in
o
r
ity
class
in
s
tan
ce
s
,
it
ca
n
also
in
t
r
o
d
u
ce
n
o
is
e
an
d
o
v
er
lap
b
etwe
en
class
es,
wh
ich
m
ay
af
f
ec
t c
lass
if
ier
p
er
f
o
r
m
an
ce
.
T
o
m
ek
l
in
k
s
,
p
r
o
p
o
s
ed
b
y
T
o
m
ek
[
1
1
]
,
h
elp
s
clea
n
th
e
d
at
aset
b
y
r
em
o
v
in
g
am
b
ig
u
o
u
s
in
s
tan
ce
s
n
ea
r
th
e
d
ec
is
io
n
b
o
u
n
d
a
r
y
.
W
h
en
co
m
b
i
n
ed
with
SMOT
E
,
th
e
SMOT
E
T
o
m
ek
tech
n
i
q
u
e
[
1
2
]
n
o
t
o
n
ly
au
g
m
en
ts
th
e
m
in
o
r
ity
class
b
u
t
also
r
ef
in
es
th
e
d
ec
is
io
n
b
o
u
n
d
ar
y
b
y
elim
in
atin
g
th
ese
a
m
b
ig
u
o
u
s
in
s
tan
ce
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
E
n
h
a
n
ci
n
g
p
h
is
h
in
g
w
eb
s
ite
d
etec
tio
n
:
a
co
m
p
a
r
a
tive
s
tu
d
y
o
f S
MOTETo
mek
-
X
GB
a
n
d
…
(
K
a
ma
l O
ma
r
i
)
2937
r
esu
ltin
g
in
clea
r
er
class
s
ep
ar
atio
n
.
On
th
e
o
th
er
h
an
d
,
SMOT
E
E
NN
[
1
3
]
co
m
b
in
es
SMOT
E
with
E
NN
[
5
]
,
a
tech
n
iq
u
e
th
at
r
ed
u
ce
s
n
o
is
e
b
y
r
em
o
v
in
g
in
s
tan
ce
s
m
is
class
if
ied
b
y
th
eir
n
ea
r
est
n
eig
h
b
o
r
s
[
1
4
]
.
T
h
is
h
y
b
r
id
ap
p
r
o
ac
h
b
alan
ce
s
th
e
d
ataset
wh
ile
also
clea
n
in
g
it,
wh
ich
ca
n
en
h
a
n
ce
th
e
p
er
f
o
r
m
an
ce
o
f
class
if
icatio
n
m
o
d
els.
B
o
th
SMOT
E
E
NN
an
d
SM
OT
E
T
o
m
ek
h
a
v
e
b
ee
n
wid
el
y
u
s
ed
to
f
ix
class
im
b
alan
ce
s
in
m
an
y
f
ield
s
,
s
h
o
win
g
th
at
th
ey
wo
r
k
well
to
im
p
r
o
v
e
m
o
d
el
p
e
r
f
o
r
m
a
n
ce
.
I
n
m
ed
icin
e
,
SMOT
E
E
NN
h
as
b
ee
n
wid
ely
u
s
ed
f
o
r
ea
r
l
y
d
etec
t
io
n
task
s
,
lik
e
p
r
ed
ictin
g
w
h
en
s
ep
tic
s
h
o
ck
will
s
tar
t,
d
iag
n
o
s
in
g
m
is
s
ed
ab
o
r
tio
n
,
a
n
d
p
r
ed
ictin
g
ch
r
o
n
ic
co
n
d
itio
n
s
lik
e
Par
k
in
s
o
n
'
s
d
is
ea
s
e
an
d
ch
r
o
n
ic
h
ea
r
t
f
ailu
r
e.
T
h
is
h
as
led
to
b
etter
d
iag
n
o
s
tic
ac
cu
r
ac
y
[
1
5
]
,
[
1
6
]
.
I
n
th
e
f
in
a
n
cial
s
ec
to
r
,
SMOT
E
E
NN
h
as
also
b
ee
n
s
u
cc
ess
f
u
lly
u
s
ed
to
p
r
ed
ict
f
r
a
u
d
with
b
etter
r
esu
lts
[
1
7
]
,
[
1
8
]
.
T
h
ese
s
tu
d
ie
s
to
g
eth
er
s
h
o
w
th
at
th
e
tech
n
iq
u
e
ca
n
b
alan
ce
d
atasets
an
d
m
ak
e
p
r
e
d
ictiv
e
m
o
d
els m
o
r
e
r
eliab
le.
I
n
ad
d
itio
n
,
SMOT
E
T
o
m
e
k
h
as
b
ee
n
ex
ten
s
iv
ely
im
p
lem
en
ted
in
s
ev
er
al
f
ield
s
,
wer
e
i
m
b
alan
ce
d
d
ata
p
o
s
es
ch
allen
g
es
to
m
o
d
e
l
ac
cu
r
ac
y
.
Fo
r
in
s
tan
ce
,
in
m
ed
icin
e,
SMOT
E
T
o
m
ek
h
as
b
ee
n
u
s
ed
to
d
ev
elo
p
p
r
ed
ictiv
e
m
o
d
els
f
o
r
illn
ess
e
s
lik
e
d
iab
etes
an
d
h
ig
h
b
lo
o
d
p
r
ess
u
r
e,
as
well
as
in
ca
n
c
er
s
tu
d
ies
to
p
r
ed
ict
th
e
d
ev
elo
p
m
en
t
o
f
p
r
o
s
tate
a
n
d
ce
r
v
ical
ca
n
ce
r
ca
s
es
b
ased
o
n
im
b
ala
n
ce
d
d
ata
s
ets
[
1
9
]
,
[
2
0
]
.
Similar
ly
,
in
co
m
p
u
ter
s
cien
ce
,
SMOT
E
T
o
m
ek
h
as
p
r
o
v
ed
to
b
e
u
s
ef
u
l
in
r
ec
o
m
m
en
d
er
s
y
s
tem
s
an
d
s
o
f
twar
e
b
u
g
p
r
ed
ictio
n
,
as
well
as
in
h
an
d
li
n
g
h
ig
h
ly
u
n
b
alan
ce
d
d
ata
s
ets
in
p
e
r
s
o
n
ality
r
ec
o
g
n
itio
n
[
1
2
]
.
T
h
ese
ex
a
m
p
les
illu
s
tr
ate
th
e
co
n
s
id
er
ab
le
a
d
v
an
tag
es
th
at
ca
n
b
e
o
b
tain
ed
f
r
o
m
u
s
in
g
th
e
SMOT
E
T
o
m
ek
alg
o
r
ith
m
,
w
h
ich
is
wh
y
th
is
tech
n
iq
u
e
m
ay
b
e
co
n
s
id
er
ed
u
s
ef
u
l
f
o
r
d
ea
lin
g
wi
th
th
e
p
r
o
b
lem
o
f
class
im
b
alan
ce
s
in
n
u
m
er
o
u
s
ca
s
es.
I
n
co
n
clu
s
io
n
,
alth
o
u
g
h
m
ac
h
in
e
lear
n
in
g
alg
o
r
ith
m
s
ca
n
p
r
o
v
id
e
m
o
r
e
s
o
p
h
is
ticated
m
eth
o
d
s
o
f
p
h
is
h
in
g
d
etec
tio
n
,
th
e
p
r
o
b
lem
o
f
class
im
b
alan
ce
s
r
eq
u
i
r
es sp
ec
ial
atten
tio
n
an
d
r
esam
p
lin
g
.
3.
M
E
T
H
O
D
T
h
is
s
ec
ti
o
n
p
r
es
e
n
t
s
th
e
r
e
s
ea
r
c
h
m
e
th
o
d
o
l
o
g
y
f
o
r
im
p
r
o
v
i
n
g
p
h
is
h
i
n
g
we
b
s
it
e
d
e
t
ec
ti
o
n
b
y
in
t
eg
r
ati
n
g
a
d
v
a
n
c
ed
r
esa
m
p
li
n
g
te
ch
n
i
q
u
es wit
h
t
h
e
XGB
o
o
s
t a
l
g
o
r
it
h
m
.
Sp
ec
i
f
i
ca
l
ly
,
w
e
d
es
cr
ib
e
t
h
e
d
atas
et
u
s
e
d
i
n
o
u
r
e
x
p
e
r
i
m
e
n
ts
an
d
d
etai
l
t
h
e
p
r
e
p
r
o
ce
s
s
in
g
s
te
p
s
a
p
p
lie
d
to
a
d
d
r
ess
c
lass
i
m
b
al
a
n
c
e
t
h
r
o
u
g
h
h
y
b
r
i
d
r
es
am
p
li
n
g
m
et
h
o
d
s
s
u
c
h
as
SMOT
E
T
o
m
e
k
a
n
d
SM
OT
E
E
NN
.
F
u
r
th
er
m
o
r
e,
we
o
u
tl
i
n
e
t
h
e
ex
p
er
im
en
tal
f
r
am
ew
o
r
k
d
esi
g
n
e
d
to
c
o
m
p
a
r
e
t
h
e
p
e
r
f
o
r
m
an
ce
o
f
SMO
T
E
T
o
m
e
k
-
X
GB
a
n
d
SMO
T
E
E
NN
-
X
GB
,
i
n
cl
u
d
i
n
g
ev
al
u
at
io
n
m
et
r
ics
a
n
d
v
ali
d
at
i
o
n
p
r
o
ce
d
u
r
es
to
e
n
s
u
r
e
a
r
o
b
u
s
t a
n
d
r
e
lia
b
l
e
a
n
al
y
s
is
.
3
.
1
.
Da
t
a
s
et
des
cr
iptio
n
T
h
is
s
tu
d
y
u
s
ed
a
p
u
b
licly
a
v
ailab
le
b
en
ch
m
ar
k
p
h
is
h
in
g
web
s
ite
d
ataset
f
r
o
m
th
e
UC
I
Ma
ch
in
e
L
ea
r
n
in
g
R
ep
o
s
ito
r
y
[
2
1
]
.
T
h
e
d
ataset
co
m
p
r
is
es 1
1
,
0
5
5
s
am
p
les with
3
1
attr
ib
u
tes,
in
clu
d
i
n
g
3
0
f
ea
tu
r
es a
n
d
o
n
e
class
lab
el.
T
h
e
class
im
b
alan
ce
r
atio
is
ca
lcu
lated
u
s
in
g
(
1
)
.
=
(
1
)
T
h
e
d
ataset
ca
teg
o
r
ized
in
t
o
t
wo
class
es:
“
p
h
is
h
in
g
web
s
ite”
an
d
“
leg
itima
te
web
s
ite.
”
Sp
ec
if
ically
,
th
er
e
ar
e
6
,
1
5
7
in
s
tan
ce
s
o
f
p
h
is
h
in
g
web
s
ites
an
d
4
,
8
9
8
i
n
s
tan
ce
s
o
f
leg
itima
te
web
s
ites
.
Sp
ec
if
ically
,
th
er
e
ar
e
6
,
1
5
7
in
s
tan
ce
s
o
f
p
h
is
h
in
g
web
s
ites
an
d
4
,
8
9
8
in
s
tan
ce
s
o
f
leg
itima
te
web
s
ites
.
T
h
e
im
b
alan
ce
r
atio
,
ca
lcu
lated
as th
e
r
atio
o
f
th
e
n
u
m
b
er
o
f
leg
itima
te
web
s
ites
to
p
h
is
h
in
g
web
s
ites
,
is
1
.
2
5
7
0
.
T
h
e
s
el
ec
t
ed
d
a
tase
t
in
tr
o
d
u
ce
d
s
e
v
e
r
al
i
n
n
o
v
ati
v
e
f
e
at
u
r
es
,
in
c
lu
d
i
n
g
t
h
e
e
x
p
e
r
i
m
e
n
t
al
i
n
t
r
o
d
u
c
ti
o
n
o
f
n
o
v
e
l
r
u
les
f
o
r
s
p
ec
if
ic
wel
l
-
es
ta
b
lis
h
e
d
p
ar
am
et
er
s
[
2
1
]
.
A
t
o
ta
l
o
f
3
0
p
ar
am
ete
r
s
we
r
e
c
o
n
s
i
d
er
e
d
i
n
t
h
is
an
al
y
s
is
,
w
h
i
c
h
ar
e
lis
t
e
d
in
T
ab
le
1
.
