I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
pu
t
er
Science
Vo
l.
4
3
,
No
.
3
,
Sep
tem
b
er
202
6
,
p
p
.
9
0
8
~
9
2
7
I
SS
N:
2
5
0
2
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/ijeecs.v
4
3
.i
3
.
pp
908
-
9
2
7
908
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
URL
-
ba
sed p
hish
ing
det
ec
tion usin
g
XG
Bo
o
st wi
th
e
ng
ineered
features
J
a
wa
her
Alha
rbi,
M
a
na
l Ba
y
o
us
ef
,
H
ind
Alm
is
ba
hi
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
m
a
t
i
o
n
Te
c
h
n
o
l
o
g
y
,
F
a
c
u
l
t
y
o
f
C
o
m
p
u
t
i
n
g
a
n
d
I
n
f
o
r
m
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
,
K
i
n
g
A
b
d
u
l
a
z
i
z
U
n
i
v
e
r
s
i
t
y
,
Jed
d
a
h
,
S
a
u
d
i
A
r
a
b
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ap
r
11
,
2
0
2
5
R
ev
is
ed
J
u
l
20
,
2
0
2
6
Acc
ep
ted
Au
g
19
,
2
0
2
6
URL
-
b
a
se
d
p
h
ish
i
n
g
i
n
v
o
l
v
e
s fak
e
u
n
if
o
rm
re
so
u
rc
e
lo
c
a
to
rs
(URL
s) cre
a
ted
b
y
a
tt
a
c
k
e
rs
to
tri
c
k
u
se
rs
i
n
to
b
e
li
e
v
in
g
t
h
e
y
a
re
v
isit
i
n
g
a
leg
it
ima
te
we
b
site
a
n
d
th
e
re
b
y
ste
a
l
th
e
ir
c
o
n
fid
e
n
ti
a
l
in
f
o
rm
a
ti
o
n
.
Wh
il
e
se
v
e
ra
l
p
o
we
rfu
l
m
a
c
h
i
n
e
lea
rn
in
g
(M
L
)
a
n
d
d
e
e
p
lea
rn
in
g
(DL)
stu
d
ie
s
e
x
ist
to
d
e
tec
t
p
h
is
h
in
g
,
t
h
e
y
st
il
l
fa
c
e
li
m
it
a
ti
o
n
s.
M
a
n
y
st
u
d
ies
re
ly
o
n
t
h
ird
-
p
a
rt
y
in
terv
e
n
t
io
n
to
e
x
trac
t
fe
a
tu
re
s,
wh
ich
i
n
tro
d
u
c
e
s
d
e
lay
s
t
h
a
t
m
a
k
e
th
e
m
u
n
su
i
tab
le
fo
r
fa
st
d
e
tec
ti
o
n
.
An
o
th
e
r
li
m
it
a
ti
o
n
is
t
h
a
t
e
x
isti
n
g
st
u
d
ies
o
ften
u
se
sm
a
ll
d
a
tas
e
ts,
a
n
d
trad
it
i
o
n
a
l
fe
a
tu
re
s
h
i
n
d
e
r
m
o
d
e
ls'
a
b
il
it
y
to
lea
rn
n
e
w
p
h
is
h
in
g
tec
h
n
iq
u
e
s,
re
su
lt
in
g
in
p
o
o
r
g
e
n
e
ra
li
z
a
ti
o
n
.
Th
e
re
fo
re
,
d
e
v
e
lo
p
in
g
n
e
w
fe
a
tu
re
s is
c
ru
c
ia
l
to
e
n
s
u
re
t
h
a
t
a
n
ti
-
p
h
ish
i
n
g
t
o
o
l
s
c
a
n
k
e
e
p
p
a
c
e
with
e
v
o
l
v
i
n
g
p
h
ish
i
n
g
tac
ti
c
s.
In
a
d
d
i
ti
o
n
,
th
e
e
x
isti
n
g
stu
d
i
e
s
d
o
n
o
t
re
p
o
rt
d
e
tec
ti
o
n
t
ime
,
wh
ich
is
imp
o
rtan
t
fo
r
fa
st
d
e
tec
ti
o
n
,
a
n
d
re
d
u
c
e
s
m
e
th
o
d
o
lo
g
ica
l
c
larity
.
T
h
is
p
a
p
e
r
a
ims
to
a
d
d
re
ss
t
h
e
se
li
m
it
a
ti
o
n
s
b
y
a
p
p
ly
in
g
a
n
e
u
ra
l
n
e
two
r
k
m
o
d
e
l
a
n
d
trad
it
i
o
n
a
l
M
L
c
l
a
ss
ifi
c
a
ti
o
n
a
lg
o
rit
h
m
s
to
s
u
p
p
o
rt
b
r
o
ws
e
r
-
b
a
se
d
p
h
ish
in
g
d
e
tec
ti
o
n
th
a
t
b
a
lan
c
e
s
h
ig
h
a
c
c
u
ra
c
y
with
fa
st
d
e
tec
ti
o
n
.
Ou
r
XG
Bo
o
st
m
o
d
e
l
a
c
h
iev
e
d
9
8
%
a
c
c
u
ra
c
y
o
n
th
e
tes
t
se
t,
u
ti
li
z
i
n
g
4
0
t
h
ir
d
-
p
a
rty
-
in
d
e
p
e
n
d
e
n
t
fe
a
tu
re
s.
Ad
d
it
io
n
a
ll
y
,
we
a
c
h
iev
e
d
a
n
a
v
e
ra
g
e
re
sp
o
n
se
ti
m
e
o
f
0
.
0
2
6
2
2
5
se
c
o
n
d
s
a
n
d
a
n
a
v
e
ra
g
e
c
o
m
p
u
tati
o
n
ti
m
e
o
f
1
.
6
9
7
7
×
1
0
⁻
6
se
c
o
n
d
s
p
e
r
URL
,
w
h
ich
d
e
m
o
n
stra
tes
c
o
m
p
e
ti
ti
v
e
s
p
e
e
d
.
We
p
ro
v
id
e
d
a
tab
le
o
f
fe
a
tu
re
s
fr
o
m
re
c
e
n
t
stu
d
ies
,
to
g
e
t
h
e
r
with
t
h
e
ir
d
o
c
u
m
e
n
ted
so
u
rc
e
s,
to
su
p
p
o
rt
fu
tu
re
re
se
a
rc
h
a
n
d
a
n
a
ly
z
e
k
e
y
URL
-
b
a
se
d
fe
a
tu
re
s.
K
ey
w
o
r
d
s
:
C
y
b
er
s
e
cu
r
ity
Ma
ch
in
e
lear
n
in
g
Ph
is
h
in
g
web
s
ite
d
etec
tio
n
UR
L
-
b
ased
f
ea
tu
r
es
XGBo
o
s
t m
o
d
el
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
:
J
awa
h
er
Alh
ar
b
i
Dep
ar
tm
en
t o
f
I
n
f
o
r
m
atio
n
T
e
ch
n
o
lo
g
y
,
Facu
lty
o
f
C
o
m
p
u
tin
g
an
d
I
n
f
o
r
m
atio
n
T
ec
h
n
o
lo
g
y
Kin
g
Ab
d
u
laziz
Un
iv
er
s
ity
4
3
4
0
0
Ser
d
an
g
,
Selan
g
o
r
,
Ma
l
ay
s
ia
E
m
ail:
jalh
ar
b
i0
0
8
6
@
s
tu
.
k
au
.
ed
u
.
s
a
1.
I
NT
RO
D
UCT
I
O
N
C
y
b
er
th
r
ea
ts
h
av
e
b
ec
o
m
e
m
o
r
e
s
o
p
h
is
ticated
an
d
co
m
p
le
x
,
m
ak
in
g
th
eir
d
etec
tio
n
a
n
d
an
aly
s
is
a
co
n
s
tan
t
ch
allen
g
e.
Ph
is
h
in
g
web
s
ites
ar
e
am
o
n
g
th
e
m
o
s
t
p
r
o
m
in
en
t
o
f
th
ese
th
r
ea
ts
[
1
]
.
I
n
2
0
2
2
,
s
ev
e
r
al
d
o
m
ain
s
wer
e
r
e
g
is
ter
ed
f
o
r
p
h
is
h
in
g
we
b
s
ites
im
p
er
s
o
n
atin
g
th
e
Ab
s
h
er
p
latf
o
r
m
,
a
n
elec
tr
o
n
ic
p
o
r
tal
af
f
iliated
with
th
e
Sau
d
i
Min
i
s
tr
y
o
f
I
n
ter
io
r
[
2
]
.
I
n
2
0
2
3
,
f
in
an
cial
p
h
is
h
in
g
attac
k
s
ac
co
u
n
ted
f
o
r
2
3
.
5
%
o
f
all
o
th
er
attac
k
s
[
3
]
,
wh
ile
in
2
0
2
4
,
c
r
ed
en
tial p
h
is
h
in
g
attac
k
s
in
cr
ea
s
ed
b
y
7
0
3
%
[
4
]
.
W
h
at
m
ak
es th
e
m
atter
ev
en
m
o
r
e
s
er
io
u
s
is
th
e
p
r
o
jecte
d
r
is
e
in
th
e
an
n
u
al
co
s
t
o
f
cy
b
e
r
cr
im
es
to
$
1
0
.
5
tr
i
llio
n
b
y
2
0
2
5
[
5
]
.
T
h
er
ef
o
r
e,
th
ese
s
tatis
tics
r
ef
l
ec
t
th
at
th
e
r
ap
id
a
n
d
ac
c
u
r
at
e
d
etec
tio
n
o
f
p
h
is
h
in
g
is
an
im
p
o
r
tan
t
r
esear
c
h
ch
allen
g
e.
I
n
r
ec
en
t
y
ea
r
s
,
m
a
n
y
s
tu
d
ies
h
av
e
r
elied
o
n
ML
class
if
ier
s
to
d
ev
elo
p
ef
f
ec
ti
v
e
m
ec
h
an
is
m
s
f
o
r
d
etec
tin
g
p
h
is
h
in
g
attac
k
s
[
1
]
.
Alth
o
u
g
h
t
h
ese
class
if
icati
o
n
s
h
av
e
ac
h
iev
e
d
h
ig
h
d
ete
ctio
n
p
er
f
o
r
m
a
n
ce
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
UR
L
-
b
a
s
ed
p
h
is
h
in
g
d
etec
tio
n
u
s
in
g
X
GB
o
o
s
t wi
th
en
g
in
ee
r
ed
fea
tu
r
es
(
Ja
w
a
h
er A
lh
a
r
b
i
)
909
cu
r
r
en
t
s
tu
d
ies
s
till
s
u
f
f
er
f
r
o
m
s
ev
er
al
lim
itatio
n
s
.
A
m
ajo
r
lim
itat
io
n
is
th
e
r
elian
ce
o
n
th
ir
d
-
p
ar
t
y
s
er
v
ices
d
u
r
in
g
f
ea
t
u
r
e
a
n
aly
s
is
an
d
ex
tr
ac
tio
n
,
wh
ic
h
lead
s
to
d
elay
s
in
d
etec
tio
n
[
6
]
.
W
ith
th
e
ev
o
lu
tio
n
o
f
p
h
is
h
in
g
attac
k
s
,
tr
ad
itio
n
al
UR
L
-
b
ase
d
f
ea
tu
r
es,
s
u
ch
as
UR
L
len
g
t
h
an
d
lex
ical
f
ea
tu
r
es,
ar
e
n
o
lo
n
g
er
s
u
f
f
i
cien
t
t
o
ac
h
iev
e
ef
f
ec
tiv
e
d
etec
tio
n
[
7
]
.
Ho
wev
e
r
,
m
o
s
t
s
tu
d
ies
s
till
r
ely
o
n
tr
ad
itio
n
al
f
ea
tu
r
e
s
,
wh
ile
f
ew
h
a
v
e
ad
d
r
ess
ed
ad
v
an
ce
d
UR
L
-
b
ased
f
ea
tu
r
es,
s
u
ch
as
i
n
ter
p
lan
etar
y
f
ile
s
y
s
tem
(
I
PF
S)
d
ec
e
n
tr
alize
d
p
h
is
h
in
g
h
o
s
tin
g
,
p
er
ce
n
t
-
en
co
d
in
g
,
an
d
Pu
n
y
co
d
e
.
I
n
ad
d
itio
n
,
p
r
e
v
io
u
s
s
tu
d
ies
o
f
ten
r
ely
o
n
a
lim
ited
an
d
o
u
td
ate
d
d
ataset.
T
h
er
ef
o
r
e,
th
e
r
esear
c
h
er
s
in
[
8
]
p
o
in
te
d
o
u
t
th
e
im
p
o
r
tan
ce
o
f
tr
ai
n
in
g
ML
class
i
f
ier
s
o
n
a
lar
g
e
an
d
up
-
to
-
d
ate
d
ataset
th
at
in
cl
u
d
es
ad
v
an
ce
d
f
ea
tu
r
es
to
co
u
n
t
er
co
m
p
le
x
p
h
is
h
in
g
attac
k
s
.
T
h
is
s
tu
d
y
aim
s
to
ad
d
r
ess
th
ese
lim
itatio
n
s
b
y
i
n
tr
o
d
u
ci
n
g
a
UR
L
-
b
ased
d
etec
tio
n
s
y
s
tem
u
s
in
g
ML
class
if
ier
s
f
o
r
p
h
is
h
in
g
d
etec
tio
n
,
b
ala
n
cin
g
s
p
ee
d
an
d
d
etec
tio
n
ac
cu
r
ac
y
w
ith
o
u
t
r
ely
in
g
o
n
th
ir
d
-
p
ar
ty
s
er
v
ic
es.
Ad
d
itio
n
ally
,
it
p
r
esen
ts
a
d
ataset
th
at
in
clu
d
es
b
o
th
tr
ad
itio
n
al
an
d
ad
v
an
ce
d
f
e
atu
r
es.
Fu
r
t
h
er
m
o
r
e,
we
ev
alu
ate
d
th
e
ef
f
ec
tiv
en
ess
o
f
th
e
s
elec
ted
f
ea
tu
r
es
u
s
in
g
s
ix
tr
ad
itio
n
al
ML
alg
o
r
ith
m
s
:
d
ec
is
io
n
tr
ee
,
k
-
n
ea
r
est
n
eig
h
b
o
r
s
(
KNN)
,
r
an
d
o
m
f
o
r
est
(
R
F),
ad
ap
tiv
e
b
o
o
s
tin
g
(
A
d
aBo
o
s
t)
,
ex
tr
em
e
g
r
ad
ien
t
b
o
o
s
tin
g
(
XGBo
o
s
t)
,
lig
h
t
g
r
ad
ien
t
b
o
o
s
tin
g
m
ac
h
in
e
(
L
i
g
h
tGB
M)
,
an
d
a
n
eu
r
al
n
etwo
r
k
m
o
d
el,
th
e
m
u
lti
-
lay
e
r
p
er
c
ep
tr
o
n
(
ML
P).
Of
th
ese
m
o
d
els,
XGBo
o
s
t
ac
h
iev
ed
ef
f
ec
ti
v
e
r
esu
lts
;
th
er
ef
o
r
e,
it
h
as
b
ee
n
s
elec
ted
as
th
e
b
ase
m
o
d
el
f
o
r
o
u
r
r
esear
ch
.
T
h
e
m
ain
co
n
tr
ib
u
tio
n
s
in
th
is
r
esear
ch
ar
e
as f
o
llo
ws:
−
W
e
cr
ea
ted
a
d
ataset
co
m
p
ile
d
f
r
o
m
r
ec
en
t,
p
u
b
licly
av
aila
b
le
s
o
u
r
ce
s
th
at
in
clu
d
es
UR
L
-
b
ased
f
ea
tu
r
es
in
d
ep
en
d
en
t
o
f
th
ir
d
-
p
ar
ty
s
er
v
ices.
T
h
is
d
ataset
c
o
n
s
is
ts
o
f
4
0
lex
ical
an
d
s
tr
u
ctu
r
al
f
ea
t
u
r
es,
in
clu
d
in
g
th
r
ee
f
ea
tu
r
es
th
at
ar
e
n
o
t
c
o
m
m
o
n
l
y
u
s
ed
in
p
r
e
v
io
u
s
s
tu
d
ies
b
u
t
wer
e
id
en
tifie
d
as
im
p
o
r
tan
t
an
d
ef
f
ec
tiv
e
in
o
u
r
m
o
d
el.
−
Su
p
p
lem
en
tar
y
1
,
T
a
b
le
1
s
u
m
m
ar
izes
th
e
7
3
f
ea
tu
r
es
with
th
eir
r
ef
er
en
ce
s
f
r
o
m
r
ec
en
t
s
tu
d
ies
an
d
d
escr
ib
es
th
e
c
o
r
r
esp
o
n
d
in
g
f
ea
tu
r
e
ex
t
r
ac
tio
n
m
eth
o
d
s
.
T
h
is
p
r
o
v
id
es
a
u
s
ef
u
l
c
o
n
tr
ib
u
tio
n
to
p
h
is
h
in
g
d
etec
tio
n
r
esear
ch
,
as it su
p
p
o
r
ts
r
esear
ch
er
s
in
id
en
tify
i
n
g
a
n
d
ap
p
l
y
in
g
f
ea
tu
r
e
in
f
o
r
m
atio
n
.
−
Ou
r
XGBo
o
s
t
m
o
d
el
ac
h
iev
e
d
a
b
alan
ce
b
etwe
en
h
ig
h
ac
cu
r
ac
y
an
d
s
p
ee
d
.
I
t
ac
h
iev
ed
9
8
%
ac
cu
r
ac
y
o
n
th
e
test
s
e
t
with
an
av
er
ag
e
co
m
p
u
tatio
n
tim
e
o
f
1
.
6
9
7
7
×1
0
⁻
6
s
ec
o
n
d
s
p
er
UR
L
.
Fu
r
th
er
m
o
r
e,
th
e
m
o
d
el
r
ec
o
r
d
e
d
an
a
v
er
ag
e
r
esp
o
n
s
e
tim
e
o
f
0
.
0
2
6
2
2
5
s
ec
o
n
d
s
p
e
r
UR
L
in
th
e
o
n
lin
e
s
ettin
g
.
T
h
e
p
ap
er
is
o
r
g
a
n
ized
as
f
o
llo
ws:
Sectio
n
2
r
ev
iews
th
e
r
elev
an
t
liter
atu
r
e.
Sectio
n
3
ex
p
lain
s
th
e
r
esear
ch
m
eth
o
d
o
lo
g
y
,
d
etailin
g
th
e
d
ataset,
f
ea
tu
r
e
e
x
tr
ac
tio
n
an
d
s
elec
tio
n
,
ML
al
g
o
r
ith
m
s
u
s
ed
,
an
d
ev
alu
atio
n
c
r
iter
ia.
T
h
e
n
,
Se
ctio
n
4
p
r
esen
ts
th
e
e
x
p
er
im
en
tal
r
esu
lts
.
Sectio
n
5
an
aly
s
es
th
e
r
esu
lts
an
d
ev
alu
ates
th
e
p
er
f
o
r
m
a
n
ce
.
A
d
d
itio
n
ally
,
t
h
is
s
ec
tio
n
co
m
p
ar
es
th
e
r
esu
lts
with
th
e
ex
is
ti
n
g
liter
atu
r
e.
Af
ter
th
at,
th
e
r
esu
lts
ar
e
d
is
cu
s
s
ed
,
an
d
f
u
tu
r
e
r
ec
o
m
m
en
d
atio
n
s
a
r
e
p
r
esen
ted
in
Sectio
n
6
.
Fin
ally
,
Sectio
n
7
co
n
clu
d
es th
e
p
ap
er
a
n
d
p
r
esen
ts
f
u
tu
r
e
wo
r
k
.
2.
RE
L
AT
E
D
WO
RK
Pre
v
io
u
s
s
tu
d
ies h
av
e
ex
p
lo
r
e
d
p
h
is
h
in
g
web
s
ites
ex
ten
s
iv
ely
u
s
in
g
tr
ad
itio
n
al
ML
an
d
D
L
m
eth
o
d
s
.
Ko
cy
ig
it
et
a
l.
[
9
]
f
o
cu
s
ed
o
n
d
ev
elo
p
in
g
a
d
etec
tio
n
s
y
s
tem
b
y
s
elec
tin
g
o
p
tim
al
f
ea
tu
r
es
u
s
in
g
o
p
tim
ize
d
g
en
etic
alg
o
r
ith
m
f
ea
t
u
r
e
s
ele
ctio
n
(
GAFS).
T
h
e
r
esear
ch
er
s
u
s
ed
a
d
ataset
f
r
o
m
a
s
tu
d
y
in
[
1
0
]
,
c
o
n
tain
in
g
3
6
,
1
7
3
p
h
is
h
in
g
an
d
5
1
,
3
1
6
leg
itima
te
web
s
ites
.
Af
ter
ap
p
ly
in
g
GAFS,
th
ey
o
b
tain
ed
4
4
o
p
tim
al
UR
L
-
b
ased
f
ea
tu
r
es.
T
h
e
r
esu
lts
s
h
o
wed
th
at
th
e
R
F
m
o
d
el
ac
h
iev
e
d
an
ac
cu
r
ac
y
o
f
9
2
.
9
3
%
o
n
th
e
test
s
et,
with
a
co
m
p
u
tatio
n
tim
e
o
f
0
.
6
3
8
0
s
e
co
n
d
s
.
Ho
wev
er
,
th
e
r
esear
c
h
e
r
s
r
ep
o
r
ted
th
at
s
o
m
e
o
f
th
e
s
e
lecte
d
f
ea
tu
r
es
ar
e
no
t
s
u
itab
le
f
o
r
r
ea
l
-
tim
e
d
ep
l
o
y
m
en
t
d
u
e
to
th
e
h
ig
h
p
r
o
ce
s
s
in
g
co
s
t.
I
n
co
n
tr
ast,
Gu
p
ta
e
t
a
l.
[
1
1
]
p
r
o
p
o
s
ed
a
lig
h
tweig
h
t
p
h
is
h
in
g
d
etec
t
io
n
s
y
s
tem
co
m
p
atib
le
with
in
ter
n
et
o
f
th
in
g
s
(
I
o
T
)
d
ev
ic
es.
T
h
eir
ap
p
r
o
ac
h
r
elies
o
n
th
e
lex
ical
f
ea
t
u
r
es
o
f
UR
L
s
wit
h
o
u
t
r
ely
in
g
o
n
th
i
r
d
-
p
ar
t
y
s
er
v
ices.
T
h
e
r
esear
ch
er
s
u
s
ed
a
d
ataset
f
r
o
m
I
SC
XURL
2
0
1
6
p
r
o
v
id
ed
b
y
th
e
Un
iv
e
r
s
ity
o
f
Ne
w
B
r
u
n
s
wick
(
UNB),
wh
ic
h
co
n
s
is
ts
o
f
1
0
,
0
0
0
leg
itima
te
UR
L
s
an
d
9
,
9
6
4
p
h
is
h
in
g
UR
L
s
,
with
o
n
ly
9
lex
ical
f
ea
tu
r
es
ex
tr
ac
ted
.
T
h
e
ex
p
er
im
e
n
tatio
n
r
esu
lts
in
d
icate
th
at
th
e
R
F
m
o
d
el
ac
h
iev
ed
9
9
.
5
7
%
o
n
th
e
t
est
s
et,
wi
th
a
r
esp
o
n
s
e
tim
e
o
f
5
1
.
5
m
illi
s
ec
o
n
d
s
(
0
.
0
5
1
5
s
ec
o
n
d
s
)
.
T
h
e
r
esear
ch
er
s
in
d
icate
d
t
h
at
f
u
t
u
r
e
wo
r
k
wo
u
ld
in
v
o
lv
e
e
x
p
an
d
i
n
g
th
e
wo
r
k
u
s
in
g
ad
v
an
ce
d
DL
alg
o
r
ith
m
s
.
T
h
i
s
f
u
tu
r
e
d
i
r
ec
tio
n
r
ef
lects
th
e
co
n
tin
u
in
g
in
ter
est
in
e
x
p
lo
r
in
g
m
o
r
e
a
d
v
an
ce
d
lear
n
in
g
ap
p
r
o
ac
h
es f
o
r
p
h
is
h
i
n
g
d
etec
tio
n
.
Similar
ly
,
Ah
am
m
ad
et
a
l.
[
1
2
]
ev
alu
ate
d
s
ev
er
al
ML
clas
s
if
ier
s
an
d
tr
ain
ed
t
h
em
o
n
U
R
L
-
b
ased
f
ea
tu
r
es
to
i
d
en
tify
p
h
is
h
in
g
UR
L
s
.
T
h
ey
ex
tr
ac
ted
1
5
f
ea
t
u
r
es
f
r
o
m
a
d
ataset
o
f
1
,
5
0
0
p
h
is
h
in
g
UR
L
s
f
r
o
m
Ph
is
h
T
an
k
an
d
1
,
5
0
0
leg
itima
te
UR
L
s
f
r
o
m
UNB.
Af
ter
s
ev
er
al
ex
p
er
im
en
ts
,
th
e
L
ig
h
tGB
M
m
o
d
el
ac
h
iev
ed
an
ac
cu
r
a
cy
o
f
8
6
%.
T
h
e
r
esear
ch
er
s
n
o
ted
lim
itatio
n
s
i
n
th
e
s
tu
d
y
,
n
am
ely
th
at
s
o
m
e
U
R
L
s
d
o
n
o
t
co
n
tain
s
u
f
f
icien
t
in
f
o
r
m
ati
o
n
in
W
HOI
S
d
atab
ases
,
r
esu
ltin
g
in
th
e
in
ab
ilit
y
to
ex
tr
ac
t
all
th
e
r
eq
u
ir
ed
f
ea
tu
r
es.
T
h
e
y
r
ep
o
r
ted
th
e
n
ee
d
to
g
at
h
er
n
e
wer
UR
L
-
b
ased
f
ea
tu
r
es to
im
p
r
o
v
e
ac
c
u
r
ac
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
4
3
,
No
.
3
,
Sep
tem
b
er
20
2
6
:
9
0
8
-
9
2
7
910
T
ab
le
1
.
Su
m
m
a
r
y
o
f
r
elate
d
wo
r
k
R
e
f
.