T
h
e
d
esc
r
i
p
ti
v
e
s
t
atis
ti
cs
s
u
m
m
a
r
iz
e
d
ab
o
v
e
s
u
ch
as
co
u
n
t
,
m
e
a
n
,
s
tan
d
ar
d
d
ev
iati
o
n
,
m
in
im
u
m
,
a
n
d
m
ax
im
u
m
o
f
f
e
r
cr
u
cia
l
i
n
s
i
g
h
ts
i
n
t
o
th
e
ce
n
t
r
al
te
n
d
e
n
cy
,
d
is
p
e
r
s
i
o
n
,
a
n
d
o
v
e
r
al
l d
is
tr
ib
u
t
io
n
o
f
ea
c
h
a
tt
r
ib
u
te
wit
h
i
n
t
h
e
d
at
aset
.
T
h
is
a
n
al
y
s
is
p
r
o
v
i
d
es a
n
in
-
d
e
p
t
h
u
n
d
e
r
s
ta
n
d
i
n
g
o
f
t
h
e
d
at
ase
t'
s
c
h
a
r
a
cte
r
is
t
ics
,
wh
ic
h
is
ess
e
n
t
ial
f
o
r
d
ev
el
o
p
i
n
g
r
o
b
u
s
t
m
o
d
els
a
n
d
d
r
aw
i
n
g
r
eli
a
b
l
e
c
o
n
c
lu
s
io
n
s
.
3
.
2
.
Adv
a
nced
da
t
a
-
ba
la
ncing
t
ec
hn
iqu
e
s
f
o
r
ph
is
hi
ng
websi
t
e
det
ec
t
io
n
T
h
is
s
tu
d
y
ex
p
lo
r
ed
two
ad
v
a
n
ce
d
d
ata
-
b
alan
cin
g
tech
n
i
q
u
e
s
to
ad
d
r
ess
class
im
b
alan
ce
i
n
p
h
is
h
in
g
web
s
ite
d
etec
tio
n
.
T
h
e
tec
h
n
iq
u
es
u
s
ed
ar
e
i)
SMO
T
E
E
NN
h
y
b
r
id
s
am
p
lin
g
m
eth
o
d
[
1
3
]
a
n
d
ii)
SMOT
E
T
o
m
ek
h
y
b
r
id
s
am
p
lin
g
m
eth
o
d
[
1
2
]
.
T
h
ese
tech
n
iq
u
es
ar
e
d
esig
n
e
d
to
e
n
h
an
c
e
th
e
p
er
f
o
r
m
a
n
ce
o
f
m
ac
h
in
e
lear
n
i
n
g
m
o
d
el
s
b
y
cr
ea
tin
g
a
m
o
r
e
b
ala
n
ce
d
d
ataset,
wh
ich
is
cr
u
cial
f
o
r
im
p
r
o
v
in
g
class
if
icatio
n
ac
cu
r
ac
y
.
3
.
2
.
1
.
SM
O
T
E
E
NN
hy
brid s
a
m
pli
ng
SMOT
E
E
NN
h
y
b
r
id
s
am
p
lin
g
[
1
3
]
is
a
tech
n
iq
u
e
th
at
co
m
b
in
es
two
m
eth
o
d
s
SMOT
E
[
1
0
]
an
d
E
NN
[
5
]
u
n
d
er
-
s
am
p
lin
g
to
t
ac
k
le
class
im
b
alan
ce
e
f
f
ec
tiv
ely
.
SMOT
E
s
y
n
t
h
esizes
m
in
o
r
ity
e
x
am
p
les
b
y
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.
1
5
,
No
.
3
,
J
u
n
e
2
0
2
6
:
2
9
3
5
-
2
9
4
5
2938
ch
o
o
s
in
g
o
n
e
m
i
n
o
r
ity
ex
am
p
le
an
d
cr
ea
tin
g
m
o
r
e
ex
am
p
le
s
o
n
th
e
lin
e
jo
in
in
g
it
to
its
k
-
n
ea
r
est
m
in
o
r
ity
ex
am
p
les.
Alth
o
u
g
h
th
is
tech
n
iq
u
e
m
ak
es
s
u
r
e
th
er
e
is
a
b
alan
ce
in
th
e
d
ataset
b
y
ad
d
i
n
g
ex
am
p
les
o
f
th
e
m
in
o
r
ity
class
,
SMOT
E
m
ig
h
t
co
n
s
id
er
th
at
an
y
p
o
in
t
jo
in
in
g
an
y
two
ex
am
p
les
o
f
th
e
m
in
o
r
ity
class
is
al
s
o
an
ex
am
p
le
o
f
th
e
m
in
o
r
ity
class
.
T
o
ad
d
r
ess
th
is
,
E
NN
s
tep
s
i
n
to
clea
n
th
e
d
ataset.
I
t
ev
al
u
ates
ea
ch
s
am
p
le
b
ased
o
n
i
ts
n
ea
r
est
n
eig
h
b
o
r
s
an
d
r
em
o
v
es
s
am
p
les
th
at
ar
e
o
u
t
o
f
p
lace
,
m
e
an
in
g
th
o
s
e
wh
o
s
e
class
lab
els
d
o
n
'
t
m
atch
th
e
m
ajo
r
ity
class
lab
els
o
f
th
eir
n
eig
h
b
o
r
s
.
T
h
is
p
r
o
ce
s
s
h
elp
s
t
o
elim
in
ate
n
o
is
e,
esp
ec
ially
wh
en
n
ew
s
y
n
th
etic
s
am
p
les
f
r
o
m
SMOT
E
m
ig
h
t
o
v
er
lap
with
th
e
m
aj
o
r
ity
clas
s
s
p
ac
e
as
s
h
o
wn
in
Fig
u
r
e
1
.
B
y
f
ir
s
t
ap
p
ly
i
n
g
SMOT
E
to
b
alan
ce
th
e
d
ataset
an
d
th
en
u
s
in
g
E
NN
to
cle
an
it
u
p
,
SMOT
E
E
NN
en
s
u
r
e
s
th
at
th
e
d
ataset
is
well
-
b
alan
ce
d
an
d
m
in
im
izes th
e
r
is
k
o
f
i
n
tr
o
d
u
cin
g
n
o
is
e.
T
ab
le
1
.
Ph
is
h
in
g
web
s
ite
d
ata
attr
ib
u
tes
C
a
t
e
g
o
r
y
D
a
t
a
a
t
t
r
i
b
u
t
e
s
M
e
a
n
S
t
d
C
o
u
n
t
M
i
n
2
5
%
5
0
%
7
5
%
M
a
x
U
R
L
-
b
a
se
d
U
R
L_
Le
n
g
t
h
-
0
.
6
3
3
1
9
8
0
.
7
6
6
0
9
5
1
1
0
5
5
-
1
.
0
-
1
.
0
-
1
.
0
-
1
.
0
1
.
0
U
R
L
-
b
a
se
d
h
a
v
i
n
g
_
I
P
_
A
d
d
r
e
ss
0
.
3
1
3
7
9
5
0
.
9
4
9
5
3
4
1
1
0
5
5
-
1
.
0
-
1
.
0
1
.
0
1
.
0
1
.
0
U
R
L
-
b
a
se
d
S
h
o
r
t
e
n
i
n
g
_
S
e
r
v
i
c
e
0
.
7
3
8
7
6
1
0
.
6
7
3
9
9
8
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
U
R
L
-
b
a
se
d
h
a
v
i
n
g
_
A
t
_
S
y
m
b
o
l
0
.
7
0
0
5
8
8
0
.
7
1
3
5
9
8
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
U
R
L
-
b
a
se
d
d
o
u
b
l
e
_
s
l
a
s
h
_
r
e
d
i
r
e
c
t
i
n
g
0
.
7
4
1
4
7
4
0
.
6
7
1
0
1
1
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
U
R
L
-
b
a
se
d
P
r
e
f
i
x
_
S
u
f
f
i
x
-
0
.
7
3
4
9
6
2
0
.
6
7
8
1
3
9
1
1
0
5
5
-
1
.
0
-
1
.
0
-
1
.
0
-
1
.
0
1
.
0
D
o
ma
i
n
-
b
a
s
e
d
D
o
ma
i
n
_
r
e
g
i
s
t
r
a
t
i
o
n
_
l
e
n
g
t
h
-
0
.
3
3
6
7
7
1
0
.
9
4
1
6
2
9
1
1
0
5
5
-
1
.
0
-
1
.
0
-
1
.
0
1
.
0
1
.
0
D
o
ma
i
n
-
b
a
s
e
d
a
g
e
_
o
f
_
d
o
ma
i
n
0
.
0
6
1
2
3
9
0
.
9
9
8
1
6
8
1
1
0
5
5
-
1
.
0
-
1
.
0
1
.
0
1
.
0
1
.
0
D
o
ma
i
n
-
b
a
s
e
d
D
N
S
R
e
c
o
r
d
0
.
2
8
7
2
9
1
0
.
8
2
7
7
3
3
1
1
0
5
5
-
1
.
0
-
1
.
0
1
.
0
1
.
0
1
.
0
D
o
ma
i
n
-
b
a
s
e
d
P
a
g
e
_
R
a
n
k
-
0
.
4
8
3
6
7
3
0
.
8
7
5
2
8
9
1
1
0
5
5
-
1
.
0
-
1
.
0
-
1
.
0
1
.
0
1
.
0
D
o
ma
i
n
-
b
a
s
e
d
G
o
o
g
l
e
_
I
n
d
e
x
0
.
7
2
1
5
7
4
0
.
6
9
2
3
6
9
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
C
o
n
t
e
n
t
-
b
a
se
d
S
S
Lf
i
n
a
l
_
S
t
a
t
e
0
.
2
5
0
9
2
7
0
.
9
1
1
8
9
2
1
1
0
5
5
-
1
.
0
-
1
.
0
1
.
0
1
.
0
1
.
0
C
o
n
t
e
n
t
-
b
a
se
d
F
a
v
i
c
o
n
0
.
6
2
8
5
8
4
0
.
7
7
7
7
7
7
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
C
o
n
t
e
n
t
-
b
a
se
d
R
e
q
u
e
st
_
U
R
L
0
.
1
8
6
7
9
3
0
.
9
8
2
4
4
4
1
1
0
5
5
-
1
.
0
-
1
.
0
1
.
0
1
.
0
1
.
0
C
o
n
t
e
n
t
-
b
a
se
d
U
R
L_
o
f
_
A
n
c
h
o
r
-
0
.
0
7
6
5
2
6
0
.
7
1
5
1
3
8
1
1
0
5
5
-
1
.
0
-
1
.
0
0
.
0
0
.
0
1
.
0
C
o
n
t
e
n
t
-
b
a
se
d
Li
n
k
s
_
i
n
_
t
a
g
s
-
0
.
1
1
8
1
3
7
0
.
7
6
3
9
7
3
1
1
0
5
5
-
1
.
0
-
1
.
0
0
.
0
0
.
0
1
.
0
C
o
n
t
e
n
t
-
b
a
se
d
SFH
-
0
.
5
9
5
7
4
9
0
.
7
5
9
1
4
3
1
1
0
5
5
-
1
.
0
-
1
.
0
-
1
.
0
-
1
.
0
1
.
0
C
o
n
t
e
n
t
-
b
a
se
d
S
u
b
m
i
t
t
i
n
g
_
t
o
_
e
ma
i
l
0
.
6
3
5
6
4
0
0
.
7
7
2
0
2
1
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
C
o
n
t
e
n
t
-
b
a
se
d
A
b
n
o
r
ma
l
_
U
R
L
0
.
7
0
5
2
9
2
0
.
7
0
8
9
4
9
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
B
e
h
a
v
i
o
r
a
l
R
e
d
i
r
e
c
t
0
.
1
1
5
6
9
4
0
.
3
1
9
8
7
2
1
1
0
5
5
0
.
0
0
.
0
0
.
0
0
.
0
1
.
0
B
e
h
a
v
i
o
r
a
l
o
n
_
m
o
u
se
o
v
e
r
0
.
7
6
2
0
9
9
0
.
6
4
7
4
9
0
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
B
e
h
a
v
i
o
r
a
l
R
i
g
h
t
C
l
i
c
k
0
.
9
1
3
8
8
5
0
.
4
0
5
9
9
1
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
B
e
h
a
v
i
o
r
a
l
p
o
p
U
p
W
i
n
d
o
w
0
.
6
1
3
3
8
8
0
.
7
8
9
8
1
8
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
B
e
h
a
v
i
o
r
a
l
I
f
r
a
me
0
.
8
1
6
9
1
5
0
.
5
7
6
7
8
4
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
Tr
a
f
f
i
c
&
p
o
p
u
l
a
r
i
t
y
w
e
b
_
t
r
a
f
f
i
c
0
.
2
8
7
2
9
1
0
.
8
2
7
7
3
3
1
1
0
5
5
-
1
.
0
0
.
0
1
.
0
1
.
0
1
.
0
Tr
a
f
f
i
c
&
p
o
p
u
l
a
r
i
t
y
Li
n
k
s
_
p
o
i
n
t
i
n
g
_
t
o
_
p
a
g
e
0
.
3
4
4
0
0
7
0
.
5
6
9
9
4
4
1
1
0
5
5
-
1
.
0
0
.
0
0
.
0
1
.
0
1
.
0
Tr
a
f
f
i
c
&
p
o
p
u
l
a
r
i
t
y
S
t
a
t
i
st
i
c
a
l
_
r
e
p
o
r
t
0
.
7
1
9
5
8
4
0
.
6
9
4
4
3
7
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
Tr
a
f
f
i
c
&
p
o
p
u
l
a
r
i
t
y
H
TTPS
_
t
o
k
e
n
0
.
6
7
5
0
7
9
0
.
7
3
7
7
7
9
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
Tr
a
f
f
i
c
&
p
o
p
u
l
a
r
i
t
y
p
o
r
t
0
.
7
2
8
2
6
8
0
.
6
8
5
3
2
4
1
1
0
5
5
-
1
.
0
1
.
0
1
.
0
1
.
0
1
.
0
Fig
u
r
e
1
.
Sch
em
atic
d
iag
r
am
o
f
th
e
SMOT
E
E
NN
m
eth
o
d
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
E
n
h
a
n
ci
n
g
p
h
is
h
in
g
w
eb
s
ite
d
etec
tio
n
:
a
co
m
p
a
r
a
tive
s
tu
d
y
o
f S
MOTETo
mek
-
X
GB
a
n
d
…
(
K
a
ma
l O
ma
r
i
)
2939
T
ab
le
2
o
u
tlin
es
th
e
p
ar
a
m
eter
s
u
s
ed
in
th
e
SMOT
E
E
NN
h
y
b
r
id
s
am
p
lin
g
m
e
th
o
d
.
T
h
e
"sam
p
lin
g
_
s
tr
ateg
y
"
is
s
et
t
o
"a
u
to
",
en
s
u
r
in
g
th
at
th
e
m
i
n
o
r
ity
class
is
ad
e
q
u
ately
r
esam
p
led
to
ac
h
ie
v
e
d
ataset
b
alan
ce
.
A
"r
an
d
o
m
_
s
tate"
o
f
4
2
is
u
s
ed
to
m
ain
tain
r
ep
r
o
d
u
cib
ilit
y
b
y
c
o
n
tr
o
llin
g
th
e
r
an
d
o
m
n
u
m
b
er
g
en
er
atio
n
p
r
o
ce
s
s
.
T
h
e
"sm
o
te"
p
ar
am
eter
co
r
r
e
s
p
o
n
d
s
to
th
e
SMOT
E
o
b
ject,
co
n
f
ig
u
r
ed
with
5
n
ea
r
est
n
eig
h
b
o
r
s
"k
_
n
eig
h
b
o
r
s
=5
"
to
g
u
id
e
t
h
e
cr
ea
tio
n
o
f
s
y
n
th
etic
s
am
p
les
f
o
r
th
e
m
in
o
r
ity
class
.
T
h
e
"e
n
n
"
p
ar
am
ete
r
r
elate
s
to
th
e
E
NN
m
eth
o
d
,
wh
ich
is
s
et
to
ev
alu
ate
1
n
ea
r
est n
eig
h
b
o
r
"
n
_
n
eig
h
b
o
r
s
=1
"
an
d
ap
p
ly
th
e
'
m
o
d
e'
s
elec
tio
n
s
tr
a
teg
y
"k
in
d
_
s
el
=
"m
o
d
e"
to
ef
f
icien
tly
elim
in
ate
n
o
is
y
in
s
tan
ce
s
.
T
h
e
"n
_
jo
b
s
"
p
ar
am
eter
is
s
p
ec
if
ied
as
"
No
n
e
",
in
d
icatin
g
th
e
u
s
e
o
f
a
s
in
g
le
C
PU
co
r
e,
th
o
u
g
h
t
h
is
ca
n
b
e
ad
ju
s
ted
b
ased
o
n
th
e
c
o
m
p
u
tatio
n
al
r
eso
u
r
ce
s
av
ailab
le
(
T
ab
le
2
)
.
T
ab
le
2
.
SMOT
E
E
NN
p
ar
am
e
ter
s
P
a
r
a
me
t
e
r
V
a
l
u
e
D
e
scri
p
t
i
o
n
samp
l
i
n
g
_
s
t
r
a
t
e
g
y
'au
t
o
'
R
e
sa
mp
l
e
t
h
e
m
i
n
o
r
i
t
y
c
l
a
ss
t
o
b
a
l
a
n
c
e
t
h
e
d
a
t
a
se
t
.
r
a
n
d
o
m
_
s
t
a
t
e
42
S
e
e
d
u
s
e
d
b
y
t
h
e
r
a
n
d
o
m
n
u
mb
e
r
g
e
n
e
r
a
t
o
r
f
o
r
r
e
p
r
o
d
u
c
i
b
i
l
i
t
y
.
smo
t
e
S
M
O
TE
o
b
j
e
c
t
(
k
_
n
e
i
g
h
b
o
r
s
=
5
)
S
M
O
TE
o
b
j
e
c
t
w
i
t
h
sp
e
c
i
f
i
e
d
p
a
r
a
me
t
e
r
s.
e
n
n
EN
N
o
b
j
e
c
t
(
n
_
n
e
i
g
h
b
o
r
s
=
1
,
k
i
n
d
_
s
e
l
=
'm
o
d
e
')
EN
N
o
b
j
e
c
t
w
i
t
h
sp
e
c
i
f
i
e
d
p
a
r
a
me
t
e
r
s
.
n
_
j
o
b
s
N
o
n
e
N
u
mb
e
r
o
f
C
P
U
c
o
r
e
s
u
se
d
d
u
r
i
n
g
p
r
o
c
e
ssi
n
g
(
N
o
n
e
me
a
n
s
u
s
i
n
g
o
n
e
c
o
r
e
)
.
3
.
2
.
2
.
SM
O
T
E
T
o
m
ek
hy
brid s
a
m
pli
ng
SMOT
E
T
o
m
ek
h
y
b
r
i
d
s
am
p
li
n
g
is
a
tech
n
iq
u
e
th
at
c
o
m
b
i
n
es
SMOT
E
an
d
T
o
m
ek
l
in
k
s
to
tack
le
class
im
b
alan
ce
an
d
b
o
o
s
t
class
if
icatio
n
p
er
f
o
r
m
an
ce
.
Her
e’
s
h
o
w
it
wo
r
k
s
T
o
m
e
k
l
in
k
s
is
an
u
n
d
er
-
s
am
p
lin
g
m
eth
o
d
th
at
f
i
n
d
s
an
d
r
em
o
v
e
s
p
air
s
o
f
s
am
p
les
th
at
ar
e
cl
o
s
e
to
ea
c
h
o
t
h
er
b
u
t
b
elo
n
g
t
o
d
if
f
er
en
t
class
es.
T
h
is
h
elp
s
clea
n
u
p
t
h
e
ar
ea
b
etwe
en
class
es,
r
ed
u
cin
g
o
v
e
r
lap
an
d
n
o
is
e.
SMOT
E
,
o
n
th
e
o
th
e
r
h
an
d
,
h
elp
s
b
y
cr
ea
tin
g
ar
tific
ial
d
ata
p
o
in
ts
f
o
r
th
e
m
in
o
r
ity
class
.
T
h
e
tech
n
iq
u
e
in
v
o
lv
es
ch
o
o
s
in
g
a
m
in
o
r
ity
class
p
o
in
t
a
n
d
f
o
r
m
in
g
n
ew
p
o
in
ts
alo
n
g
th
e
li
n
e
s
eg
m
en
t
jo
in
in
g
th
at
p
o
in
t
to
its
n
eig
h
b
o
r
in
g
p
o
in
ts
.
T
h
is
h
elp
s
cr
ea
te
a
b
alan
ce
d
d
ata
s
et
an
d
en
s
u
r
es
p
r
o
p
er
d
is
tr
ib
u
tio
n
,
wh
ich
is
ex
tr
em
ely
im
p
o
r
tan
t i
n
b
o
o
s
tin
g
m
ac
h
in
e
lea
r
n
in
g
m
o
d
el
p
er
f
o
r
m
a
n
ce
.
Af
ter
ap
p
ly
in
g
SMOT
E
to
in
cr
ea
s
e
th
e
n
u
m
b
er
o
f
m
in
o
r
ity
class
s
am
p
les,
T
o
m
ek
l
i
n
k
s
f
u
r
th
er
r
ef
in
es
th
e
d
ataset
b
y
r
em
o
v
in
g
th
e
p
r
o
b
lem
atic
p
air
s
i
d
en
tifie
d
ea
r
lier
.
T
h
is
s
tep
h
elp
s
clea
r
u
p
an
y
r
em
ain
in
g
n
o
is
e
an
d
s
h
a
r
p
en
s
th
e
d
ec
is
io
n
b
o
u
n
d
ar
y
b
etwe
en
class
es
as
s
h
o
wn
in
Fig
u
r
e
2
.
B
y
co
m
b
in
in
g
th
ese
tech
n
iq
u
es,
SMOT
E
T
o
m
ek
h
y
b
r
id
s
am
p
lin
g
ef
f
ec
tiv
ely
ad
d
r
ess
es
clas
s
im
b
alan
ce
,
r
ed
u
ce
s
n
o
is
e,
an
d
r
ef
in
es
d
ec
is
io
n
b
o
u
n
d
ar
ies,
s
ig
n
if
ican
tly
en
h
an
cin
g
th
e
p
e
r
f
o
r
m
a
n
ce
o
f
m
ac
h
in
e
lear
n
in
g
m
o
d
els,
s
u
c
h
as
th
o
s
e
u
s
ed
f
o
r
d
etec
tin
g
p
h
is
h
i
n
g
web
s
ites
.
Fig
u
r
e
2
.
Sch
em
atic
d
iag
r
am
o
f
th
e
SMOT
E
-
T
o
m
e
k
m
eth
o
d
T
h
e
k
ey
d
if
f
er
en
ce
b
etwe
en
T
ab
les
2
an
d
3
lies
in
th
e
s
u
b
s
titu
tio
n
o
f
th
e
E
NN
m
eth
o
d
with
th
e
T
o
m
ek
l
in
k
s
m
eth
o
d
.
W
h
ile
t
h
e
co
r
e
p
ar
a
m
eter
s
s
u
ch
as
"s
am
p
lin
g
_
s
tr
ateg
y
"
,
"r
an
d
o
m
_
s
tate"
,
an
d
"sm
o
te"
r
em
ain
co
n
s
is
ten
t
ac
r
o
s
s
b
o
th
ap
p
r
o
ac
h
es,
th
e
in
tr
o
d
u
ctio
n
o
f
th
e
"to
m
ek
"
p
ar
am
eter
i
n
T
a
b
le
3
m
ar
k
s
a
k
ey
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.
1
5
,
No
.
3
,
J
u
n
e
2
0
2
6
:
2
9
3
5
-
2
9
4
5
2940
ad
ju
s
tm
en
t.
T
h
is
p
ar
am
eter
lev
er
ag
es
T
o
m
ek
l
in
k
s
to
id
en
ti
f
y
an
d
elim
in
ate
p
air
s
o
f
s
am
p
les
th
at
ar
e
n
ea
r
est
n
eig
h
b
o
r
s
b
u
t
b
elo
n
g
to
d
if
f
er
en
t
class
es.
T
h
i
s
r
ef
in
em
en
t
p
r
o
ce
s
s
is
cr
u
cial,
as
it
ef
f
ec
tiv
ely
s
h
ar
p
en
s
class
b
o
u
n
d
ar
ies,
th
e
r
eb
y
r
ed
u
cin
g
o
v
er
lap
an
d
en
h
an
cin
g
th
e
o
v
er
all
p
r
ec
is
io
n
o
f
th
e
class
if
ie
r
.
B
y
in
c
o
r
p
o
r
atin
g
T
o
m
ek
l
in
k
s
,
th
e
SMOT
E
T
o
m
ek
m
eth
o
d
o
f
f
er
s
a
m
o
r
e
r
e
f
in
ed
ap
p
r
o
ac
h
to
ad
d
r
ess
in
g
class
im
b
alan
ce
.