M
o
d
e
l
M
e
t
h
o
d
D
a
t
a
s
e
t
s
i
z
e
N
u
mb
e
r
o
f
f
e
a
t
u
r
e
s
A
c
c
u
r
a
c
y
D
e
t
e
c
t
i
o
n
t
i
m
e
(
sec
o
n
d
s)
D
e
p
e
n
d
e
n
c
e
o
n
t
h
i
r
d
p
a
r
t
i
e
s
Li
mi
t
a
t
i
o
n
s
[
9
]
,
2
0
2
4
RF
U
R
L
-
b
a
s
e
d
3
6
,
1
7
3
p
h
i
s
h
i
n
g
U
R
Ls
a
n
d
5
1
,
3
1
6
l
e
g
i
t
i
m
a
t
e
U
R
Ls
.
44
9
2
.
9
3
%
0
.
6
3
8
0
S
o
me
sel
e
c
t
e
d
f
e
a
t
u
r
e
s
a
r
e
n
o
t
su
i
t
a
b
l
e
f
o
r
r
e
a
l
-
t
i
me
d
e
p
l
o
y
me
n
t
d
u
e
t
o
t
h
e
h
i
g
h
p
r
o
c
e
ssi
n
g
c
o
s
t
.
N
o
t
f
u
l
l
y
r
e
p
o
r
t
e
d
e
x
t
r
a
c
t
e
d
f
e
a
t
u
r
e
.
[
1
1
]
,
2
0
2
1
RF
U
R
L
-
b
a
s
e
d
9
,
9
6
4
p
h
i
s
h
i
n
g
U
R
Ls
a
n
d
1
0
,
0
0
0
l
e
g
i
t
i
m
a
t
e
U
R
Ls
.
9
9
9
.
5
7
%
0
.
0
5
1
5
Th
e
d
a
t
a
se
t
i
s
o
l
d
,
l
i
m
i
t
e
d
,
a
n
d
c
o
n
t
a
i
n
s
a
sma
l
l
se
t
o
f
t
r
a
d
i
t
i
o
n
a
l
f
e
a
t
u
r
e
s
.
[
1
2
]
,
2
0
2
2
Li
g
h
t
G
B
M
U
R
L
-
b
a
s
e
d
1
,
5
0
0
p
h
i
s
h
i
n
g
U
R
Ls
a
n
d
1
,
5
0
0
l
e
g
i
t
i
m
a
t
e
U
R
Ls
.
15
8
6
%
N
o
t
men
t
i
o
n
e
d
✓
R
e
l
i
a
n
c
e
o
n
t
h
i
r
d
-
p
a
r
t
y
s
e
r
v
i
c
e
s.
U
n
sp
e
c
i
f
i
e
d
d
e
t
e
c
t
i
o
n
t
i
m
e
.
L
o
w
a
c
c
u
r
a
c
y
r
a
t
e
.
[
1
3
]
,
2
0
2
0
RF
U
R
L
-
b
a
s
e
d
a
n
d
c
o
n
t
e
n
t
-
b
a
s
e
d
6
,
1
5
7
p
h
i
s
h
i
n
g
U
R
Ls
a
n
d
4
,
8
9
8
l
e
g
i
t
i
m
a
t
e
U
R
Ls
.
30
9
9
.
3
6
%
4
✓
Th
e
st
u
d
y
d
e
p
e
n
d
s
o
n
t
h
i
r
d
-
p
a
r
t
y
d
a
t
a
,
d
e
l
a
y
s i
n
d
e
t
e
c
t
i
o
n
,
a
n
d
a
sma
l
l
d
a
t
a
se
t
.
[
1
4
]
,
2
0
2
3
LSD
mo
d
e
l
w
i
t
h
C
a
n
o
p
y
f
e
a
t
u
r
e
sel
e
c
t
i
o
n
U
R
L
-
b
a
s
e
d
5
,
0
0
0
p
h
i
s
h
i
n
g
U
R
Ls
a
n
d
6
,
0
0
0
l
e
g
i
t
i
m
a
t
e
U
R
Ls
.
33
9
8
.
1
2
%
3
7
0
.
9
8
4
5
N
o
t
men
t
i
o
n
e
d
N
o
t
f
u
l
l
y
r
e
p
o
r
t
e
d
e
x
t
r
a
c
t
e
d
f
e
a
t
u
r
e
.
D
e
t
e
c
t
i
o
n
t
i
m
e
r
e
p
o
r
t
e
d
w
i
t
h
o
u
t
sp
e
c
i
f
y
i
n
g
t
h
e
mea
s
u
r
e
me
n
t
u
n
i
t
.
[
1
5
]
,
2
0
2
1
GA
–
G
r
a
d
i
e
n
t
B
o
o
s
t
,
G
A
–
B
a
g
g
i
n
g
,
a
n
d
G
A
–
X
G
B
o
o
st
U
R
L
-
b
a
s
e
d
a
n
d
c
o
n
t
e
n
t
-
b
a
s
e
d
D
a
t
a
s
e
t
1
:
4
,
8
9
8
p
h
i
s
h
i
n
g
U
R
Ls
,
6
,
1
5
7
l
e
g
i
t
i
m
a
t
e
U
R
Ls.
D
a
t
a
set
2
:
5
,
0
0
0
p
h
i
s
h
i
n
g
U
R
Ls,
5
,
0
0
0
l
e
g
i
t
i
m
a
t
e
U
R
Ls
,
a
n
d
D
a
t
a
s
e
t
3
:
3
0
,
6
4
7
p
h
i
s
h
i
n
g
U
R
Ls,
5
8
,
0
0
0
l
e
g
i
t
i
m
a
t
e
U
R
Ls
.
D
a
t
a
s
e
t
1
:
3
0
,
d
a
t
a
se
t
2
:
4
8
d
a
t
a
se
t
3
:
1
1
1
D
a
t
a
s
e
t
1
:
9
7
.
1
6
%
,
d
a
t
a
se
t
2
:
9
8
.
5
8
%
,
d
a
t
a
se
t
3
:
9
7
.
3
9
%
N
o
t
men
t
i
o
n
e
d
✓
Th
e
sy
s
t
e
m
s
u
f
f
e
r
s
f
r
o
m h
i
g
h
c
o
m
p
u
t
a
t
i
o
n
a
l
c
o
s
t
s,
d
e
p
e
n
d
e
n
c
e
o
n
t
h
i
r
d
-
p
a
r
t
y
s
e
r
v
i
c
e
s,
a
n
d
u
n
s
p
e
c
i
f
i
e
d
d
e
t
e
c
t
i
o
n
t
i
m
e
.
[
7
]
,
2
0
2
1
RF
U
R
L
-
b
a
s
e
d
Le
g
i
t
i
ma
t
e
:
3
5
,
3
7
8
,
p
h
i
s
h
i
n
g
:
9
,
9
6
5
,
mal
w
a
r
e
:
1
1
,
5
6
6
,
a
n
d
sp
a
m
:
1
1
,
9
4
2
.
20
9
9
%
4
.
7
0
9
Th
e
d
a
t
a
se
t
i
s
o
u
t
d
a
t
e
d
.
L
o
n
g
d
e
t
e
c
t
i
o
n
t
i
me
.
[
1
6
]
,
2
0
2
1
R
N
N
-
G
R
U
U
R
L
-
b
a
s
e
d
6
0
,
0
0
0
p
h
i
s
h
i
n
g
U
R
Ls
a
n
d
6
0
,
0
0
0
l
e
g
i
t
i
m
a
t
e
U
R
Ls
.
1
0
0
9
9
.
1
8
%
0
.
2
Li
mi
t
a
t
i
o
n
s i
n
p
r
o
c
e
ss
i
n
g
U
R
Ls
o
v
e
r
2
0
0
c
h
a
r
a
c
t
e
r
s
a
n
d
a
l
a
c
k
o
f
U
R
L
sh
o
r
t
e
n
i
n
g
s
u
p
p
o
r
t
.
Lo
n
g
d
e
t
e
c
t
i
o
n
t
i
me.
[
1
7
]
,
2
0
2
2
l
o
g
i
s
t
i
c
r
e
g
r
e
ss
i
o
n
U
R
L
-
b
a
s
e
d
3
0
,
0
0
0
p
h
i
s
h
i
n
g
U
R
Ls
a
n
d
3
0
,
0
0
0
l
e
g
i
t
i
m
a
t
e
U
R
Ls
.
38
9
6
.
5
0
%
N
o
t
men
t
i
o
n
e
d
S
l
i
g
h
t
d
e
c
r
e
a
s
e
i
n
o
v
e
r
a
l
l
a
c
c
u
r
a
c
y
w
h
e
n
u
s
i
n
g
l
o
g
i
n
U
R
Ls.
O
u
r
st
u
d
y
X
G
B
o
o
st
U
R
L
-
b
a
s
e
d
5
6
,
3
6
9
p
h
i
s
h
i
n
g
U
R
Ls
a
n
d
3
9
,
3
6
7
l
e
g
i
t
i
m
a
t
e
U
R
Ls
.
40
9
8
%
0
.
0
2
6
2
2
5
Th
e
st
u
d
y
r
e
l
i
e
s
o
n
l
y
o
n
U
R
L
f
e
a
t
u
r
e
s,
l
i
m
i
t
i
n
g
i
t
s
e
f
f
e
c
t
i
v
e
n
e
ss
a
g
a
i
n
s
t
c
o
n
t
e
n
t
-
b
a
se
d
p
h
i
s
h
i
n
g
a
t
t
a
c
k
s.
HR
et
a
l.
[
1
3
]
in
tr
o
d
u
ce
d
a
n
ew
b
r
o
wser
ar
c
h
itectu
r
e
c
alled
th
e
em
b
ed
d
ed
p
h
is
h
in
g
d
etec
tio
n
b
r
o
wser
(
E
PDB
)
u
s
in
g
an
R
F
m
o
d
el.
T
h
e
m
o
d
el
was
tr
ain
ed
o
n
a
d
ataset
o
f
6
,
1
5
7
p
h
is
h
in
g
UR
L
s
f
r
o
m
Ph
is
h
T
an
k
an
d
4
,
8
9
8
leg
itima
te
UR
L
s
f
r
o
m
Miller
s
'
Mile
s
.
T
h
e
E
PDB
s
y
s
tem
u
s
es
a
r
u
le
-
b
ased
ex
tr
ac
tio
n
f
r
am
ewo
r
k
to
id
e
n
tify
3
0
web
s
ite
f
ea
tu
r
es
th
at
co
m
b
in
e
UR
L
-
b
ased
an
d
co
n
te
n
t
-
b
ased
f
e
atu
r
es,
u
s
in
g
th
ir
d
-
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
UR
L
-
b
a
s
ed
p
h
is
h
in
g
d
etec
tio
n
u
s
in
g
X
GB
o
o
s
t wi
th
en
g
in
ee
r
ed
fea
tu
r
es
(
Ja
w
a
h
er A
lh
a
r
b
i
)
911
p
ar
ty
s
er
v
ices.
R
esear
ch
er
s
h
av
e
co
n
f
ir
m
ed
th
at
th
e
p
r
o
p
o
s
ed
s
y
s
tem
d
etec
ts
n
ewly
cr
e
ated
p
h
is
h
in
g
p
ag
es
with
9
9
.
3
6
%
ac
cu
r
ac
y
,
with
a
n
av
er
ag
e
r
esp
o
n
s
e
tim
e
o
f
4
s
ec
o
n
d
s
.
Ho
wev
er
,
th
e
a
u
th
o
r
s
p
lan
to
cr
ea
te
an
u
p
d
ated
d
ataset
b
y
co
llectin
g
p
h
is
h
in
g
UR
L
s
f
r
o
m
u
s
er
s
to
tr
ain
th
e
m
o
d
els
o
n
th
e
latest
p
h
is
h
in
g
s
ites
.
Ma
in
tain
in
g
u
p
-
to
-
d
ate
p
h
is
h
in
g
d
atasets
is
ess
en
tia
l
f
o
r
im
p
r
o
v
in
g
th
e
r
o
b
u
s
tn
ess
o
f
p
h
is
h
in
g
d
etec
tio
n
m
o
d
els.
Kar
im
et
a
l.
[
1
4
]
p
r
o
p
o
s
ed
a
h
y
b
r
id
m
o
d
el
b
ased
o
n
th
r
ee
m
ain
class
if
ier
s
:
L
o
g
is
tic
r
eg
r
ess
io
n
,
s
u
p
p
o
r
t
v
ec
to
r
class
if
ier
,
an
d
d
ec
is
io
n
tr
ee
(
L
R
+SVC+
D
T
)
,
with
s
o
f
t
v
o
tin
g
an
d
h
ar
d
v
o
tin
g
to
d
etec
t
p
h
is
h
in
g
web
s
ites
.
T
h
ey
also
u
s
ed
th
e
C
an
o
p
y
tech
n
iq
u
e
f
o
r
f
ea
tu
r
e
s
elec
tio
n
to
im
p
r
o
v
e
p
er
f
o
r
m
a
n
ce
.