T
h
e
h
y
p
er
p
ar
am
eter
s
u
s
ed
f
o
r
c
o
n
f
ig
u
r
in
g
th
e
XGBo
o
s
t
clas
s
if
i
er
with
in
th
e
SMOT
E
T
o
m
ek
an
d
SMOT
E
E
NN
f
r
am
ewo
r
k
s
ar
e
o
u
tlin
e
d
.
T
h
ese
s
ett
in
g
s
wer
e
ca
r
ef
u
lly
ch
o
s
en
to
b
alan
ce
m
o
d
el
co
m
p
lex
ity
,
tr
ain
in
g
ef
f
icien
cy
,
a
n
d
p
r
ed
ictiv
e
ac
cu
r
ac
y
.
T
ab
le
3
.
SMOT
E
T
o
m
e
k
p
ar
a
m
eter
s
P
a
r
a
me
t
e
r
V
a
l
u
e
D
e
scri
p
t
i
o
n
samp
l
i
n
g
_
s
t
r
a
t
e
g
y
'au
t
o
'
R
e
sa
mp
l
e
t
h
e
m
i
n
o
r
i
t
y
c
l
a
ss
t
o
b
a
l
a
n
c
e
t
h
e
d
a
t
a
se
t
.
r
a
n
d
o
m
_
s
t
a
t
e
42
S
e
e
d
u
s
e
d
b
y
t
h
e
r
a
n
d
o
m
n
u
mb
e
r
g
e
n
e
r
a
t
o
r
f
o
r
r
e
p
r
o
d
u
c
i
b
i
l
i
t
y
.
smo
t
e
S
M
O
TE
o
b
j
e
c
t
(
k
_
n
e
i
g
h
b
o
r
s
=
5
)
S
M
O
TE
o
b
j
e
c
t
w
i
t
h
sp
e
c
i
f
i
e
d
p
a
r
a
me
t
e
r
s.
t
o
m
e
k
To
me
k
l
i
n
k
s
o
b
j
e
c
t
To
me
k
Li
n
k
s me
t
h
o
d
t
o
i
d
e
n
t
i
f
y
a
n
d
r
e
mo
v
e
sam
p
l
e
p
a
i
r
s t
h
a
t
a
r
e
n
e
a
r
e
st
n
e
i
g
h
b
o
r
s
b
u
t
b
e
l
o
n
g
t
o
d
i
f
f
e
r
e
n
t
c
l
a
sse
s.
n
_
j
o
b
s
N
o
n
e
N
u
mb
e
r
o
f
C
P
U
c
o
r
e
s
u
se
d
d
u
r
i
n
g
p
r
o
c
e
ssi
n
g
(
N
o
n
e
mea
n
s
u
si
n
g
o
n
e
c
o
r
e
)
.
3
.
2
.
3
.
XG
B
o
o
s
t
cla
s
s
if
ier
T
h
e
XGBo
o
s
t
class
if
ier
was
ch
o
s
en
f
o
r
th
is
s
tu
d
y
d
u
e
to
its
e
x
ce
p
tio
n
al
p
er
f
o
r
m
an
ce
an
d
ap
p
r
o
p
r
iaten
ess
f
o
r
th
e
task
.
XGBo
o
s
t
i
s
p
ar
ticu
lar
ly
ef
f
e
ctiv
e
in
m
an
ag
in
g
co
m
p
lex
a
n
d
n
o
n
-
lin
ea
r
d
ata
th
r
o
u
g
h
its
g
r
ad
ien
t
b
o
o
s
tin
g
f
r
am
ewo
r
k
,
wh
ich
b
u
ild
s
an
en
s
em
b
le
o
f
tr
ee
s
to
en
h
an
ce
m
o
d
el
ac
cu
r
ac
y
[
2
2
]
.
T
h
is
ch
ar
ac
ter
is
tic
is
esp
ec
ially
b
en
ef
icial
f
o
r
p
h
is
h
in
g
web
s
ite
d
etec
tio
n
,
wh
e
r
e
th
e
ch
allen
g
e
o
f
d
is
tin
g
u
is
h
in
g
b
etwe
en
th
e
m
in
o
r
ity
class
(
p
h
is
h
in
g
s
ites
)
an
d
th
e
m
ajo
r
ity
class
(
leg
itima
te
s
ites
)
i
s
ex
ac
er
b
ated
b
y
class
im
b
alan
c
e
[
2
3
]
.
XGBo
o
s
t'
s
b
u
ilt
-
in
class
weig
h
t
ad
ju
s
tm
en
t
m
ec
h
an
is
m
s
ar
e
d
esig
n
e
d
to
ad
d
r
ess
th
ese
im
b
alan
ce
s
,
th
er
eb
y
im
p
r
o
v
in
g
th
e
class
if
ier
’
s
ab
ilit
y
to
id
en
tif
y
p
h
is
h
in
g
s
ites
ac
cu
r
ately
.
Fu
r
th
er
m
o
r
e
,
XGBo
o
s
t
p
r
o
v
id
es
ex
ten
s
iv
e
h
y
p
er
p
a
r
am
ete
r
tu
n
in
g
o
p
tio
n
s
,
s
u
ch
as
ad
ju
s
tm
en
ts
to
th
e
lear
n
in
g
r
ate
an
d
tr
ee
d
ep
t
h
,
wh
ich
f
ac
ilit
ate
o
p
tim
izatio
n
f
o
r
s
p
ec
if
ic
d
atasets
[
2
4
]
.
I
ts
ef
f
icien
cy
is
also
en
h
an
ce
d
b
y
p
ar
alleliza
tio
n
a
n
d
r
eg
u
lar
izatio
n
tech
n
iq
u
es,
wh
ich
co
n
tr
ib
u
te
to
f
aster
p
r
o
ce
s
s
in
g
an
d
g
r
ea
ter
m
o
d
el
r
o
b
u
s
tn
ess
[
2
5
]
.
T
h
e
s
e
f
ea
tu
r
es,
to
g
eth
er
with
r
o
b
u
s
t
co
m
m
u
n
ity
s
u
p
p
o
r
t
a
n
d
co
m
p
r
eh
en
s
iv
e
d
o
cu
m
e
n
tatio
n
,
m
a
k
e
XGBo
o
s
t a
h
ig
h
ly
s
u
itab
le
ch
o
ice
f
o
r
en
h
an
cin
g
p
h
is
h
in
g
web
s
ite
d
etec
tio
n
.
C
ar
ef
u
l
tu
n
in
g
o
f
th
ese
h
y
p
er
p
ar
am
eter
s
is
ess
en
tial
f
o
r
m
a
x
im
izin
g
t
h
e
class
if
ier
’
s
p
er
f
o
r
m
an
ce
i
n
b
o
th
f
r
am
ewo
r
k
s
,
e
n
s
u
r
in
g
th
at
th
e
d
ataset
r
em
ain
s
well
-
b
a
lan
ce
d
an
d
o
f
h
ig
h
q
u
ality
as
s
h
o
wn
in
T
ab
le
4
.
B
y
ca
r
ef
u
lly
tu
n
in
g
th
ese
h
y
p
er
p
a
r
am
eter
s
,
we
m
ak
e
th
e
m
o
s
t
o
f
th
e
SMOT
E
T
o
m
ek
m
eth
o
d
f
o
r
d
ata
r
esam
p
lin
g
a
n
d
th
e
p
o
wer
f
u
l
ca
p
ab
ilit
ies
o
f
th
e
XGBo
o
s
t
a
lg
o
r
ith
m
to
ac
c
u
r
ately
d
etec
t
p
h
is
h
in
g
d
o
m
ai
n
s
,
ev
en
wh
en
d
ea
lin
g
with
s
ig
n
if
ican
t c
lass
im
b
alan
ce
s
in
th
e
d
ataset.
I
n
s
u
m
m
ar
y
,
th
e
s
elec
ted
h
y
p
er
p
ar
am
eter
s
f
o
r
b
o
th
SMOT
E
T
o
m
ek
-
XG
B
an
d
SMOT
E
E
NN
-
XGB
ar
e
d
esig
n
ed
to
b
alan
ce
m
o
d
el
co
m
p
lex
ity
,
tr
ain
in
g
tim
e,
an
d
p
e
r
f
o
r
m
an
ce
.
T
h
e
g
o
al
is
to
d
ev
elo
p
r
o
b
u
s
t
an
d
ef
f
ec
tiv
e
class
if
ier
s
ca
p
ab
l
e
o
f
ad
d
r
ess
in
g
th
e
ch
allen
g
es p
o
s
ed
b
y
im
b
ala
n
c
ed
d
atasets
.
T
ab
le
4
.
XGBo
o
s
t c
lass
if
ier
p
ar
am
eter
s
P
a
r
a
me
t
e
r
V
a
l
u
e
D
e
scri
p
t
i
o
n
n
_
e
st
i
ma
t
o
r
s
1
0
0
Th
e
n
u
m
b
e
r
o
f
b
o
o
s
t
i
n
g
r
o
u
n
d
s
.
T
h
i
s
p
a
r
a
m
e
t
e
r
c
o
n
t
r
o
l
s
t
h
e
n
u
mb
e
r
o
f
t
r
e
e
s i
n
t
h
e
e
n
sem
b
l
e
.
max
_
d
e
p
t
h
7
Th
e
m
a
x
i
m
u
m
d
e
p
t
h
o
f
e
a
c
h
t
r
e
e
.
D
e
e
p
e
r
t
r
e
e
s
c
a
n
mo
d
e
l
mo
r
e
c
o
mp
l
e
x
p
a
t
t
e
r
n
s
b
u
t
ma
y
l
e
a
d
t
o
o
v
e
r
f
i
t
t
i
n
g
.
l
e
a
r
n
i
n
g
_
r
a
t
e
0
.
6
Th
e
st
e
p
si
z
e
s
h
r
i
n
k
a
g
e
u
se
d
i
n
e
a
c
h
b
o
o
s
t
i
n
g
st
e
p
t
o
p
r
e
v
e
n
t
o
v
e
r
f
i
t
t
i
n
g
.
A
h
i
g
h
e
r
l
e
a
r
n
i
n
g
r
a
t
e
s
p
e
e
d
s
u
p
l
e
a
r
n
i
n
g
b
u
t
ma
y
r
e
q
u
i
r
e
c
a
r
e
f
u
l
t
u
n
i
n
g
.
r
a
n
d
o
m
_
s
t
a
t
e
42
Th
e
s
e
e
d
u
s
e
d
b
y
t
h
e
r
a
n
d
o
m n
u
m
b
e
r
g
e
n
e
r
a
t
o
r
f
o
r
r
e
p
r
o
d
u
c
i
b
i
l
i
t
y
.
En
s
u
r
e
s
t
h
a
t
t
h
e
r
e
su
l
t
s
c
a
n
b
e
r
e
p
l
i
c
a
t
e
d
.
o
b
j
e
c
t
i
v
e
'b
i
n
a
r
y
'
Lo
ss f
u
n
c
t
i
o
n
f
o
r
b
i
n
a
r
y
c
l
a
s
si
f
i
c
a
t
i
o
n
t
a
s
k
s.
b
o
o
s
t
e
r
'g
b
t
r
e
e
'
S
p
e
c
i
f
i
e
s t
h
e
l
e
a
r
n
e
r
t
y
p
e
a
s '
g
b
t
r
e
e
',
i
n
d
i
c
a
t
i
n
g
t
h
a
t
t
h
e
mo
d
e
l
w
i
l
l
b
e
a
n
e
n
semb
l
e
o
f
DT
.
4.
P
RO
P
O
SE
D
F
RA
M
E
WO
R
K
T
h
is
s
t
u
d
y
i
n
t
r
o
d
u
ce
s
tw
o
ad
v
a
n
c
e
d
a
p
p
r
o
ac
h
es
d
esi
g
n
e
d
to
a
d
d
r
ess
c
lass
im
b
ala
n
c
e
in
p
h
is
h
i
n
g
web
s
ite
d
et
ec
ti
o
n
d
a
tas
ets:
SMOT
E
E
NN
-
XGB
an
d
SMOT
E
T
o
m
e
k
-
XGB.