T
o
tr
ain
th
e
m
o
d
els,
th
e
y
u
s
ed
a
d
ataset
co
llected
f
r
o
m
th
e
Ka
g
g
le
p
latf
o
r
m
,
co
m
p
r
is
in
g
5
,
0
0
0
p
h
is
h
in
g
UR
L
s
an
d
6
,
0
0
0
leg
itima
te
UR
L
s
,
with
3
3
UR
L
-
b
ased
f
ea
tu
r
es.
T
h
e
r
esu
lts
s
h
o
wed
th
at
th
e
L
SD
m
o
d
el
with
C
an
o
p
y
f
ea
t
u
r
e
s
elec
tio
n
ac
h
i
ev
ed
an
ac
cu
r
ac
y
o
f
9
8
.
1
2
%
o
n
th
e
test
s
et.
I
n
ad
d
itio
n
,
th
e
r
esear
ch
er
s
r
ep
o
r
ted
an
ex
ec
u
tio
n
tim
e
o
f
3
7
0
.
9
8
4
5
,
wh
ich
m
ea
s
u
r
es
th
e
tim
e
r
eq
u
ir
ed
f
o
r
th
eir
m
o
d
el
to
co
m
p
lete
a
d
etec
tio
n
task
.
Ho
wev
er
,
th
ey
d
id
n
o
t
s
p
ec
if
y
th
e
ex
ac
t
tim
e
u
n
it,
lea
v
in
g
th
e
ex
ac
t
m
ea
s
u
r
em
en
t
u
n
clea
r
.
T
h
e
a
u
th
o
r
s
s
u
g
g
ested
in
teg
r
atin
g
b
lack
li
s
t
an
d
wh
itelis
t
s
y
s
tem
s
with
ML
s
y
s
tem
s
in
th
e
f
u
t
u
r
e
,
with
th
e
aim
o
f
im
p
r
o
v
in
g
th
e
ef
f
icien
cy
o
f
d
etec
tin
g
p
h
is
h
in
g
web
s
ites
.
T
h
is
f
u
tu
r
e
d
ir
ec
tio
n
r
e
f
lects
th
e
co
n
tin
u
in
g
i
n
ter
est
in
im
p
r
o
v
in
g
p
h
is
h
in
g
d
etec
tio
n
ef
f
icien
c
y
.
Al
-
Sar
em
et
a
l.
[
1
5
]
p
r
o
p
o
s
ed
th
e
o
p
tim
ized
s
tack
in
g
e
n
s
em
b
le
m
o
d
el
u
s
in
g
a
g
en
etic
alg
o
r
ith
m
(
GA)
f
o
r
d
ev
elo
p
in
g
p
h
is
h
in
g
web
s
ite
d
etec
tio
n
s
y
s
tem
s
.
T
h
e
r
esear
ch
er
s
u
s
ed
th
r
ee
p
u
b
li
c
d
atasets
:
th
e
f
ir
s
t,
f
r
o
m
th
e
Un
iv
er
s
ity
o
f
C
alif
o
r
n
ia,
I
r
v
in
e
(
UC
I
)
,
co
n
tain
e
d
6
,
1
5
7
leg
itima
te
UR
L
s
an
d
4
,
8
9
8
p
h
is
h
in
g
UR
L
s
,
with
3
0
f
ea
tu
r
es.
T
h
e
s
ec
o
n
d
,
f
r
o
m
Me
n
d
eley
d
ata
,
co
n
tain
ed
5
,
0
0
0
leg
itima
te
UR
L
s
an
d
5
,
0
0
0
p
h
is
h
in
g
UR
L
s
,
wi
th
4
8
f
ea
tu
r
es.
T
h
e
t
h
ir
d
d
ataset,
also
f
r
o
m
Me
n
d
e
ley
d
ata
,
co
n
tain
ed
5
8
,
0
0
0
leg
itima
te
UR
L
s
an
d
3
0
,
6
4
7
p
h
is
h
in
g
UR
L
s
,
with
1
1
1
f
ea
tu
r
es.
All
d
atasets
co
n
s
is
t
o
f
UR
L
-
b
ased
an
d
c
o
n
te
n
t
-
ba
s
ed
attr
ib
u
tes.
T
h
e
r
esear
ch
er
s
r
elied
o
n
th
i
r
d
-
p
ar
t
y
s
er
v
ices,
th
e
W
HOI
S
d
atab
ase,
to
ex
tr
ac
t
s
o
m
e
f
ea
tu
r
e
in
f
o
r
m
atio
n
,
s
u
ch
as
d
o
m
ain
n
am
e
s
y
s
tem
(
DNS)
,
in
th
e
d
etec
tio
n
p
r
o
ce
s
s
.
B
ased
o
n
ex
p
er
i
m
en
tal
r
esu
lts
,
GA
-
Gr
ad
ien
tB
o
o
s
t,
GA
-
B
ag
g
in
g
,
an
d
GA
-
XGBo
o
s
t
wer
e
s
elec
ted
as
th
e
to
p
th
r
ee
alg
o
r
ith
m
s
as
b
ase
lear
n
er
s
,
an
d
th
e
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
SVM)
m
o
d
el
was
ad
o
p
te
d
as
an
ad
d
itio
n
al
class
if
ier
f
o
r
th
e
s
y
s
tem
.
T
h
e
p
r
o
p
o
s
ed
m
o
d
els
ac
h
iev
ed
t
h
e
f
o
llo
win
g
r
esu
lts
:
9
7
.
1
6
%,
9
8
.
5
8
%,
an
d
9
7
.
3
9
%
f
o
r
th
e
d
at
asets
,
r
esp
ec
tiv
ely
.
T
h
e
r
esear
ch
e
r
s
n
o
ted
th
at
t
h
eir
f
u
tu
r
e
wo
r
k
f
o
cu
s
es
o
n
cr
ea
tin
g
a
lig
h
tweig
h
t
d
etec
tio
n
s
y
s
tem
s
u
itab
le
f
o
r
I
o
T
en
v
ir
o
n
m
en
ts
to
p
r
o
tect
a
g
ain
s
t p
h
is
h
in
g
attac
k
s
.
T
h
is
o
b
s
er
v
atio
n
h
ig
h
lig
h
ts
th
e
im
p
o
r
tan
ce
o
f
ac
h
iev
in
g
b
o
th
co
m
p
u
tatio
n
al
e
f
f
icien
cy
an
d
r
ap
i
d
p
h
is
h
in
g
d
etec
tio
n
in
r
eso
u
r
ce
-
co
n
s
tr
ain
ed
en
v
ir
o
n
m
en
ts
.
T
h
er
ef
o
r
e,
R
aja
et
a
l.
[
7
]
p
r
o
v
id
ed
a
lig
h
tweig
h
t
s
y
s
tem
t
o
d
etec
t
p
h
is
h
in
g
web
s
ites
.
T
h
e
au
t
h
o
r
s
u
s
ed
a
d
ataset
f
r
o
m
th
e
UNB
d
ataset
2
0
1
6
co
n
tain
in
g
f
o
u
r
class
es
:
3
5
,
3
7
8
leg
itima
te
UR
L
s
,
9
,
9
6
5
p
h
is
h
in
g
UR
L
s
,
1
1
,
5
6
6
m
alwa
r
e
UR
L
s
,
an
d
1
1
,
9
4
2
s
p
am
UR
L
s
.
2
0
UR
L
-
b
ased
f
ea
t
u
r
es
in
d
e
p
e
n
d
en
t
o
f
t
h
ir
d
-
p
ar
ty
s
er
v
ices
wer
e
ex
tr
ac
ted
.
T
h
e
ex
p
er
im
en
tal
r
esu
lts
s
h
o
wed
th
at
th
e
R
F
m
o
d
el
ac
h
iev
ed
an
ac
cu
r
ac
y
o
f
9
9
%
with
a
co
m
p
u
tatio
n
tim
e
o
f
4
.
7
0
9
s
ec
o
n
d
s
.
T
h
e
r
esear
ch
er
s
p
o
in
ted
to
th
e
n
ee
d
to
ex
tr
ac
t
n
ew
f
ea
tu
r
es
an
d
ex
p
an
d
t
h
e
d
ataset
to
s
u
p
p
o
r
t
s
o
p
h
is
ticated
p
h
is
h
in
g
d
etec
tio
n
.
T
h
is
r
ec
o
m
m
e
n
d
ati
o
n
h
ig
h
lig
h
ts
th
e
im
p
o
r
tan
ce
o
f
ad
v
a
n
ce
d
f
ea
t
u
r
e
ex
tr
ac
tio
n
f
o
r
d
etec
tin
g
ev
o
lv
in
g
p
h
is
h
in
g
attac
k
s
.
A
n
o
th
er
ap
p
r
o
ac
h
was
p
r
o
p
o
s
ed
b
y
T
an
g
an
d
Ma
h
m
o
u
d
[
1
6
]
,
wh
o
p
r
esen
ted
a
DL
-
b
ased
f
r
am
ewo
r
k
th
at
o
p
er
at
es
in
d
ep
en
d
en
tly
o
f
th
ir
d
-
p
ar
t
y
s
er
v
ices
f
o
r
d
etec
t
in
g
p
h
is
h
in
g
we
b
s
ites
in
a
r
ea
l
-
tim
e
b
r
o
wser
p
lu
g
i
n
s
ettin
g
.
T
h
e
r
esear
ch
e
r
s
ev
alu
ated
s
ev
er
al
d
atasets
,
th
e
b
est o
f
wh
ich
was th
e
KPT
-
1
2
d
ataset,
wh
ich
co
n
tain
s
6
0
,
0
0
0
leg
itima
te
UR
L
s
f
r
o
m
Kag
g
le
an
d
6
0
,
0
0
0
p
h
i
s
h
in
g
UR
L
s
f
r
o
m
Ph
is
h
T
an
k
.
R
eg
ar
d
in
g
f
ea
tu
r
es,
th
ey
u
s
e
a
ch
ar
ac
ter
-
lev
el
to
k
en
izatio
n
ap
p
r
o
ac
h
,
lim
itin
g
ea
ch
UR
L
to
1
0
0
c
o
m
m
o
n
ly
u
s
ed
ch
ar
ac
ter
s
in
UR
L
s
an
d
ex
clu
d
in
g
UR
L
s
lo
n
g
er
th
an
2
0
0
c
h
ar
ac
ter
s
to
r
ed
u
ce
co
m
p
u
tatio
n
al
r
e
q
u
i
r
em
en
ts
.
Af
ter
co
n
d
u
ctin
g
th
e
ex
p
er
im
en
ts
,
th
e
r
esu
lts
d
em
o
n
s
tr
ated
t
h
at
th
e
R
NN
-
GR
U
m
o
d
el
ac
h
ie
v
ed
a
n
ac
cu
r
ac
y
o
f
9
9
.
1
8
%
an
d
a
r
esp
o
n
s
e
tim
e
o
f
less
th
an
2
0
0
m
illi
s
ec
o
n
d
s
(
0
.
2
s
ec
o
n
d
s
)
.
Nev
e
r
th
eless
,
th
e
au
th
o
r
s
r
ep
o
r
ted
th
at
th
e
m
o
d
el
d
o
e
s
n
o
t
s
u
p
p
o
r
t
s
h
o
r
t
UR
L
s
an
d
th
at
tr
u
n
ca
tin
g
UR
L
s
at
2
0
0
ch
ar
ac
ter
s
m
a
y
r
es
u
lt
in
th
e
l
o
s
s
o
f
s
o
m
e
f
ea
tu
r
es
f
o
u
n
d
in
l
o
n
g
e
r
UR
L
s
.
Sán
ch
ez
-
Pan
iag
u
a
et
a
l.
[
1
7
]
p
r
o
p
o
s
ed
a
f
r
am
ewo
r
k
f
o
r
d
et
ec
t
in
g
p
h
is
h
in
g
UR
L
s
b
y
r
e
p
r
esen
tin
g
a
r
ea
lis
tic
s
ce
n
ar
io
b
y
u
s
in
g
U
R
L
s
f
o
r
leg
itima
te
lo
g
in
p
ag
es
an
d
p
h
is
h
in
g
o
n
es
i
n
s
tead
o
f
ju
s
t
UR
L
s
f
o
r
h
o
m
ep
a
g
es.
T
o
im
p
lem
en
t
t
h
e
p
r
o
p
o
s
ed
s
y
s
tem
,
th
e
r
esear
ch
er
s
cr
ea
ted
a
d
ataset
co
n
s
is
tin
g
o
f
3
0
,
0
0
0
p
h
is
h
in
g
UR
L
s
co
l
lecte
d
f
r
o
m
Ph
is
h
T
an
k
an
d
3
0
,
0
0
0
le
g
itima
te
UR
L
s
o
b
tain
ed
f
r
o
m
th
e
to
p
m
illi
o
n
Qu
an
tcast
web
s
ites
.
T
h
is
d
ataset
in
co
r
p
o
r
ates
3
8
lex
ical
f
ea
tu
r
es
an
d
n
atu
r
al
lan
g
u
ag
e
p
r
o
ce
s
s
in
g
(
NL
P)
p
r
o
p
er
ties
,
s
u
c
h
as
wo
r
d
an
aly
s
is
an
d
p
atter
n
s
.
Af
ter
tr
ain
in
g
,
th
e
lo
g
is
tic
r
eg
r
ess
io
n
m
o
d
el
ac
h
iev
ed
th
e
b
est
p
er
f
o
r
m
an
ce
with
a
n
ac
cu
r
a
c
y
o
f
9
6
.
5
0
%.
Ho
wev
er
,
th
e
a
u
th
o
r
s
h
ig
h
lig
h
ted
s
o
m
e
lim
itatio
n
s
,
in
clu
d
in
g
a
s
lig
h
t
d
ec
r
ea
s
e
in
o
v
er
all
ac
c
u
r
ac
y
wh
en
u
s
in
g
lo
g
i
n
UR
L
s
d
u
e
to
th
eir
h
ig
h
s
im
ilar
ity
b
etwe
en
leg
itima
te
an
d
p
h
is
h
in
g
UR
L
s
.
T
h
is
lim
i
tatio
n
h
ig
h
lig
h
ts
th
e
im
p
o
r
tan
ce
o
f
ef
f
ec
tiv
e
f
ea
t
u
r
e
en
g
i
n
ee
r
in
g
to
im
p
r
o
v
e
th
e
d
is
cr
im
in
ativ
e
ca
p
ab
ilit
y
b
etw
ee
n
h
ig
h
l
y
s
im
ilar
leg
itima
te
an
d
p
h
is
h
in
g
UR
L
s.
I
n
g
en
er
al,
t
h
e
co
m
p
a
r
is
o
n
in
T
ab
le
1
in
d
icate
s
th
at
m
an
y
cu
r
r
en
t
ap
p
r
o
ac
h
es
h
av
e
ac
h
i
ev
ed
h
ig
h
d
etec
tio
n
ac
cu
r
ac
y
.
Ho
we
v
er
,
th
r
ee
m
ain
r
esear
ch
g
ap
s
ca
n
b
e
id
en
tifie
d
:
i
)
a
lim
ited
f
o
cu
s
o
n
co
m
p
u
tatio
n
al
ef
f
icien
cy
a
n
d
p
r
ac
tical
d
e
p
lo
y
m
en
t,
s
u
ch
as
r
el
ian
ce
o
n
th
ir
d
-
p
ar
ty
s
er
v
ices
o
r
r
ep
o
r
tin
g
r
elativ
ely
l
o
n
g
e
r
r
esp
o
n
s
e
tim
es,
ii
)
th
e
ev
alu
at
io
n
o
f
s
ev
e
r
al
s
tu
d
ies
is
lim
it
ed
b
y
th
e
u
s
e
o
f
s
m
all
-
s
ca
le
o
r
o
u
td
ated
d
atasets
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
4
3
,
No
.
3
,
Sep
tem
b
er
20
2
6
:
9
0
8
-
9
2
7
912
th
at
p
r
im
ar
ily
in
clu
d
e
tr
a
d
itio
n
al
f
ea
tu
r
es,
iii
)
lim
ited
tr
a
n
s
p
ar
en
cy
an
d
r
e
p
r
o
d
u
c
ib
ilit
y
,
p
ar
ticu
lar
ly
d
u
e
t
o
in
co
m
p
lete
r
ep
o
r
tin
g
o
f
th
e
ex
tr
ac
ted
f
ea
tu
r
e
s
ets
an
d
d
etec
tio
n
tim
e.
Ou
r
m
eth
o
d
o
lo
g
y
ai
m
s
to
ad
d
r
ess
th
ese
lim
itatio
n
s
b
y
b
alan
cin
g
h
ig
h
s
p
ee
d
an
d
ac
cu
r
ac
y
u
s
in
g
th
e
XGBo
o
s
t
m
o
d
el.
T
h
is
i
s
ac
h
iev
ed
b
y
u
s
in
g
a
r
ec
en
t
d
ataset
w
ith
ad
v
an
ce
d
f
ea
tu
r
es
th
at
s
u
p
p
o
r
t
th
e
d
et
ec
tio
n
o
f
s
o
p
h
is
ticated
p
h
is
h
in
g
attac
k
s
with
o
u
t
r
ely
in
g
o
n
th
ir
d
-
p
ar
t
y
s
er
v
ices.
Hav
in
g
a
r
o
b
u
s
t
d
ataset
an
d
m
o
d
er
n
f
ea
tu
r
es
ca
n
im
p
r
o
v
e
r
esil
ien
ce
to
ze
r
o
-
d
ay
attac
k
s
.
3.
M
E
T
H
O
D
I
n
th
is
s
tu
d
y
,
we
r
ev
iewe
d
r
e
ce
n
t
liter
atu
r
e
th
at
ad
d
r
ess
es
th
e
s
am
e
p
r
o
b
lem
o
f
d
etec
tin
g
p
h
is
h
in
g
web
s
ites
.
W
e
f
o
u
n
d
th
at
UR
L
-
b
ased
s
o
lu
tio
n
s
ar
e
e
f
f
ec
tiv
e
in
d
etec
tin
g
p
h
is
h
in
g
we
b
s
ites
[
1
1
]
.
T
h
er
ef
o
r
e,
th
e
p
r
im
ar
y
o
b
jectiv
e
o
f
th
is
r
esear
ch
is
to
d
e
v
elo
p
an
d
ev
al
u
ate
tr
ad
itio
n
al
ML
an
d
n
e
u
r
al
n
etwo
r
k
m
o
d
els
to
s
u
p
p
o
r
t
UR
L
-
b
ased
p
h
is
h
in
g
d
etec
tio
n
u
s
in
g
en
g
in
ee
r
ed
f
e
atu
r
es
wh
ile
m
ain
tain
in
g
ac
cu
r
ac
y
an
d
ef
f
icien
c
y
.
T
h
e
f
o
llo
win
g
s
u
b
s
ec
tio
n
p
r
o
v
id
es
d
etailed
in
f
o
r
m
atio
n
ab
o
u
t
th
e
s
o
u
r
c
es
an
d
s
tr
u
ctu
r
e
o
f
th
e
d
ataset.
I
n
th
is
r
esear
c
h
,
we
u
s
ed
a
s
et
o
f
f
ea
tu
r
es
th
at
ar
e
in
d
ep
e
n
d
en
t
o
f
t
h
ir
d
-
p
ar
ty
s
er
v
ices
an
d
h
ig
h
ly
e
f
f
ec
tiv
e
i
n
d
etec
tin
g
p
h
is
h
in
g
attac
k
s
.
I
n
ad
d
itio
n
,
we
tr
ain
e
d
t
h
e
m
o
d
e
ls
o
n
a
m
o
d
er
n
an
d
r
ea
lis
tic
d
ataset
th
at
in
clu
d
es
n
ew
an
d
m
o
r
e
co
m
p
lex
p
h
i
s
h
in
g
tech
n
iq
u
es.
T
o
ev
alu
at
e
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
n
eu
r
al
n
etwo
r
k
an
d
tr
ad
itio
n
al
ML
m
o
d
els,
we
u
s
ed
Py
th
o
n
to
p
er
f
o
r
m
q
u
a
n
titativ
e
an
aly
s
is
an
d
o
b
tain
s
t
atis
tical
r
e
s
u
lts
f
o
r
co
m
p
ar
in
g
th
e
m
o
d
els'
p
er
f
o
r
m
an
ce
an
d
id
en
tif
y
in
g
th
e
m
o
s
t
ef
f
ec
tiv
e
alg
o
r
ith
m
f
o
r
d
e
tectin
g
UR
L
-
b
ase
d
p
h
is
h
in
g
attac
k
s
.
T
h
e
m
o
d
el
d
ev
elo
p
m
e
n
t
p
h
ase
is
s
u
m
m
ar
ized
in
Fig
u
r
e
1
.
T
h
is
s
ec
tio
n
d
is
cu
s
s
es
th
e
r
esear
ch
m
eth
o
d
o
lo
g
y
,
wh
ich
co
m
p
r
is
es
th
e
f
o
llo
win
g
f
o
u
r
s
tep
s
:
d
ata
co
llectio
n
;
d
ataset
p
r
e
-
p
r
o
ce
s
s
in
g
,
in
clu
d
in
g
d
ata
clea
n
in
g
,
f
ea
t
u
r
e
ex
t
r
ac
tio
n
,
an
d
f
ea
tu
r
e
s
elec
tio
n
;
th
e
u
s
e
o
f
n
e
u
r
al
n
etwo
r
k
m
o
d
els
an
d
tr
ad
itio
n
al
ML
m
o
d
els
;
an
d
m
o
d
el
e
v
alu
atio
n
c
r
iter
ia.
T
h
en
,
th
e
f
in
al
d
ec
is
io
n
is
m
ad
e:
leg
itima
te
o
r
p
h
is
h
in
g
.
Fig
u
r
e
1
.
Ov
e
r
v
iew
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
lo
g
y
(
m
o
d
el
d
e
v
elo
p
m
en
t
p
h
ase
an
d
o
u
tp
u
t c
l
ass
if
icatio
n
)
3.1.
Da
t
a
co
llect
io
n
T
h
e
f
ir
s
t
s
tep
in
d
ev
elo
p
in
g
ML
m
o
d
els
is
to
u
s
e
a
h
ig
h
-
q
u
ality
d
ataset,
esp
ec
ially
i
n
p
h
is
h
in
g
d
etec
tio
n
.
W
e
co
llected
7
2
,
8
3
2
p
h
is
h
in
g
UR
L
s
f
r
o
m
Ph
is
h
T
an
k
,
a
web
s
ite
th
at
u
p
d
ates
p
h
is
h
in
g
d
atasets
h
o
u
r
ly
[
1
8
]
.
T
h
is
d
ataset
was
ac
ce
s
s
ed
in
Ma
y
2
0
2
3
.
L
eg
i
tim
ate
UR
L
s
wer
e
o
b
tain
ed
f
r
o
m
two
p
u
b
licly
av
ailab
le
s
o
u
r
ce
s
o
n
Me
n
d
eley
d
ata
[
1
9
]
an
d
Kag
g
le
[
2
0
]
p
latf
o
r
m
s
.
B
o
th
d
atasets
wer
e
ac
ce
s
s
ed
in
J
u
n
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
UR
L
-
b
a
s
ed
p
h
is
h
in
g
d
etec
tio
n
u
s
in
g
X
GB
o
o
s
t wi
th
en
g
in
ee
r
ed
fea
tu
r
es
(
Ja
w
a
h
er A
lh
a
r
b
i
)
913
2
0
2
3
.
T
h
e
Ka
g
g
le
d
ataset
co
n
tain
s
w
ell
-
k
n
o
wn
d
o
m
ain
s
,
w
h
ile
th
e
Me
n
d
eley
d
ataset
in
cl
u
d
es
less
co
m
m
o
n
d
o
m
ain
s
a
n
d
UR
L
s
r
ep
r
esen
t
in
g
d
i
f
f
er
en
t
web
s
ite
lev
els,
in
clu
d
in
g
lo
g
i
n
p
a
g
es.
T
h
is
ap
p
r
o
ac
h
aim
e
d
to
cr
ea
te
a
b
r
o
ad
a
n
d
u
n
b
iased
d
ataset.
T
o
cr
ea
te
a
le
g
itima
te
d
ataset,
we
em
p
lo
y
e
d
a
u
n
if
o
r
m
r
an
d
o
m
s
am
p
lin
g
tech
n
iq
u
e
o
n
b
o
th
s
ets,
r
esu
ltin
g
in
th
e
s
elec
tio
n
o
f
7
3
,
6
3
7
UR
L
s
f
r
o
m
Kag
g
le
an
d
3
,
3
8
3
UR
L
s
f
r
o
m
Me
n
d
eley
d
ata
.
3
.
2
.
Da
t
a
s
et
pre
-
pro
ce
s
s
ing
T
h
e
p
r
e
-
p
r
o
ce
s
s
in
g
s
tep
is
e
s
s
en
tial
f
o
r
ML
m
o
d
els,
as
it
p
r
ep
ar
es
th
e
d
ata
to
b
e
s
u
itab
le
f
o
r
class
if
icatio
n
tech
n
iq
u
es.
T
h
is
s
tep
in
clu
d
es
d
ata
clea
n
in
g
,
f
ea
tu
r
e
ex
tr
ac
tio
n
,
an
d
f
ea
t
u
r
e
s
elec
tio
n
.
W
e
ex
p
lain
ed
th
ese
s
tep
s
b
elo
w.
3
.
2
.
1
.
Da
t
a
clea
nin
g
Th
e
p
h
is
h
in
g
d
ataset
r
eq
u
ir
e
d
m
in
im
al
p
r
e
-
p
r
o
ce
s
s
in
g
,
as
o
n
ly
th
e
UR
L
co
lu
m
n
was
r
etain
ed
.
I
n
co
n
tr
ast,
th
e
leg
itima
te
d
ata
s
et
u
n
d
er
wen
t
ad
d
itio
n
al
p
r
e
-
p
r
o
ce
s
s
in
g
to
e
x
tr
ac
t
o
n
l
y
UR
L
s
lab
eled
as
leg
itima
te
b
ef
o
r
e
th
ey
wer
e
m
er
g
ed
.
Fo
llo
win
g
th
at
,
we
a
p
p
l
ied
UR
L
s
tan
d
ar
d
izatio
n
to
t
h
e
leg
itima
te
d
ataset
b
ef
o
r
e
f
ea
tu
r
e
ex
tr
ac
tio
n
.