B
o
t
h
m
e
th
o
d
s
i
n
t
e
g
r
ate
s
o
p
h
is
tic
at
ed
r
esa
m
p
li
n
g
t
ec
h
n
iq
u
es
wit
h
a
p
o
we
r
f
u
l c
lass
i
f
i
c
ati
o
n
a
lg
o
r
it
h
m
t
o
en
h
an
ce
d
et
ec
ti
o
n
p
e
r
f
o
r
m
a
n
c
e
as
s
h
o
wn
i
n
F
ig
u
r
e
3
.
T
h
e
f
i
g
u
r
e
ill
u
s
t
r
at
es
th
e
o
v
e
r
a
ll
wo
r
k
f
lo
w
f
o
r
a
d
d
r
ess
i
n
g
c
lass
im
b
al
a
n
ce
i
n
a
d
atas
et
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
E
n
h
a
n
ci
n
g
p
h
is
h
in
g
w
eb
s
ite
d
etec
tio
n
:
a
co
m
p
a
r
a
tive
s
tu
d
y
o
f S
MOTETo
mek
-
X
GB
a
n
d
…
(
K
a
ma
l O
ma
r
i
)
2941
an
d
e
v
al
u
ati
n
g
t
h
e
X
GB
o
o
s
t
cl
ass
i
f
ie
r
'
s
p
e
r
f
o
r
m
a
n
ce
.
I
t
b
e
g
i
n
s
wit
h
s
o
u
r
ci
n
g
t
h
e
d
at
aset
f
r
o
m
t
h
e
UC
I
d
at
a
b
ase
,
f
o
ll
o
we
d
b
y
t
h
e
id
e
n
ti
f
i
ca
t
io
n
o
f
i
m
b
ala
n
ce
d
c
las
s
d
is
t
r
i
b
u
ti
o
n
s
.
T
o
r
ec
ti
f
y
th
i
s
,
h
y
b
r
i
d
s
a
m
p
li
n
g
tec
h
n
i
q
u
es s
u
ch
as
SMO
T
E
E
NN
a
n
d
SMO
T
E
T
o
m
e
k
a
r
e
a
p
p
li
ed
,
r
esu
lti
n
g
in
a
c
lass
-
b
al
an
ce
d
d
a
tase
t.
Fig
u
r
e
3
.
Pro
p
o
s
ed
f
r
am
ewo
r
k
T
h
is
d
ataset
is
s
p
lit
in
to
tr
ain
in
g
an
d
test
in
g
s
ets,
with
8
0
%
o
f
th
e
d
ata
u
s
ed
f
o
r
tr
ain
in
g
an
d
2
0
%
r
eser
v
ed
f
o
r
test
in
g
,
f
o
llo
win
g
wid
ely
ac
ce
p
ted
r
ec
o
m
m
en
d
a
tio
n
s
in
m
ac
h
in
e
lear
n
in
g
liter
atu
r
e.
T
h
e
tr
ai
n
in
g
d
ata,
n
o
w
b
alan
ce
d
,
is
u
s
ed
to
tr
ain
t
h
e
XGBo
o
s
t
class
if
ier
,
wh
ile
th
e
test
in
g
d
ata
is
r
eser
v
ed
f
o
r
m
o
d
el
ev
alu
atio
n
.
Po
s
t
-
tr
ain
in
g
,
th
e
XGBo
o
s
t
m
o
d
el
is
ap
p
lied
to
class
if
y
n
ew
d
ata
in
to
p
h
i
s
h
in
g
o
r
leg
itima
te
ca
teg
o
r
ies.
T
h
e
m
o
d
el'
s
ef
f
ec
tiv
en
ess
is
th
en
ass
es
s
ed
u
s
in
g
m
etr
ics
s
u
ch
as
ac
cu
r
ac
y
,
F1
-
s
co
r
e,
r
ec
all,
p
r
ec
is
io
n
,
r
ec
eiv
e
r
o
p
er
atin
g
ch
ar
ac
ter
is
tic
-
ar
ea
u
n
d
er
th
e
cu
r
v
e
(
R
OC
-
AUC)
,
an
d
g
e
o
m
etr
ic
m
ea
n
s
co
r
e
(
G
-
m
ea
n
)
,
d
em
o
n
s
tr
atin
g
th
e
i
m
p
ac
t o
f
h
y
b
r
id
s
am
p
lin
g
tec
h
n
iq
u
es o
n
im
p
r
o
v
in
g
m
o
d
el
p
e
r
f
o
r
m
a
n
ce
.
4
.
1
.
C
o
mp
a
r
i
s
o
n
o
f
c
l
a
s
s
d
i
s
t
r
i
b
u
t
i
o
n
s
b
e
f
o
r
e
a
n
d
a
f
t
e
r
a
p
p
l
y
S
M
O
T
E
E
N
N
a
n
d
S
M
O
T
E
T
o
m
e
k
t
e
c
h
n
i
q
u
e
s
T
ab
le
5
illu
s
tr
ates
th
e
class
d
is
tr
ib
u
tio
n
o
f
th
e
o
r
ig
i
n
al
d
atas
et
as
well
as
th
e
d
atasets
o
b
ta
in
ed
af
ter
ap
p
ly
in
g
th
e
p
r
o
p
o
s
ed
SMOT
E
E
NN
an
d
SMOT
E
T
o
m
ek
m
eth
o
d
s
.
T
a
b
le
5
p
r
esen
ts
a
co
m
p
ar
is
o
n
o
f
s
am
p
le
d
is
tr
ib
u
tio
n
s
b
ef
o
r
e
an
d
a
f
ter
ap
p
ly
in
g
two
h
y
b
r
id
-
s
am
p
lin
g
tech
n
iq
u
es:
SMOT
E
E
NN
an
d
SMOT
E
T
o
m
ek
.
T
h
e
"Or
ig
in
al
s
am
p
les"
co
lu
m
n
s
h
o
ws
th
e
in
itial
co
u
n
t
a
n
d
p
er
ce
n
tag
e
d
is
tr
ib
u
tio
n
o
f
th
e
leg
itima
te
an
d
p
h
is
h
in
g
class
es,
with
leg
itima
te
s
am
p
les
at
4
4
.
3
0
1
%
an
d
p
h
is
h
in
g
s
am
p
les
at
5
5
.
6
9
8
%.
Af
ter
ap
p
l
y
in
g
SMOT
E
E
NN,
th
e
d
is
tr
ib
u
tio
n
o
f
leg
itima
te
s
am
p
les
in
c
r
ea
s
es
to
5
0
.
7
3
6
%
an
d
p
h
is
h
in
g
s
am
p
les
d
ec
r
ea
s
e
to
4
9
.
2
6
3
%,
r
esu
ltin
g
in
a
m
o
r
e
b
alan
ce
d
d
ataset.
Similar
ly
,
ap
p
ly
in
g
SMOT
E
T
o
m
ek
r
e
s
u
lts
in
an
eq
u
al
d
is
tr
ib
u
tio
n
o
f
5
0
.
0
0
%
f
o
r
b
o
th
leg
itima
te
an
d
p
h
is
h
in
g
s
am
p
les,
f
u
r
th
er
b
alan
cin
g
th
e
d
ataset.
T
h
is
co
m
p
ar
is
o
n
h
ig
h
lig
h
ts
h
o
w
ea
ch
s
am
p
lin
g
tech
n
iq
u
e
a
d
ju
s
ts
th
e
class
d
is
tr
ib
u
tio
n
to
ad
d
r
ess
class
im
b
alan
ce
.
T
ab
le
5
.
T
h
e
class
d
is
tr
ib
u
tio
n
C
l
a
s
s
Le
g
i
t
i
ma
t
e
P
h
i
s
h
i
n
g
O
r
i
g
i
n
a
l
s
a
mp
l
e
s
4
,
8
9
8
6
,
1
5
7
D
i
st
r
i
b
u
t
i
o
n
(
%)
4
4
.
3
0
1
5
5
.
6
9
8
S
M
O
TEENN
s
a
m
p
l
e
s
4
,
9
2
6
4
,
7
8
3
D
i
st
r
i
b
u
t
i
o
n
(
%)
5
0
.
7
3
6
4
9
.
2
6
3
S
M
O
TET
o
m
e
k
sam
p
l
e
s
4
,
9
5
0
4
,
9
5
0
D
i
st
r
i
b
u
t
i
o
n
(
%)
5
0
.
0
0
5
0
.
0
0
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.
1
5
,
No
.
3
,
J
u
n
e
2
0
2
6
:
2
9
3
5
-
2
9
4
5
2942
4
.
2
.
E
v
a
lua
t
i
o
n m
et
rics
E
v
alu
atio
n
m
etr
ics
p
lay
a
v
ita
l
r
o
le
i
n
ass
ess
in
g
th
e
p
er
f
o
r
m
an
ce
o
f
m
ac
h
in
e
lear
n
in
g
m
o
d
els.
T
h
ey
p
r
o
v
id
e
q
u
an
titativ
e
m
ea
s
u
r
e
s
s
u
ch
as
ac
c
u
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
F1
-
s
co
r
e,
R
OC
-
AUC
,
an
d
th
e
G
-
m
ea
n
,
wh
ich
ar
e
ess
en
tial
f
o
r
co
m
p
ar
in
g
d
if
f
e
r
en
t
m
o
d
els.
T
h
ese
m
etr
ics
h
elp
r
esear
ch
er
s
d
eter
m
in
e
wh
ich
ap
p
r
o
ac
h
is
b
est
s
u
ited
f
o
r
th
eir
s
p
ec
if
ic
task
s
,
en
s
u
r
in
g
th
at
th
e
ch
o
s
en
m
o
d
el
ef
f
ec
tiv
ely
h
an
d
les
th
e
ch
allen
g
es o
f
th
e
p
r
o
b
lem
at
h
an
d
.
i)
Acc
u
r
ac
y
:
ac
cu
r
ac
y
m
ea
s
u
r
es
th
e
o
v
er
all
co
r
r
ec
tn
ess
o
f
a
m
o
d
el,
ca
lcu
lated
as
th
e
r
atio
o
f
co
r
r
ec
tly
p
r
ed
icted
in
s
tan
ce
s
to
th
e
to
tal
in
s
tan
ce
s
in
th
e
d
ataset.
W
h
ile
it
is
in
tu
itiv
e
an
d
s
tr
aig
h
tf
o
r
war
d
,
ac
cu
r
ac
y
m
ay
n
o
t
b
e
id
ea
l
f
o
r
im
b
alan
ce
d
d
atasets
as
it
ca
n
b
e
m
is
lead
in
g
wh
en
o
n
e
class
d
o
m
in
ates
,
as
d
ef
in
ed
in
(
2
)
.
=
(
2
)
ii)
F1
-
s
co
r
e:
th
e
F1
-
s
co
r
e
is
a
b
alan
ce
d
m
etr
ic
th
at
co
n
s
id
er
s
b
o
th
p
r
ec
is
io
n
a
n
d
r
ec
all.
I
t
is
esp
ec
ially
u
s
ef
u
l
f
o
r
im
b
alan
ce
d
d
atasets
,
as
it
ca
lcu
lates
th
e
h
ar
m
o
n
i
c
m
ea
n
o
f
p
r
ec
is
io
n
an
d
r
ec
al
l,
p
r
o
v
id
in
g
a
s
in
g
le
v
alu
e
th
at
b
ala
n
ce
s
th
e
tr
ad
e
-
o
f
f
b
etwe
en
f
alse p
o
s
itiv
es a
n
d
f
alse n
eg
ativ
es
,
as sh
o
wn
in
(
3
)
.
1
−
=
2
×
×
+
(
3
)
iii)
R
ec
all:
r
ec
all,
also
k
n
o
wn
as
s
en
s
itiv
ity
o
r
tr
u
e
p
o
s
itiv
e
r
at
e,
m
ea
s
u
r
es
th
e
p
r
o
p
o
r
tio
n
o
f
ac
tu
al
p
o
s
itiv
e
in
s
tan
ce
s
co
r
r
ec
tly
id
e
n
tifie
d
b
y
th
e
m
o
d
el.
I
t
is
cr
u
cial
w
h
en
m
in
im
izin
g
f
alse
n
e
g
ativ
es
is
ess
en
tial,
f
o
cu
s
in
g
o
n
co
r
r
ec
tly
id
e
n
tify
i
n
g
as m
an
y
p
o
s
itiv
e
in
s
tan
ce
s
as p
o
s
s
ib
le
,
as d
ef
in
ed
in
(
4
)
.
=
+
(
4
)
iv
)
Pre
cisi
o
n
:
p
r
ec
is
io
n
r
ep
r
esen
ts
th
e
p
r
o
p
o
r
tio
n
o
f
tr
u
e
p
o
s
itiv
e
p
r
ed
ictio
n
s
am
o
n
g
all
p
o
s
itiv
e
p
r
ed
ictio
n
s
m
ad
e
b
y
th
e
m
o
d
el.