Fo
r
Kag
g
le
d
ata
co
n
tain
in
g
o
n
ly
h
o
s
tn
am
es
(
s
u
ch
as
"ww
w.
y
o
u
t
u
b
e.
co
m
"
)
with
o
u
t
a
p
r
o
to
co
l,
we
ad
d
ed
an
h
ttp
:
//
p
r
ef
ix
to
en
ab
le
UR
L
r
eq
u
ests
an
d
d
eter
m
in
e
th
e
ac
tu
al
p
r
o
to
co
l
(
HT
T
P
o
r
HT
T
PS
)
b
ased
o
n
th
e
s
er
v
er
'
s
r
esp
o
n
s
e.
T
h
is
s
tep
was
p
er
f
o
r
m
ed
s
o
lely
f
o
r
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
d
i
d
n
o
t
alter
th
e
s
em
an
tic
s
tr
u
ctu
r
e
o
f
th
e
UR
L
s
.
Sev
er
al
p
r
elim
in
ar
y
ex
p
er
i
m
en
ts
wer
e
co
n
d
u
ct
ed
to
d
eter
m
in
e
a
n
ap
p
r
o
p
r
iate
co
n
n
ec
tio
n
tim
eo
u
t.
B
ased
o
n
th
ese
e
x
p
er
im
en
ts
,
a
tim
eo
u
t o
f
1
2
0
s
ec
o
n
d
s
was
ch
o
s
en
to
p
r
o
v
id
e
s
u
f
f
icien
t
tim
e
f
o
r
UR
L
r
eq
u
ests
with
v
ar
y
in
g
s
er
v
er
r
esp
o
n
s
e
tim
es,
esp
ec
ially
th
o
s
e
h
o
s
ted
o
n
s
lo
we
r
s
er
v
er
s
.
C
o
n
s
eq
u
en
tly
,
th
e
leg
itima
te
d
ataset
co
n
tain
ed
7
6
,
9
9
3
UR
L
s
.
Fin
ally
,
th
e
p
h
is
h
in
g
an
d
leg
i
tim
ate
d
atasets
wer
e
m
er
g
e
d
.
Af
ter
r
em
o
v
in
g
d
u
p
licates
an
d
m
is
s
in
g
d
ata,
th
e
r
esu
ltin
g
d
ataset
co
n
tain
ed
9
5
,
7
3
6
UR
L
s
,
o
f
wh
ich
3
9
,
3
6
7
wer
e
leg
itima
te
UR
L
s
(
lab
el
0
)
,
an
d
5
6
,
3
6
9
wer
e
p
h
is
h
in
g
UR
L
s
(
lab
el
1
)
.
T
h
e
d
ataset
was
th
en
s
h
u
f
f
led
a
n
d
s
p
lit
i
n
t
o
9
0
%
tr
ain
in
g
a
n
d
1
0
%
test
in
g
u
s
in
g
s
tr
atif
ied
s
am
p
li
n
g
(
r
an
d
o
m
_
s
tate
=
4
2
)
.
I
n
o
u
r
ap
p
r
o
ac
h
,
we
ap
p
lied
o
n
e
-
h
o
t
en
co
d
in
g
,
wh
ich
co
n
v
er
ts
ca
teg
o
r
ical
f
ea
tu
r
es
i
n
to
a
n
u
m
er
ical
f
o
r
m
at
th
at
ML
m
o
d
els
ca
n
p
r
o
ce
s
s
to
av
o
id
th
e
p
r
o
b
lem
o
f
o
r
d
in
ality
.
I
n
b
in
a
r
y
class
if
icatio
n
,
d
ataset
b
alan
ce
is
p
ar
ticu
lar
ly
im
p
o
r
tan
t
wh
en
ac
cu
r
ac
y
is
u
s
ed
to
ass
es
s
m
o
d
el
p
e
r
f
o
r
m
an
ce
[
2
1
]
.
Alth
o
u
g
h
p
er
f
ec
tly
b
alan
ce
d
d
atasets
ar
e
u
n
co
m
m
o
n
,
an
im
b
alan
ce
r
atio
o
f
ap
p
r
o
x
im
ately
1
0
:1
is
g
en
e
r
ally
co
n
s
id
er
e
d
m
o
d
estl
y
im
b
alan
ce
d
,
wh
er
ea
s
r
atio
s
ex
c
ee
d
in
g
1
0
0
0
:1
ar
e
r
eg
ar
d
e
d
as
h
ig
h
ly
im
b
alan
ce
d
[
2
2
]
.
Fo
r
o
u
r
d
ataset,
af
ter
clea
n
in
g
it
b
y
r
em
o
v
in
g
m
is
s
in
g
an
d
d
u
p
licate
v
alu
es,
th
e
r
atio
o
f
t
h
e
m
ajo
r
i
ty
(
p
h
is
h
in
g
)
class
to
th
e
m
i
n
o
r
ity
(
leg
itima
te)
class
was
ap
p
r
o
x
im
ately
1
.
4
:1
(
5
6
,
3
6
9
/3
9
,
3
6
7
)
.
T
h
is
r
atio
f
alls
with
in
th
e
"m
ar
g
in
ally
im
b
a
lan
ce
d
"
r
an
g
e
,
wh
ich
m
ea
n
s
it
ca
n
b
e
co
n
s
id
er
e
d
s
af
ely
b
alan
ce
d
.
3
.
2
.
2
.
F
ea
t
ure
e
x
t
ra
ct
i
o
n
T
h
e
s
tu
d
y
r
elied
o
n
a
wid
e
r
an
g
e
o
f
liter
atu
r
e
r
elate
d
to
d
etec
tin
g
p
h
is
h
in
g
web
s
ites
t
o
id
en
tify
ef
f
ec
tiv
e
UR
L
-
b
ased
f
ea
tu
r
es.
I
n
ad
d
itio
n
,
we
c
o
n
d
u
cted
a
n
in
-
d
e
p
th
a
n
aly
s
is
o
f
th
e
p
h
is
h
in
g
an
d
leg
itima
cy
d
ataset
to
id
en
tify
ad
d
itio
n
al
UR
L
-
b
ased
d
is
cr
im
in
ato
r
y
f
ea
tu
r
es.
As
a
r
esu
lt,
we
id
en
tifie
d
n
ew
p
h
is
h
in
g
tech
n
iq
u
es
n
o
t
y
et
d
o
cu
m
e
n
te
d
in
th
e
liter
atu
r
e,
d
is
cu
s
s
ed
in
Sectio
n
6
,
Dis
cu
s
s
io
n
with
r
ec
o
m
m
en
d
atio
n
s
.
W
e
in
ten
tio
n
ally
ex
clu
d
e
d
f
e
atu
r
es
th
at
r
ely
o
n
th
ir
d
-
p
ar
t
y
s
er
v
ices
s
u
ch
as
Go
o
g
le
i
n
d
ex
an
d
Pag
eRan
k
b
ec
au
s
e
th
ey
ar
e
n
o
t
alwa
y
s
a
cc
ess
ib
le
th
r
o
u
g
h
f
r
ee
API
s
,
m
ak
in
g
t
h
em
im
p
r
ac
tical
f
o
r
r
ea
l
-
tim
e
d
etec
tio
n
.
W
e
also
ex
clu
d
ed
DNS
an
d
d
o
m
ain
a
g
e
f
ea
tu
r
es,
s
u
ch
as
th
e
cr
ea
tio
n
d
ate
a
n
d
e
x
p
ir
at
io
n
d
ate
o
f
web
s
ite
in
f
o
r
m
atio
n
o
b
tain
e
d
f
r
o
m
th
e
W
HOI
S
d
atab
ase.
T
h
e
d
o
m
ain
ag
e
f
ea
tu
r
e
is
m
o
r
e
h
elp
f
u
l
in
d
etec
tin
g
lo
n
g
-
ter
m
p
h
is
h
in
g
web
s
ites
,
b
u
t
it
is
in
ef
f
ec
tiv
e
f
o
r
ze
r
o
-
d
a
y
att
ac
k
s
.
T
h
is
d
ec
is
io
n
aim
s
to
s
p
ee
d
u
p
th
e
f
ea
tu
r
e
ex
tr
ac
tio
n
p
r
o
ce
s
s
an
d
h
elp
im
p
r
o
v
e
th
e
d
etec
tio
n
o
f
ze
r
o
-
d
ay
attac
k
s
.
Ad
d
itio
n
ally
,
it
r
ed
u
ce
s
p
o
ten
tial
p
r
iv
ac
y
c
o
n
ce
r
n
s
b
y
av
o
id
in
g
th
e
tr
an
s
m
is
s
io
n
o
f
u
s
er
d
ata
to
ex
ter
n
al
s
er
v
e
r
s
f
o
r
a
n
aly
s
is
.
Acc
o
r
d
in
g
ly
,
we
ex
tr
ac
ted
7
3
f
ea
tu
r
es
th
at
co
m
b
in
e
lex
i
ca
l
f
ea
tu
r
es
with
s
ec
u
r
ity
an
d
p
r
o
to
c
o
l
f
ea
tu
r
es.
L
ex
ical
f
ea
tu
r
es
r
ef
e
r
to
th
e
s
tr
u
ct
u
r
e
o
f
UR
L
s
tr
in
g
s
,
n
u
m
b
e
r
o
f
ch
ar
ac
te
r
s
,
s
y
m
b
o
ls
s
u
ch
as
(
@
,
/,
-
,
)
,
as
well
as
an
aly
s
is
o
f
UR
L
co
m
p
o
n
e
n
ts
s
u
ch
as
p
ath
an
d
q
u
er
y
p
ar
t
[
2
3
]
.
Mo
r
eo
v
e
r
,
lex
ical
f
ea
tu
r
es
ca
n
d
etec
t
p
h
is
h
in
g
UR
L
s
g
en
er
a
ted
b
y
AI
[
2
4
]
.
B
ec
au
s
e
AI
r
elies
o
n
lex
ical
f
ea
tu
r
es
to
s
im
u
late
leg
itima
te
UR
L
s
.
R
eg
ar
d
in
g
s
ec
u
r
ity
an
d
p
r
o
to
c
o
ls
,
th
ese
f
ea
tu
r
es
h
elp
d
etec
t
p
h
is
h
in
g
UR
L
s
m
o
r
e
s
o
p
h
is
ticated
ly
.
Fo
r
ex
am
p
le,
p
h
is
h
in
g
UR
L
s
u
s
e
in
ter
n
et
p
r
o
to
co
l
(
I
P)
ad
d
r
ess
es
in
th
e
d
o
m
ai
n
f
ield
,
em
p
l
o
y
n
o
n
-
s
tan
d
ar
d
p
o
r
t
n
u
m
b
er
s
,
ex
p
lo
it
d
ec
en
tr
alize
d
p
r
o
to
c
o
ls
s
u
ch
as
I
PF
S,
an
d
lev
er
ag
e
f
r
ee
s
er
v
ices
lik
e
f
r
ee
h
o
s
tin
g
d
o
m
ai
n
(
FHD)
s
er
v
ices.
C
o
m
b
in
in
g
lex
ical
f
ea
tu
r
es
with
s
ec
u
r
ity
an
d
p
r
o
t
o
co
l
f
ea
tu
r
es
h
elp
s
e
x
p
an
d
th
e
d
etec
tio
n
r
an
g
e
an
d
p
r
o
v
id
es b
r
o
ad
e
r
an
aly
s
is
o
f
p
h
is
h
in
g
UR
L
s
.
At
t
h
is
s
tag
e,
we
id
en
tifie
d
7
3
UR
L
-
b
ased
f
ea
tu
r
es,
p
r
esen
ted
in
a
s
tr
u
ct
u
r
ed
ta
b
le
th
at
in
cl
u
d
es
th
eir
r
ef
er
en
ce
s
an
d
a
n
ex
p
lan
atio
n
o
f
t
h
eir
p
u
r
p
o
s
e
an
d
ca
teg
o
r
y
:
ca
teg
o
r
ical,
co
u
n
t,
a
n
d
n
u
m
er
ical
f
ea
tu
r
es
(
s
ee
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
4
3
,
No
.
3
,
Sep
tem
b
er
20
2
6
:
9
0
8
-
9
2
7
914
T
ab
le
1
in
Su
p
p
lem
en
tar
y
1
)
.
T
h
ese
ex
tr
ac
ted
f
e
atu
r
es
r
ef
lect
th
e
co
m
p
lex
ity
o
f
d
etec
tin
g
n
ew
an
d
ad
v
an
ce
d
UR
L
-
b
ased
p
h
is
h
in
g
attac
k
s
.
3
.
2
.
3
.
F
ea
t
ure
s
elec
t
io
n
Af
ter
ex
tr
ac
tin
g
7
3
f
ea
t
u
r
es,
we
ap
p
lied
Pear
s
o
n
'
s
co
r
r
elatio
n
co
e
f
f
icien
t
[
2
5
]
to
id
en
ti
f
y
h
i
g
h
ly
co
r
r
elate
d
f
ea
tu
r
es
an
d
r
e
d
u
ce
r
ed
u
n
d
an
cy
.
Acc
o
r
d
i
n
g
ly
,
f
e
atu
r
es
with
a
co
r
r
elatio
n
c
o
ef
f
icien
t
g
r
ea
ter
th
a
n
0
.
8
wer
e
ex
clu
d
ed
,
wh
ile
th
e
m
o
r
e
im
p
o
r
tan
t
f
ea
tu
r
es
wer
e
r
etain
ed
.
Ho
wev
er
,
c
o
r
r
elatio
n
an
aly
s
is
r
ev
ea
led
th
at
th
e
"c
o
u
n
t_
h
ttp
"
a
n
d
"c
o
u
n
t_
h
ttp
s
"
f
ea
tu
r
es,
as
well
as
th
e
"SDT
_
wo
r
d
s
_
d
o
t"
a
n
d
"
SDT_
wo
r
d
s
_
s
lash
"
f
ea
tu
r
es,
wer
e
ass
o
ciate
d
with
"lo
n
g
_
u
r
l".
R
em
o
v
in
g
o
r
c
o
m
b
in
in
g
th
ese
f
ea
tu
r
es
r
esu
l
ted
in
a
lo
wer
F1
-
s
co
r
e,
in
d
icatin
g
th
at
th
ey
p
r
o
v
id
e
co
m
p
lem
en
tar
y
p
r
ed
ict
iv
e
in
f
o
r
m
atio
n
.
T
h
er
e
f
o
r
e,
t
h
ese
f
ea
tu
r
es
wer
e
r
etain
ed
.
Fo
llo
win
g
t
h
at,
w
e
u
s
ed
r
ec
u
r
s
iv
e
f
ea
tu
r
e
elim
in
atio
n
with
cr
o
s
s
-
v
alid
atio
n
(
R
FECV)
with
XGBo
o
s
t
as
th
e
b
ase
esti
m
ato
r
to
i
d
e
n
tify
th
e
o
p
tim
al
s
u
b
s
et
o
f
f
ea
tu
r
es
[
2
6
]
.
No
tab
l
y
,
th
e
f
e
atu
r
e
s
elec
tio
n
was
p
er
f
o
r
m
ed
ex
cl
u
s
iv
ely
o
n
th
e
tr
ain
in
g
s
et
to
p
r
ev
en
t
d
ata
lea
k
ag
e,
an
d
th
e
s
elec
ted
f
ea
t
u
r
e
s
wer
e
th
en
ap
p
lie
d
to
b
o
th
th
e
tr
ain
i
n
g
an
d
test
s
ets.
Featu
r
es
wer
e
r
ec
u
r
s
iv
ely
elim
in
ated
o
n
e
at
a
tim
e
(
s
tep
=
1
)
an
d
e
v
alu
ated
u
s
in
g
s
tr
atif
ied
5
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
with
ac
cu
r
ac
y
as
th
e
s
co
r
in
g
m
etr
ic.
T
o
q
u
an
ti
tativ
ely
ass
e
s
s
th
e
co
n
tr
ib
u
tio
n
o
f
th
e
s
elec
ted
f
ea
tu
r
es,
Gain
-
b
ased
f
ea
t
u
r
e
i
m
p
o
r
tan
ce
s
co
r
es
wer
e
ex
tr
ac
ted
f
r
o
m
th
e
tr
ain
e
d
XGBo
o
s
t
esti
m
ato
r
an
d
an
a
ly
ze
d
.
C
o
n
s
eq
u
e
n
tly
,
4
0
f
ea
tu
r
es
wer
e
s
elec
ted
as
th
e
o
p
tim
al
s
u
b
s
et,
as
illu
s
tr
ated
in
Fig
u
r
e
2
.
T
h
e
m
o
s
t im
p
o
r
tan
t f
e
atu
r
es a
r
e
m
ar
k
ed
with
an
aster
is
k
(
*
)
,
wh
ile
t
h
e
d
elete
d
f
ea
tu
r
es
ar
e
m
ar
k
ed
with
a
p
l
u
s
(
+)
in
th
e
n
o
v
el
tab
le
(
S
u
p
p
lem
e
n
tar
y
1
,
T
a
b
le
1
)
.
T
h
e
4
0
s
e
lecte
d
f
ea
tu
r
es
ar
e
ex
p
lain
ed
b
elo
w,
with
'
#
F'
d
en
o
tin
g
th
e
f
ea
t
u
r
e
n
u
m
b
er
in
T
a
b
le
1
.
Fig
u
r
e
2
.
Gain
-
b
ased
f
ea
tu
r
e
i
m
p
o
r
tan
ce
o
b
tain
ed
f
r
o
m
th
e
XGBo
o
s
t m
o
d
el
B
r
an
d
n
am
es
an
d
s
u
s
p
icio
u
s
wo
r
d
s
:
Ph
is
h
in
g
attac
k
er
s
f
r
eq
u
en
tly
u
s
e
well
-
k
n
o
wn
b
r
an
d
n
am
es
an
d
s
u
s
p
icio
u
s
wo
r
d
s
in
UR
L
s
.
Featu
r
e
2
F
co
u
n
ts
th
e
f
r
eq
u
e
n
cy
o
f
b
r
an
d
n
a
m
es
with
in
UR
L
s
.
As
s
h
o
wn
in
E
x
am
p
le
1
o
f
a
p
h
is
h
in
g
UR
L
f
r
o
m
o
u
r
d
ataset,
th
e
Pay
Pa
l
b
r
an
d
is
u
s
ed
twice
in
d
if
f
e
r
en
t
p
o
s
itio
n
s
with
in
th
e
UR
L
.
T
h
e
y
also
u
s
e
s
u
s
p
icio
u
s
wo
r
d
s
lik
e
"sig
n
in
"
to
e
n
t
ice
u
s
er
s
to
v
is
it
th
e
p
h
is
h
in
g
p
ag
e
(
s
ig
n
in
.
h
tm
l)
.
T
h
er
ef
o
r
e,
Featu
r
e
8
F
c
o
u
n
ts
th
e
f
r
e
q
u
en
c
y
o
f
s
en
s
itiv
e
wo
r
d
s
in
UR
L
s
.
T
h
is
r
e
v
ea
led
t
h
at
th
e
attac
k
er
s
u
s
e
m
o
r
e
th
an
o
n
e
s
en
s
itiv
e
wo
r
d
in
a
UR
L
,
wh
ile
leg
it
im
ate
UR
L
s
u
s
e
at
m
o
s
t
o
n
e.
B
ased
o
n
o
u
r
d
ataset,
we
o
b
s
er
v
ed
th
at
attac
k
er
s
u
s
e
b
r
an
d
n
am
es
at
th
e
d
o
m
ain
lev
el
b
u
t
with
a
f
a
k
e
to
p
-
le
v
el
d
o
m
ai
n
(
T
L
D)
.
T
h
er
ef
o
r
e,
f
ea
tu
r
e
5
5
F
d
etec
ts
th
e
p
r
esen
ce
o
f
b
r
an
d
n
am
es in
a
UR
L
.
Als
o
,
f
ea
tu
r
e
3
5
F
co
u
n
ts
th
e
n
u
m
b
er
o
f
tim
es
co
n
s
ec
u
tiv
e
ch
ar
ac
ter
s
ar
e
r
ep
ea
ted
in
UR
L
s
,
s
u
ch
as
'
p
ay
p
al.
co
m
'
to
'
p
ay
p
all.
co
m
'
.
Du
e
to
th
e
in
h
e
r
en
t
r
e
p
etitio
n
o
f
ch
ar
ac
ter
s
in
UR
L
k
ey
wo
r
d
s
,
we
h
a
v
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
UR
L
-
b
a
s
ed
p
h
is
h
in
g
d
etec
tio
n
u
s
in
g
X
GB
o
o
s
t wi
th
en
g
in
ee
r
ed
fea
tu
r
es
(
Ja
w
a
h
er A
lh
a
r
b
i
)
915
in
ten
tio
n
ally
e
x
clu
d
ed
p
r
o
to
c
o
l
in
d
icato
r
s
s
u
c
h
as
HT
T
P,
HT
T
PS
,
an
d
W
W
W
.
T
h
is
ap
p
r
o
ac
h
ca
n
s
u
p
p
o
r
t
ef
f
icien
t d
etec
tio
n
o
f
p
h
is
h
in
g
attac
k
s
.
T
o
p
-
lev
el
d
o
m
ai
n
(
T
L
D)
:
C
y
b
er
cr
im
in
als
u
tili
ze
T
L
Ds
t
o
cr
ea
te
UR
L
s
th
at
ap
p
ea
r
l
eg
itima
te.
Netc
r
af
t
[
2
7
]
id
en
tifie
d
5
0
T
L
Ds
ass
o
ciate
d
with
p
h
is
h
in
g
at
tack
s
,
f
o
r
e
x
am
p
le,
(
.
s
h
o
p
,
.
t
o
p
,
.
cf
d
,
an
d
.
clo
u
d
)
.
Fu
r
th
er
m
o
r
e
,
th
ey
h
ig
h
lig
h
ted
h
o
w
attac
k
er
s
ex
p
lo
ited
n
ew
T
L
Ds
r
elea
s
ed
b
y
Go
o
g
le
in
Ma
r
ch
2
0
2
3
[
2
8
]
,
n
o
tab
ly
th
e
T
L
Ds
with
n
am
es
s
im
ilar
to
f
ile
e
x
ten
s
io
n
s
,
s
u
ch
as
'
.
zip
'
an
d
'
.
m
o
v
'
.
T
h
ese
d
o
m
ain
s
we
r
e
s
u
b
s
eq
u
en
tly
ass
o
ciate
d
with
b
r
an
d
n
a
m
es
an
d
s
en
s
itiv
e
wo
r
d
s
to
attr
ac
t
u
s
er
s
[
2
9
]
.
T
h
er
e
f
o
r
e,
we
d
ev
elo
p
ed
f
ea
tu
r
e
5
3
F to
d
etec
t sen
s
itiv
e
wo
r
d
s
ass
o
ciat
ed
with
s
u
s
p
icio
u
s
T
L
Ds.
Featu
r
es
7
F
an
d
6
F
co
u
n
t
th
e
f
r
eq
u
en
c
y
o
f
T
L
Ds
in
h
o
s
tn
am
es
an
d
ex
ten
s
io
n
s
(
s
u
ch
as
.
h
tm
l
an
d
.
p
h
p
)
ap
p
ea
r
in
g
in
th
e
p
ath
p
o
r
tio
n
o
f
a
UR
L
.
Attack
er
s
u
s
e
th
e
T
L
D
'
co
m
'
b
ec
au
s
e
it
is
f
r
eq
u
e
n
tly
u
s
ed
in
leg
itima
te
UR
L
s
.
As
s
h
o
wn
in
E
x
am
p
le
1
,
a
p
h
is
h
in
g
UR
L
f
r
o
m
o
u
r
d
ataset
s
h
o
ws
th
e
f
r
e
q
u
en
cy
o
f
t
h
e
'
co
m
'
in
th
e
h
o
s
tn
am
e
an
d
p
ath
p
o
r
t
io
n
s
.
T
h
e
r
ef
o
r
e
,
Featu
r
e
3
F
tr
ac
k
s
th
e
f
r
eq
u
en
cy
o
f
th
e
'
co
m
'
T
L
D
ap
p
ea
r
i
n
g
in
UR
L
s
.
E
x
am
p
le
1
:
h
ttp
://www.
asas
k
o
f
o
r
d
h
a
m
.
co
m
/
p
ay
p
al/w
ww.
p
ay
p
al.
co
m
/s
ig
n
i
n
.
h
tm
l
.
Sp
ec
ial
ch
ar
ac
ter
s
:
Ph
is
h
in
g
UR
L
s
o
f
ten
u
s
e
s
y
m
b
o
ls
s
tr
ateg
ically
to
ex
p
lo
it
b
r
o
wser
b
eh
av
io
r
.
B
ased
o
n
o
u
r
an
aly
s
is
,
p
h
is
h
i
n
g
UR
L
s
ex
ce
s
s
iv
ely
u
s
e
th
e
'@
'
s
y
m
b
o
l
ac
r
o
s
s
m
u
ltip
le
U
R
L
p
ar
ts
,
as
s
h
o
wn
in
E
x
am
p
le
2
o
f
a
p
h
is
h
in
g
U
R
L
f
r
o
m
o
u
r
d
ataset.
As
a
r
es
u
lt,
f
ea
tu
r
e
4
F
co
u
n
ts
th
e
n
u
m
b
er
o
f
tim
es
th
e
'@
'
s
y
m
b
o
l a
p
p
ea
r
s
in
a
UR
L
.
E
x
am
p
le
2
:
h
ttp
s
://p
r
ev
en
tin
g
o
v
er
d
iag
n
o
s
is
.
n
et/@
@
/
Glo
b
alSo
u
r
ce
s
/?em
ail=a
b
u
s
e@
ch
ase
.
co
m
.
Ad
d
itio
n
ally
,
we
e
x
am
in
ed
th
e
o
cc
u
r
r
en
ce
s
o
f
s
y
m
b
o
ls
s
u
ch
as
'
-
'
(
2
5
F)
in
h
o
s
tn
am
es
an
d
(
2
4
F)
in
q
u
er
y
,
w
h
ile
'
/
'
(
3
4
F),
an
d
'
//'
(
6
4
F)
in
th
e
p
ath
,
an
d
'
.