I
t
is
v
ital
wh
en
th
e
co
s
t
o
f
f
alse
p
o
s
it
i
v
es
is
h
ig
h
,
aim
in
g
to
r
ed
u
ce
th
e
n
u
m
b
er
o
f
in
co
r
r
ec
tly
class
if
ied
p
o
s
itiv
e
in
s
tan
ce
s
,
as sh
o
wn
in
(
5
)
.
=
+
(
5
)
v)
R
OC
-
AU
C
:
th
e
R
OC
-
AU
C
m
ea
s
u
r
es
th
e
m
o
d
el'
s
ab
ilit
y
to
d
is
tin
g
u
is
h
b
etwe
en
class
es
[
2
6
]
.
I
t
p
lo
ts
t
h
e
tr
u
e
p
o
s
itiv
e
r
ate
(
r
ec
all)
ag
ai
n
s
t
th
e
f
alse
p
o
s
itiv
e
r
ate,
p
r
o
v
id
in
g
a
s
in
g
le
v
alu
e
t
h
at
s
u
m
m
ar
izes
th
e
m
o
d
el'
s
p
er
f
o
r
m
an
c
e
ac
r
o
s
s
a
ll
class
if
icatio
n
th
r
esh
o
ld
s
.
A
h
ig
h
er
R
OC
-
AUC
in
d
icate
s
b
etter
m
o
d
e
l
p
er
f
o
r
m
an
ce
.
v
i)
G
-
m
ea
n
:
th
e
G
-
m
ea
n
is
a
m
etr
ic
th
at
b
alan
ce
s
th
e
p
er
f
o
r
m
an
ce
o
f
a
m
o
d
el
ac
r
o
s
s
b
o
th
class
es,
esp
ec
ially
in
im
b
alan
ce
d
d
ata
s
ets.
I
t
is
ca
lcu
lated
as
th
e
s
q
u
ar
e
r
o
o
t
o
f
th
e
p
r
o
d
u
ct
o
f
th
e
tr
u
e
p
o
s
itiv
e
r
ate
(
r
ec
all)
an
d
th
e
tr
u
e
n
e
g
ativ
e
r
ate
(
s
p
ec
if
icity
)
.
G
-
m
e
an
en
s
u
r
es
th
at
t
h
e
m
o
d
el
p
er
f
o
r
m
s
well
o
n
b
o
th
th
e
m
in
o
r
ity
a
n
d
m
aj
o
r
ity
class
es,
p
r
o
v
id
in
g
a
b
alan
ce
d
ev
alu
atio
n
[
2
7
]
.
I
n
s
u
m
m
ar
y
,
a
c
o
m
b
in
atio
n
o
f
m
u
ltip
le
ev
alu
atio
n
m
etr
ics
p
r
o
v
id
es
a
c
o
m
p
r
e
h
en
s
iv
e
an
d
r
eliab
le
ass
es
s
m
en
t
o
f
m
o
d
el
p
er
f
o
r
m
an
ce
.
Sp
ec
if
ically
,
ac
cu
r
ac
y
m
ea
s
u
r
es
o
v
er
all
co
r
r
ec
tn
ess
,
wh
ile
th
e
F1
-
s
co
r
e
b
alan
ce
s
p
r
ec
is
io
n
an
d
r
ec
al
l
to
h
an
d
le
im
b
alan
ce
d
d
ata
s
ets
ef
f
ec
tiv
ely
.
I
n
ad
d
itio
n
,
r
ec
all
em
p
h
asize
s
m
in
im
izin
g
f
alse
n
eg
ativ
es,
p
r
ec
is
io
n
f
o
cu
s
es
o
n
r
ed
u
cin
g
f
alse
p
o
s
itiv
es,
R
O
C
-
AU
C
ev
alu
ates
th
e
m
o
d
el’
s
ab
ilit
y
to
d
is
tin
g
u
is
h
b
etwe
en
class
es,
an
d
th
e
G
-
m
ea
n
en
s
u
r
es b
alan
ce
d
p
er
f
o
r
m
a
n
ce
ac
r
o
s
s
all
clas
s
es.
5.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
is
s
tu
d
y
co
m
p
ar
e
s
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
XGB
m
o
d
el,
wh
ich
d
o
esn
'
t
ad
d
r
ess
clas
s
im
b
alan
ce
,
to
two
ad
v
an
ce
d
m
eth
o
d
s
s
p
ec
if
ically
d
esig
n
ed
to
h
an
d
le
th
is
is
s
u
e
in
p
h
is
h
in
g
web
s
ite
d
etec
tio
n
:
SMOT
E
E
NN
an
d
SMOT
E
T
o
m
e
k
.
T
h
ese
m
eth
o
d
s
co
m
b
in
e
s
o
p
h
is
ticated
r
esam
p
lin
g
tech
n
iq
u
es
with
th
e
XGB
alg
o
r
ith
m
to
im
p
r
o
v
e
d
etec
tio
n
p
er
f
o
r
m
a
n
c
e,
cr
ea
tin
g
a
m
o
r
e
b
ala
n
ce
d
cl
ass
d
is
tr
ib
u
tio
n
an
d
p
o
te
n
tially
m
ak
in
g
th
e
m
o
d
el
m
o
r
e
ef
f
ec
tiv
e.
T
h
e
r
esu
lts
o
b
tain
ed
f
r
o
m
th
e
ex
p
er
im
e
n
tatio
n
u
s
in
g
th
e
s
cik
it
-
lear
n
to
o
l
h
ig
h
lig
h
t
a
co
m
p
ar
is
o
n
b
etwe
en
t
h
r
ee
cl
ass
if
ier
s
XG
B
,
SMOT
E
E
NN
-
XGB,
an
d
SMOT
E
T
o
m
e
k
-
XGB,
as
s
h
o
wn
in
T
ab
le
6
.
Am
o
n
g
th
em
,
SMOT
E
T
o
m
ek
-
XGB m
ar
g
in
ally
o
u
t
p
er
f
o
r
m
s
th
e
o
th
er
s
in
k
ey
m
etr
ics,
p
ar
ticu
lar
ly
in
ac
cu
r
ac
y
(
0
.
9
7
8
)
,
F1
-
s
co
r
e
(
0
.
9
8
1
)
,
r
ec
all
(
0
.
9
8
2
)
,
an
d
th
e
G
-
m
ea
n
(
0
.
9
7
8
)
,
p
o
s
itio
n
in
g
it
as
th
e
m
o
s
t
b
alan
ce
d
an
d
ef
f
ec
tiv
e
ap
p
r
o
a
ch
f
o
r
ad
d
r
ess
in
g
class
im
b
ala
n
ce
in
p
h
is
h
in
g
web
s
ite
d
etec
t
io
n
.
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
E
n
h
a
n
ci
n
g
p
h
is
h
in
g
w
eb
s
ite
d
etec
tio
n
:
a
co
m
p
a
r
a
tive
s
tu
d
y
o
f S
MOTETo
mek
-
X
GB
a
n
d
…
(
K
a
ma
l O
ma
r
i
)
2943
Alth
o
u
g
h
XGB
d
o
es
n
o
t
s
p
ec
if
ically
tar
g
et
class
im
b
alan
ce
,
it
p
e
r
f
o
r
m
s
r
em
ar
k
a
b
ly
well,
ac
h
iev
in
g
a
h
ig
h
R
OC
-
AUC
(
0
.
9
9
8
)
an
d
s
tr
o
n
g
o
v
er
all
ac
cu
r
ac
y
(
0
.
9
7
7
)
,
d
e
m
o
n
s
tr
at
in
g
its
r
o
b
u
s
tn
ess
.
SMOT
E
E
NN
-
XGB
ex
ce
ls
in
p
r
ec
is
io
n
(
0
.
9
8
2
)
,
ef
f
ec
tiv
el
y
m
in
im
izin
g
f
alse
p
o
s
itiv
es,
b
u
t
ex
h
i
b
its
a
m
in
o
r
tr
ad
e
-
o
f
f
in
r
ec
all
(
0
.
9
7
2
)
a
n
d
th
e
G
-
m
ea
n
(
0
.
9
7
5
)
wh
en
co
m
p
ar
ed
to
SMOT
E
T
o
m
ek
-
XGB.
T
h
is
an
aly
s
is
s
u
g
g
ests
th
at
wh
ile
all
th
r
ee
m
eth
o
d
s
ar
e
h
ig
h
ly
ef
f
ec
tiv
e,
SMOT
E
T
o
m
ek
-
XGB
p
r
o
v
id
es
th
e
b
est
b
alan
ce
b
etwe
en
s
en
s
itiv
ity
an
d
s
p
ec
if
icity
,
m
ak
in
g
it th
e
m
o
s
t su
itab
le
ch
o
ice
f
o
r
th
is
ap
p
licatio
n
.
T
ab
le
6
.
E
v
alu
atio
n
r
esu
lts
in
(
%)
C
l
a
s
si
f
i
e
r
A
c
c
u
r
a
c
y
F1
-
sc
o
r
e
R
e
c
a
l
l
P
r
e
c
i
s
i
o
n
R
O
C
-
AUC
G
-
mea
n
XGB
0
.
9
7
7
0
.
9
8
0
0
.
9
7
9
0
.
9
8
0
0
.
9
9
8
0
.
9
7
7
S
M
O
TEENN
-
XGB
0
.
9
7
4
0
.
9
7
7
0
.
9
7
2
0
.
9
8
2
0
.
9
9
7
0
.
9
7
5
S
M
O
TET
o
m
e
k
-
XGB
0
.
9
7
8
0
.
9
8
1
0
.
9
8
2
0
.
9
7
9
0
.
9
9
8
0
.
9
7
8
Af
ter
a
th
o
r
o
u
g
h
ex
a
m
in
atio
n
o
f
o
u
r
r
esear
ch
f
in
d
in
g
s
,
we
will
p
r
o
ce
ed
t
o
co
n
d
u
ct
a
co
m
p
ar
ativ
e
an
aly
s
is
with
o
th
er
r
elev
an
t
s
tu
d
ies
in
th
e
f
ield
(
T
ab
le
7
)
.
I
n
co
m
p
ar
is
o
n
to
th
e
r
ef
er
en
ce
d
s
tu
d
ies,
o
u
r
class
if
ier
s
,
p
ar
ticu
lar
ly
SMOT
E
T
o
m
ek
-
XGB
an
d
XGB,
ex
h
ib
it
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
,
with
SMOT
E
T
o
m
ek
-
XGB
ac
h
iev
in
g
an
ac
cu
r
ac
y
o
f
9
7
.
8
%
an
d
XGB
r
ea
ch
in
g
9
7
.
7
%.
T
h
ese
r
esu
lts
s
lig
h
tly
s
u
r
p
ass
th
e
b
est
ac
cu
r
ac
ies
r
ep
o
r
ted
b
y
Als
ar
ier
a
et
a
l.
[
28
]
(
9
7
.
4
%)
a
n
d
Aln
e
m
ar
i
an
d
Als
h
am
m
ar
i
[
29
]
(
9
7
.
3
%)
u
s
in
g
ad
v
a
n
ce
d
e
n
s
em
b
le
m
eth
o
d
s
an
d
RF
.
No
tab
ly
,
th
e
m
o
d
els
n
o
t
o
n
ly
m
at
ch
o
r
ex
ce
e
d
th
ese
ac
cu
r
ac
ies
b
u
t
also
d
em
o
n
s
tr
ate
ex
ce
llen
ce
in
o
th
er
k
e
y
m
etr
ics,
s
u
ch
as
F1
-
s
co
r
e,
r
ec
all,
p
r
ec
is
io
n
,
R
OC
-
AU
C
,
an
d
G
-
m
ea
n
.
T
h
i
s
in
d
icate
s
th
at
th
e
h
y
b
r
id
s
am
p
lin
g
tech
n
iq
u
es
co
m
b
in
e
d
with
XGB
in
o
u
r
s
tu
d
y
p
r
o
v
id
e
a
m
o
r
e
b
alan
ce
d
an
d
r
o
b
u
s
t
ap
p
r
o
ac
h
to
p
h
is
h
in
g
web
s
ite
d
etec
tio
n
co
m
p
a
r
ed
to
th
e
m
et
h
o
d
s
em
p
lo
y
ed
in
th
e
r
e
f
er
en
ce
d
r
e
s
ea
r
ch
.
T
ab
le
7
.
E
v
alu
atio
n
o
f
ex
is
tin
g
p
h
is
h
in
g
d
o
m
ai
n
d
etec
tio
n
m
o
d
els
A
u
t
h
o
r
s
D
a
t
a
s
e
t
A
l
g
o
r
i
t
h
m
A
c
c
u
r
a
c
y
(
%)
U
b
i
n
g
e
t
a
l
.
[
30
]
U
C
I
En
se
mb
l
e
b
a
g
g
i
n
g
,
b
o
o
s
t
i
n
g
,
a
n
d
s
t
a
c
k
i
n
g
9
5
.
4
A
l
sari
e
r
a
e
t
a
l
.
[
28
]
U
C
I
F
o
r
e
st
P
A
-
P
W
D
M
,
B
a
g
g
e
d
-
F
o
r
e
s
t
P
A
-
P
W
D
M
,
a
n
d
A
d
a
b
-
F
o
r
e
s
t
P
A
-
P
W
D
M
9
6
.
2
6
,
9
6
.
5
,
a
n
d
9
7
.
4
La
k
s
h
m
i
e
t
a
l
.
[3
1
]
U
C
I
D
N
N
+
A
d
a
m
9
6
.
0
0
A
l
n
e
mari
a
n
d
A
l
s
h
a
mm
a
r
i
[
29
]
U
C
I
RF
9
7
.
3
O
mari
[
8
]
U
C
I
G
r
a
d
i
e
n
t
b
o
o
s
t
9
7
.
2
6.
CO
NCLU
SI
O
N
T
h
is
s
tu
d
y
h
as u
n
d
er
s
co
r
e
d
th
e
cr
itical
im
p
o
r
tan
ce
o
f
ad
d
r
ess
in
g
class
im
b
alan
ce
in
p
h
is
h
in
g
web
s
ite
d
etec
tio
n
th
r
o
u
g
h
th
e
ap
p
lic
atio
n
o
f
ad
v
an
ce
d
r
esam
p
lin
g
tech
n
iq
u
es.
B
y
co
m
p
ar
in
g
th
r
ee
ap
p
r
o
ac
h
es
XGB
o
o
s
t
,
SMOT
E
E
NN
-
X
G
B
,
an
d
SMOT
E
T
o
m
ek
-
XGB
th
is
r
esear
ch
h
ig
h
lig
h
ts
th
e
ef
f
ec
tiv
en
ess
o
f
in
teg
r
atin
g
h
y
b
r
id
s
am
p
lin
g
m
eth
o
d
s
with
a
r
o
b
u
s
t
class
if
ier
lik
e
XGBo
o
s
t.
T
h
e
r
esu
lts
in
d
icate
th
at
wh
ile
all
m
eth
o
d
s
p
er
f
o
r
m
ex
ce
p
tio
n
all
y
well,
SMOT
E
T
o
m
ek
-
XGB
o
f
f
er
s
th
e
m
o
s
t
b
alan
ce
d
p
er
f
o
r
m
an
ce
ac
r
o
s
s
k
ey
m
etr
ics,
p
r
o
v
id
in
g
an
o
p
tim
al
tr
ad
e
-
o
f
f
b
etwe
en
s
en
s
itiv
ity
an
d
s
p
ec
if
icity
.
T
h
is
m
a
k
es
it
p
ar
ticu
lar
ly
well
-
s
u
ited
f
o
r
p
h
is
h
in
g
d
etec
tio
n
task
s
,
wh
er
e
b
o
th
f
alse
p
o
s
itiv
es
an
d
f
alse
n
eg
ativ
es
ca
n
h
av
e
s
ig
n
if
ican
t
co
n
s
eq
u
en
ce
s
.
T
h
e
f
in
d
in
g
s
em
p
h
asize
th
at
ca
r
ef
u
l
co
n
s
id
er
atio
n
o
f
class
im
b
alan
ce
s
tr
ateg
ies
ca
n
s
ig
n
if
ican
tly
en
h
an
ce
th
e
p
er
f
o
r
m
an
ce
o
f
m
ac
h
in
e
lear
n
in
g
m
o
d
els
in
cy
b
er
s
ec
u
r
ity
a
p
p
l
icatio
n
s
,
u
ltima
tely
co
n
tr
ib
u
tin
g
to
m
o
r
e
r
eliab
le
a
n
d
ac
cu
r
ate
d
etec
tio
n
s
y
s
tem
s
.
Fu
tu
r
e
r
esear
ch
co
u
l
d
ex
p
l
o
r
e
th
e
in
teg
r
atio
n
o
f
th
ese
tech
n
iq
u
es
with
o
th
er
c
lass
if
ier
s
o
r
in
d
if
f
er
en
t
d
o
m
ain
s
wh
er
e
class
im
b
alan
ce
is
p
r
ev
alen
t,
th
e
r
eb
y
b
r
o
ad
e
n
in
g
th
e
ap
p
licab
ilit
y
o
f
th
ese
f
in
d
in
g
s
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
Au
th
o
r
s
s
tate
th
at
n
o
f
u
n
d
in
g
was
r
ec
eiv
ed
f
o
r
th
is
r
esear
ch
.
T
h
is
wo
r
k
was
co
n
d
u
cted
in
d
ep
en
d
en
tly
with
o
u
t
f
in
an
ci
al
s
u
p
p
o
r
t
f
r
o
m
an
y
f
u
n
d
in
g
a
g
en
cy
,
i
n
s
titu
tio
n
,
o
r
o
r
g
an
iza
tio
n
.
C
o
n
s
eq
u
en
tly
,
n
o
g
r
a
n
t n
u
m
b
er
s
o
r
ex
ter
n
al
s
p
o
n
s
o
r
s
h
ip
s
ar
e
ass
o
ciate
d
wit
h
th
is
s
tu
d
y
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
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.
1
5
,
No
.
3
,
J
u
n
e
2
0
2
6
:
2
9
3
5
-
2
9
4
5
2944
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Kam
al
Om
ar
i
✓
✓
✓
✓
✓
✓
✓
✓
Ay
o
u
b
Ou
k
h
atar
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
th
at
th
ey
h
av
e
n
o
co
n
f
lict
o
f
in
ter
est.
T
h
e
a
u
th
o
r
s
d
ec
lar
e
th
at
t
h
ey
h
a
v
e
n
o
k
n
o
wn
f
in
an
cial,
p
er
s
o
n
al,
o
r
p
r
o
f
ess
io
n
al
r
elatio
n
s
h
ip
s
th
at
co
u
ld
h
av
e
in
f
lu
en
ce
d
th
e
wo
r
k
r
e
p
o
r
t
ed
in
th
is
p
ap
e
r
.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
ata
th
at
s
u
p
p
o
r
t
th
e
f
in
d
i
n
g
s
o
f
th
is
s
tu
d
y
ar
e
o
p
en
ly
a
v
ailab
le
in
UC
I
r
v
in
e
Ma
ch
in
e
L
ea
r
n
in
g
R
ep
o
s
ito
r
y
at
h
ttp
s
://ar
ch
iv
e.
i
cs.u
ci.
ed
u
/m
l/d
atasets
/p
h
is
h
in
g
+w
eb
s
ites
,
r
ef
er
en
ce
[
2
2
]
.
RE
F
E
R
E
NC
E
S
[
1
]
A
P
W
G
,
“
P
h
i
sh
i
n
g
a
c
t
i
v
i
t
y
t
r
e
n
d
s
r
e
p
o
r
t
1
st
q
u
a
r
t
e
r
(
2
0
2
4
)
:
u
n
i
f
y
i
n
g
t
h
e
g
l
o
b
a
l
r
e
s
p
o
n
se
t
o
c
y
b
e
r
c
r
i
me,
”
d
o
c
s.
a
p
w
g
.
o
rg
.
2
0
2
4
.
A
c
c
e
ss
e
d
:
J
u
l
.
2
5
,
2
0
2
4
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
d
o
c
s.a
p
w
g
.
o
r
g
/
r
e
p
o
r
t
s
/
a
p
w
g
_
t
r
e
n
d
s
_
r
e
p
o
r
t
_
q
1
_
2
0
2
4
.
p
d
f
[
2
]
K
.
N
a
g
a
r
a
j
,
B
.
B
h
a
t
t
a
c
h
a
r
j
e
e
,
A
.
S
r
i
d
h
a
r
,
a
n
d
S
.
G
S
,
“
D
e
t
e
c
t
i
o
n
o
f
p
h
i
s
h
i
n
g
w
e
b
s
i
t
e
s
u
s
i
n
g
a
n
o
v
e
l
t
w
o
f
o
l
d
e
n
sem
b
l
e
mo
d
e
l
,
”
J
o
u
rn
a
l
o
f
S
y
st
e
m
s
a
n
d
I
n
f
o
rm
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
,
v
o
l
.
2
0
,
n
o
.
3
,
p
p
.
3
2
1
–
3
5
7
,
N
o
v
.
2
0
1
8
,
d
o
i
:
1
0
.
1
1
0
8
/
JS
I
T
-
09
-
2
0
1
7
-
0
0
7
4
.
[
3
]
M
.
B
u
d
a
,
A
.
M
a
k
i
,
a
n
d
M
.
A
.
M
a
z
u
r
o
w
sk
i
,
“
A
sy
s
t
e
m
a
t
i
c
s
t
u
d
y
o
f
t
h
e
c
l
a
ss
i
m
b
a
l
a
n
c
e
p
r
o
b
l
e
m
i
n
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s,”
N
e
u
r
a
l
N
e
t
w
o
r
k
s
,
v
o
l
.
1
0
6
,
p
p
.
2
4
9
–
2
5
9
,
O
c
t
.
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
n
e
u
n
e
t
.
2
0
1
8
.
0
7
.
0
1
1
.
[
4
]
M
.
K
h
u
sh
i
e
t
a
l
.
,
“
A
c
o
mp
a
r
a
t
i
v
e
p
e
r
f
o
r
man
c
e
a
n
a
l
y
s
i
s
o
f
d
a
t
a
r
e
sam
p
l
i
n
g
m
e
t
h
o
d
s
o
n
i
m
b
a
l
a
n
c
e
me
d
i
c
a
l
d
a
t
a
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
9
,
p
p
.
1
0
9
9
6
0
–
1
0
9
9
7
5
,
2
0
2
1
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
1
.
3
1
0
2
3
9
9
.
[
5
]
D
.
El
r
e
e
d
y
,
A
.
F
.
A
t
i
y
a
,
a
n
d
F
.
K
a
mal
o
v
,
“
A
t
h
e
o
r
e
t
i
c
a
l
d
i
s
t
r
i
b
u
t
i
o
n
a
n
a
l
y
s
i
s
o
f
s
y
n
t
h
e
t
i
c
m
i
n
o
r
i
t
y
o
v
e
r
sa
mp
l
i
n
g
t
e
c
h
n
i
q
u
e
(
S
M
O
TE)
f
o
r
i
m
b
a
l
a
n
c
e
d
l
e
a
r
n
i
n
g
,
”
M
a
c
h
i
n
e
L
e
a
rn
i
n
g
,
v
o
l
.
1
1
3
,
n
o
.
7
,
p
p
.
4
9
0
3
–
4
9
2
3
,
Ju
l
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
0
9
9
4
-
0
2
2
-
0
6
2
9
6
-
4.
[
6
]
C
.
B
e
n
t
é
j
a
c
,
A
.
C
sö
r
g
ő
,
a
n
d
G
.
M
.
-
M
u
ñ
o
z
,
“
A
c
o
m
p
a
r
a
t
i
v
e
a
n
a
l
y
si
s
o
f
g
r
a
d
i
e
n
t
b
o
o
st
i
n
g
a
l
g
o
r
i
t
h
ms,”
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
Re
v
i
e
w
,
v
o
l
.
5
4
,
n
o
.
3
,
p
p
.
1
9
3
7
–
1
9
6
7
,
M
a
r
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
0
4
6
2
-
0
2
0
-
0
9
8
9
6
-
5.
[
7
]
N
.
A
b
d
e
l
h
a
mi
d
,
F
.
T
h
a
b
t
a
h
,
a
n
d
H
.
A
.
-
J
a
b
e
r
,
“
P
h
i
s
h
i
n
g
d
e
t
e
c
t
i
o
n
:
a
r
e
c
e
n
t
i
n
t
e
l
l
i
g
e
n
t
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
c
o
m
p
a
r
i
s
o
n
b
a
s
e
d
o
n
mo
d
e
l
s
c
o
n
t
e
n
t
a
n
d
f
e
a
t
u
r
e
s,
”
i
n
2
0
1
7
I
EE
E
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
I
n
t
e
l
l
i
g
e
n
c
e
a
n
d
S
e
c
u
r
i
t
y
I
n
f
o
rm
a
t
i
c
s
(
I
S
I
)
,
J
u
l
.
2
0
1
7
,
p
p
.
7
2
–
77
,
d
o
i
:
1
0
.
1
1
0
9
/
I
S
I
.
2
0
1
7
.
8
0
0
4
8
7
7
.
[
8
]
K
.
O
m
a
r
i
,
“
C
o
m
p
a
r
a
t
i
v
e
st
u
d
y
o
f
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
ms f
o
r
p
h
i
s
h
i
n
g
w
e
b
si
t
e
d
e
t
e
c
t
i
o
n
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
A
d
v
a
n
c
e
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
1
4
,
n
o
.
9
,
2
0
2
3
,
d
o
i
:
1
0
.
1
4
5
6
9
/
I
JA
C
S
A
.
2
0
2
3
.
0
1
4
0
9
4
5
.
[
9
]
K
.
S
u
b
a
sh
i
n
i
a
n
d
V
.
N
a
r
mat
h
a
,
“
D
e
t
e
c
t
i
n
g
p
h
i
s
h
i
n
g
w
e
b
si
t
e
s
u
s
i
n
g
r
e
c
e
n
t
t
e
c
h
n
i
q
u
e
s
:
a
s
y
st
e
ma
t
i
c
l
i
t
e
r
a
t
u
r
e
r
e
v
i
e
w
,
”
I
T
M
W
e
b
o
f
C
o
n
f
e
re
n
c
e
s
,
v
o
l
.
5
7
,
N
o
v
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
5
1
/
i
t
mc
o
n
f
/
2
0
2
3
5
7
0
1
0
0
8
.
[
1
0
]
N
.
V
.
C
h
a
w
l
a
,
K
.
W
.
B
o
w
y
e
r
,
L.
O
.
H
a
l
l
,
a
n
d
W
.
P
.
K
e
g
e
l
me
y
e
r
,
“
S
M
O
TE
:
s
y
n
t
h
e
t
i
c
mi
n
o
r
i
t
y
o
v
e
r
-
s
a
m
p
l
i
n
g
t
e
c
h
n
i
q
u
e
,
”
J
o
u
rn
a
l
o
f
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
Re
s
e
a
rc
h
,
v
o
l
.
1
6
,
p
p
.
3
2
1
–
3
5
7
,
J
u
n
.
2
0
0
2
,
d
o
i
:
1
0
.
1
6
1
3
/
j
a
i
r
.
9
5
3
.
[
1
1
]
E.
F
.
S
w
a
n
a
,
W
.
D
o
o
r
sa
my
,
a
n
d
P
.
B
o
k
o
r
o
,
“
To
m
e
k
l
i
n
k
a
n
d
S
M
O
TE
a
p
p
r
o
a
c
h
e
s
f
o
r
m
a
c
h
i
n
e
f
a
u
l
t
c
l
a
ss
i
f
i
c
a
t
i
o
n
w
i
t
h
a
n
i
mb
a
l
a
n
c
e
d
d
a
t
a
se
t
,
”
S
e
n
so
rs
,
v
o
l
.
2
2
,
n
o
.
9
,
A
p
r
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
2
0
9
3
2
4
6
.
[
1
2
]
Z.
W
a
n
g
,
C
.
W
u
,
K
.
Z
h
e
n
g
,
X
.
N
i
u
,
a
n
d
X
.
W
a
n
g
,
“
S
M
O
TET
o
me
k
-
b
a
se
d
r
e
sa
mp
l
i
n
g
f
o
r
p
e
r
so
n
a
l
i
t
y
r
e
c
o
g
n
i
t
i
o
n
,
”
I
E
EE
Ac
c
e
ss
,
v
o
l
.
7
,
p
p
.
1
2
9
6
7
8
–
1
2
9
6
8
9
,
2
0
1
9
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
1
9
.
2
9
4
0
0
6
1
.
[
1
3
]
M
.
L
a
m
a
r
i
e
t
a
l
.
,
“
S
M
O
T
E
–
E
N
N
-
b
a
s
e
d
d
a
t
a
s
a
m
p
l
i
n
g
a
n
d
i
m
p
r
o
v
e
d
d
y
n
a
m
i
c
e
n
s
e
m
b
l
e
s
e
l
e
c
t
i
o
n
f
o
r
i
m
b
a
l
a
n
c
e
d
m
e
d
i
c
a
l
d
a
t
a
c
l
a
s
s
i
f
i
c
a
t
i
o
n
,
”
i
n
A
d
v
a
n
c
e
s
o
n
S
m
a
r
t
a
n
d
S
o
f
t
C
o
m
p
u
t
i
n
g
,
S
i
n
g
a
p
o
r
e
:
S
p
r
i
n
g
e
r
,
2
0
2
1
,
p
p
.
3
7
–
49
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
9
8
1
-
15
-
6
0
4
8
-
4
_
4
.
[
1
4
]
M
.
E.
L
o
k
a
n
a
n
,
“
P
r
e
d
i
c
t
i
n
g
m
o
b
i
l
e
m
o
n
e
y
t
r
a
n
sa
c
t
i
o
n
f
r
a
u
d
u
si
n
g
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
a
l
g
o
r
i
t
h
ms,”
Ap
p
l
i
e
d
AI
L
e
t
t
e
rs
,
v
o
l
.
4
,
n
o
.
2
,
A
p
r
.
2
0
2
3
,
d
o
i
:
1
0
.
1
0
0
2
/
a
i
l
2
.
8
5
.
[
1
5
]
Z.
X
u
,
D
.
S
h
e
n
,
T.
N
i
e
,
a
n
d
Y
.
K
o
u
,
“
A
h
y
b
r
i
d
sam
p
l
i
n
g
a
l
g
o
r
i
t
h
m
c
o
m
b
i
n
i
n
g
M
-
S
M
O
TE
a
n
d
EN
N
b
a
s
e
d
o
n
R
a
n
d
o
m
f
o
r
e
st
f
o
r
med
i
c
a
l
i
m
b
a
l
a
n
c
e
d
d
a
t
a
,
”
J
o
u
r
n
a
l
o
f
Bi
o
m
e
d
i
c
a
l
I
n
f
o
rm
a
t
i
c
s
,
v
o
l
.
1
0
7
,
J
u
l
.
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
j
b
i
.
2
0
2
0
.
1
0
3
4
6
5
.
[
1
6
]
F
.
Y
a
n
g
,
K
.
W
a
n
g
,
L.
S
u
n
,
M
.
Z
h
a
i
,
J.
S
o
n
g
,
a
n
d
H
.
W
a
n
g
,
“
A
h
y
b
r
i
d
s
a
mp
l
i
n
g
a
l
g
o
r
i
t
h
m
c
o
m
b
i
n
i
n
g
s
y
n
t
h
e
t
i
c
mi
n
o
r
i
t
y
o
v
e
r
-
samp
l
i
n
g
t
e
c
h
n
i
q
u
e
a
n
d
e
d
i
t
e
d
n
e
a
r
e
s
t
n
e
i
g
h
b
o
r
f
o
r
m
i
sse
d
a
b
o
r
t
i
o
n
d
i
a
g
n
o
si
s,”
B
MC
Me
d
i
c
a
l
I
n
f
o
rm
a
t
i
c
s
a
n
d
D
e
c
i
s
i
o
n
M
a
k
i
n
g
,
v
o
l
.
2
2
,
n
o
.
1
,
D
e
c
.
2
0
2
2
,
d
o
i
:
1
0
.
1
1
8
6
/
s
1
2
9
1
1
-
0
2
2
-
0
2
0
7
5
-
2.
[
1
7
]
M
.
I
san
g
e
d
i
o
k
a
n
d
K
.
G
a
j
a
ma
n
n
a
g
e
,
“
F
r
a
u
d
d
e
t
e
c
t
i
o
n
u
si
n
g
o
p
t
i
m
i
z
e
d
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
t
o
o
l
s
u
n
d
e
r
i
mb
a
l
a
n
c
e
c
l
a
ss
e
s,
”
i
n
2
0
2
2
I
EEE
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
B
i
g
D
a
t
a
(
B
i
g
D
a
t
a
)
,
D
e
c
.
2
0
2
2
,
p
p
.
4
2
7
5
–
4
2
8
4
,
d
o
i
:
1
0
.
1
1
0
9
/
B
i
g
D
a
t
a
5
5
6
6
0
.
2
0
2
2
.
1
0
0
2
0
7
2
3
.
[
1
8
]
I
.
D
.
M
i
e
n
y
e
a
n
d
Y
.
S
u
n
,
“
A
d
e
e
p
l
e
a
r
n
i
n
g
e
n
sem
b
l
e
w
i
t
h
d
a
t
a
r
e
s
a
m
p
l
i
n
g
f
o
r
c
r
e
d
i
t
c
a
r
d
f
r
a
u
d
d
e
t
e
c
t
i
o
n
,
”
I
EE
E
Ac
c
e
ss
,
v
o
l
.
1
1
,
p
p
.
3
0
6
2
8
–
3
0
6
3
8
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
3
.
3
2
6
2
0
2
0
.
[
1
9
]
N
.
L
.
F
i
t
r
i
y
a
n
i
,
M
.
S
y
a
f
r
u
d
i
n
,
G
.
A
l
f
i
a
n
,
a
n
d
J.
R
h
e
e
,
“
D
e
v
e
l
o
p
me
n
t
o
f
d
i
s
e
a
se
p
r
e
d
i
c
t
i
o
n
m
o
d
e
l
b
a
se
d
o
n
e
n
sem
b
l
e
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
f
o
r
d
i
a
b
e
t
e
s
a
n
d
h
y
p
e
r
t
e
n
s
i
o
n
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
7
,
p
p
.
1
4
4
7
7
7
–
1
4
4
7
8
9
,
2
0
1
9
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
1
9
.
2
9
4
5
1
2
9
.
[
2
0
]
L.
B
o
r
a
t
t
o
,
S
.
C
a
r
t
a
,
W
.
I
g
u
i
d
e
r
,
F
.
M
u
l
a
s
,
a
n
d
P
.
P
i
l
l
o
n
i
,
“
F
a
i
r
p
e
r
f
o
r
m
a
n
c
e
-
b
a
s
e
d
u
s
e
r
r
e
c
o
mm
e
n
d
a
t
i
o
n
i
n
e
C
o
a
c
h
i
n
g
s
y
s
t
e
ms
,
”
U
ser
M
o
d
e
l
i
n
g
a
n
d
U
ser
-
A
d
a
p
t
e
d
I
n
t
e
ra
c
t
i
o
n
,
v
o
l
.
3
2
,
n
o
.
5
,
p
p
.
8
3
9
–
8
8
1
,
N
o
v
.
2
0
2
2
,
d
o
i
:
1
0
.
1
0
0
7
/
s1
1
2
5
7
-
022
-
0
9
3
3
9
-
6.
[
2
1
]
R
.
M
o
h
a
mm
a
d
a
n
d
L.
M
c
C
l
u
s
k
e
y
,
“
P
h
i
s
h
i
n
g
w
e
b
si
t
e
s
,
”
U
C
I
M
a
c
h
i
n
e
L
e
a
r
n
i
n
g
Re
p
o
si
t
o
ry
,
2
0
1
2
,
d
o
i
:
1
0
.
2
4
4
3
2
/
C
5
1
W
2
X
.
[
2
2
]
T.
C
h
e
n
a
n
d
C
.
G
u
e
s
t
r
i
n
,
“
X
G
B
o
o
s
t
:
a
s
c
a
l
a
b
l
e
t
r
e
e
b
o
o
st
i
n
g
s
y
st
e
m,
”
i
n
Pro
c
e
e
d
i
n
g
s
o
f
t
h
e
2
2
n
d
A
C
M
S
I
G
K
D
D
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
K
n
o
w
l
e
d
g
e
D
i
s
c
o
v
e
r
y
a
n
d
D
a
t
a
Mi
n
i
n
g
,
A
u
g
.
2
0
1
6
,
p
p
.
7
8
5
–
794
,
d
o
i
:
1
0
.
1
1
4
5
/
2
9
3
9
6
7
2
.
2
9
3
9
7
8
5
.
Evaluation Warning : The document was created with Spire.PDF for Python.