'
(
2
3
F),
'
%
'
(
2
9
F),
'
_
'
(
2
2
F),
'
:
'
(
5
6
F),
'
#
'
(
1
5
F),
'
?
'
(
2
8
F)
ac
r
o
s
s
all
UR
L
p
ar
ts
.
Fu
r
th
er
m
o
r
e,
p
h
is
h
in
g
UR
L
s
m
ay
em
p
lo
y
m
u
ltip
le
en
c
o
d
in
g
tech
n
i
q
u
es,
s
u
ch
as
s
in
g
le
o
r
d
o
u
b
l
e
p
er
ce
n
t
-
e
n
co
d
i
n
g
,
to
h
id
e
s
u
s
p
icio
u
s
p
atter
n
s
o
r
b
y
p
ass
m
ec
h
an
is
m
s
th
at
r
ely
o
n
b
asic
UR
L
ch
ar
ac
ter
is
tics
.
I
n
th
e
p
h
is
h
in
g
d
ataset,
we
n
o
ticed
th
at
h
ac
k
er
s
u
s
e
th
e
q
u
esti
o
n
m
ar
k
(
?
)
in
an
e
x
ce
s
s
iv
ely
en
co
d
ed
way
,
r
ep
lacin
g
it
with
'
%3
f
'
in
U
R
L
s
.
T
h
er
ef
o
r
e,
f
ea
tu
r
e
9
F
co
u
n
ts
en
co
d
e
d
q
u
esti
o
n
m
ar
k
s
in
a
UR
L
.
Ou
r
an
aly
s
is
r
ev
ea
led
th
at
attac
k
er
s
also
u
s
e
th
e
n
o
n
-
s
tan
d
ar
d
d
ir
ec
to
r
y
s
y
m
b
o
l '
/'
w
it
h
in
th
e
q
u
e
r
y
p
ar
t,
as
s
h
o
wn
in
E
x
a
m
p
le
3
.
T
h
er
ef
o
r
e,
f
ea
tu
r
e
3
2
F
co
u
n
ts
th
e
o
cc
u
r
r
en
ce
s
o
f
t
h
e
'
/'
s
y
m
b
o
l
in
t
h
e
q
u
er
y
p
a
r
t.
E
x
am
p
le
3
:
h
ttp
s
://6
4
5
1
7
3
d
6
9
d
7
d
d
a0
e6
4
b
5
d
f
e6
.
t
r
k
.
m
ailch
ef
.
4
d
em
.
it/t
tn
.
p
h
p
?p
=e
6
0
d
4
4
f
0
4
9
b
0
3
3
5
f
0
f
4
2
5
6
5
6
5
b
9
6
7
9
4
4
/8
0
4
s
/7
cld
/r
s
/r
s
/r
s
/r
t/h
ttp
s
://co
m
p
-
ass
y
s
t.c
o
m
//o
r
g
b
k
-
b
/C
lien
t/co
n
n
ex
i
o
n
.
h
tm
l.
Qu
er
y
s
tr
in
g
s
p
lay
a
cr
u
cial
r
o
le
in
tr
an
s
m
itti
n
g
d
ata
to
t
h
e
s
er
v
er
with
in
th
e
UR
L
.
T
h
er
e
f
o
r
e,
f
ea
tu
r
e
1
0
F
co
u
n
ts
th
e
p
ar
am
ete
r
s
in
th
e
UR
L
s
'
q
u
er
y
p
ar
ts
.
I
n
a
d
d
itio
n
,
f
ea
tu
r
e
2
6
F
co
u
n
ts
a
s
p
ac
e
'
%2
0
'
with
in
UR
L
s
.
B
ased
o
n
o
u
r
d
ataset,
we
f
o
u
n
d
t
h
at
attac
k
er
s
cr
ea
te
co
m
p
lex
UR
L
s
b
y
u
s
in
g
d
o
u
b
le
-
en
co
d
e
d
s
p
ac
es,
ad
d
in
g
th
e
p
r
ef
ix
'
%2
5
'
,
wh
ich
r
esu
lts
in
'
%2
5
2
0
'
.
Ho
wev
e
r
,
p
r
e
v
io
u
s
r
esear
c
h
h
as
n
o
t
wid
ely
d
o
c
u
m
en
ted
th
is
tech
n
iq
u
e
as
a
s
tr
o
n
g
in
d
icato
r
o
f
p
h
is
h
in
g
d
etec
tio
n
.
Ou
r
m
eth
o
d
o
lo
g
y
d
o
es
n
o
t
d
e
ter
m
in
e
a
th
r
esh
o
ld
f
o
r
class
if
y
in
g
UR
L
s
,
a
s
n
o
s
p
ec
if
ic,
r
eliab
ly
in
d
icate
d
p
h
i
s
h
in
g
b
eh
av
io
r
ex
is
ts
.
T
o
en
s
u
r
e
h
ig
h
ex
tr
ac
tio
n
ac
cu
r
ac
y
,
we
e
x
clu
d
ed
th
e
co
l
o
n
in
th
e
p
r
o
t
o
co
l p
a
r
t,
e.
g
.
,
'
HT
T
P:/
/
'
o
f
th
e
UR
L
.
Pu
n
y
co
d
e:
T
h
e
i
n
ter
n
atio
n
aliz
ed
d
o
m
ain
n
a
m
es
(
I
DNs)
tech
n
o
lo
g
y
e
n
ab
les
u
s
er
s
to
r
eg
is
ter
d
o
m
ain
n
am
es
in
th
ei
r
lo
ca
l
lan
g
u
a
g
e
s
b
y
a
d
d
in
g
th
e
p
r
ef
i
x
'
x
n
--
'
a
t
th
e
b
e
g
in
n
in
g
o
f
th
e
UR
L
[
3
0
]
.
T
h
e
r
esear
c
h
er
s
in
[
3
1
]
,
[
3
2
]
ex
tr
ac
te
d
th
e
Pu
n
y
co
d
e
f
ea
tu
r
e
,
an
d
o
n
ce
p
r
es
en
t,
th
e
UR
L
is
class
if
ied
as
p
h
is
h
i
n
g
.
B
ased
o
n
o
u
r
d
ataset,
we
f
o
u
n
d
m
an
y
le
g
itima
te
UR
L
s
th
at
co
n
tain
Pu
n
y
co
d
e
o
n
ly
o
n
ce
in
t
h
e
h
o
s
tn
am
e,
m
ea
n
in
g
th
at
ju
d
g
in
g
a
UR
L
b
ased
o
n
t
h
e
ap
p
ea
r
a
n
ce
o
f
Pu
n
y
co
d
e
i
n
cr
ea
s
es
th
e
r
ate
o
f
f
alse
p
o
s
itiv
es.
C
o
n
v
er
s
ely
,
attac
k
er
s
u
s
e
m
u
ltip
le
Pu
n
y
c
o
d
es
(
x
n
-
-
)
i
n
a
UR
L
,
as
s
h
o
wn
in
E
x
am
p
le
4
f
r
o
m
o
u
r
d
ataset,
in
d
icatin
g
an
attem
p
t
to
co
m
p
licate
th
e
UR
L
b
y
en
co
d
in
g
d
i
f
f
er
en
t
p
ar
ts
o
f
th
e
h
o
s
tn
am
e.
As
a
r
esu
lt,
i
n
o
u
r
m
et
h
o
d
o
l
o
g
y
,
we
u
s
ed
f
ea
tu
r
e
5
F to
co
u
n
t th
e
n
u
m
b
er
o
f
o
cc
u
r
r
en
ce
s
o
f
Pu
n
y
co
d
e
(
x
n
--
)
with
in
t
h
e
h
o
s
tn
am
e
in
UR
L
s
.
I
t i
s
wo
r
th
n
o
tin
g
th
at
p
r
ev
io
u
s
s
tu
d
ies
h
av
e
n
o
t
ex
tr
ac
ted
th
is
f
e
atu
r
e
in
th
is
way
.
Fu
r
th
er
m
o
r
e,
th
er
e
is
a
lack
o
f
r
ec
en
t
s
tu
d
ies
o
n
ex
tr
ac
tin
g
an
d
ac
cu
r
ately
a
n
aly
zin
g
t
h
e
Pu
n
y
co
d
e
f
ea
tu
r
e
to
d
etec
t
p
h
is
h
in
g
attac
k
s
.
E
x
am
p
le
4
:
h
tt
p
://wtp
f
y
cin
2
-
xn
--
f
f
g
y
3
s
o
d
-
xn
----
c
1
ac
4
b
x
c
-
xn
----
p
1
ai.
tr
an
s
late.
x
n
--
c1
ac
4
b
x
c.
x
n
--
p
1
ai
.
Po
r
t
n
u
m
b
e
r
an
d
p
r
o
to
c
o
ls
(
HT
T
P
an
d
HT
T
PS
)
:
Secu
r
ity
m
ec
h
an
is
m
s
m
o
n
ito
r
k
n
o
wn
p
o
r
ts
,
s
u
ch
as
8
0
f
o
r
HT
T
P
an
d
4
4
3
f
o
r
HT
T
PS
,
to
en
s
u
r
e
n
etwo
r
k
s
ec
u
r
ity
.
Featu
r
e
7
2
F
r
etu
r
n
s
th
e
p
o
r
t
n
u
m
b
e
r
in
an
y
UR
L
p
ar
t.
B
ec
au
s
e
th
e
attac
k
er
s
ex
p
lo
it
m
ix
ed
co
n
ten
t
tec
h
n
iq
u
es
to
h
id
e
n
o
n
-
s
tan
d
a
r
d
p
o
r
ts
,
s
u
ch
as
8
0
8
1
,
as
s
h
o
wn
in
E
x
am
p
le
5
o
f
a
p
h
is
h
in
g
UR
L
f
r
o
m
o
u
r
d
ataset.
Ph
is
h
in
g
UR
L
s
o
f
ten
co
n
tain
m
u
ltip
le
p
r
o
to
c
o
ls
in
a
s
in
g
le
UR
L
,
u
n
lik
e
leg
it
im
ate
o
n
es.
As
s
h
o
wn
in
E
x
am
p
le
6
,
a
p
h
is
h
in
g
UR
L
f
r
o
m
o
u
r
d
ataset,
th
e
attac
k
er
s
u
s
e
HT
T
PS
th
r
ee
ti
m
es
in
a
s
in
g
le
UR
L
.
T
h
er
ef
o
r
e,
f
ea
tu
r
es
3
0
F
an
d
3
1
F
m
ea
s
u
r
e
th
e
o
cc
u
r
r
e
n
ce
s
o
f
HT
T
P
an
d
HT
T
PS
p
r
o
to
co
ls
in
UR
L
s
.
T
o
en
s
u
r
e
th
at
th
ese
f
ea
tu
r
es
ar
e
ex
tr
ac
ted
ac
c
u
r
ately
,
we
u
s
ed
t
h
e
co
m
p
lete
p
r
o
t
o
co
l
f
o
r
m
ats
(
'
h
ttp
://'
an
d
'
h
ttp
s
://
'
)
to
en
s
u
r
e
th
e
in
teg
r
ity
o
f
t
h
e
r
esu
lt
an
d
av
o
id
c
o
n
f
u
s
io
n
.
E
x
am
p
le
5
:
h
ttp
s
://www.
g
o
o
g
le.
co
m
/u
r
l?
q
=h
ttp
://www.
cu
s
to
m
er
-
p
o
r
tal.
i
n
f
o
:8
0
8
1
?r
i
d
%3
D9
2
3
6
0
Z
X%2
6
p
co
d
e%3
Dp
Z
tN
T
&
s
o
u
r
ce
=g
m
ail
-
im
a
p
&
u
s
t=1
6
6
4
9
5
3
3
7
7
0
0
0
0
0
0
&
u
s
g
=A
Ov
Vaw
0
v
p
jeHQBH1
n
m
l
-
Z
m
W
q
AQm
2
.
E
x
am
p
le
6
:
h
ttp
s
://www.
b
lo
g
g
er
.
co
m
/n
a
v
b
ar
.
g
?tar
g
etB
lo
g
I
D
\
x
3
d
5
3
7
3
5
8
2
2
8
1
5
4
2
0
8
1
4
9
9
\
x
2
6
b
lo
g
Na
m
e
\
x
3
d
ca
r
r
o
r
o
s
a
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
4
3
,
No
.
3
,
Sep
tem
b
er
20
2
6
:
9
0
8
-
9
2
7
916
b
r
ea
s
\
x
2
6
p
u
b
lis
h
Mo
d
e
\
x
3
d
PUB
L
I
SH_
MO
DE
_
B
L
OG
SP
OT
\
x
2
6
n
a
v
b
ar
T
y
p
e
\
x
3
d
B
L
UE
\
x
2
6
lay
o
u
tTy
p
e
\
x
3
d
C
L
ASSI
C
\
x
2
6
s
ea
r
ch
R
o
o
t
\
x
3
d
h
ttp
s
://car
r
o
r
o
s
ab
r
ea
s
.
b
lo
g
s
p
o
t.c
o
m
/s
ea
r
ch
\
x
2
6
b
lo
g
L
o
ca
le
\
x
3
d
p
t_
B
R
\
x26v
\
x3d2
\
x
2
6
h
o
m
ep
a
g
eUr
l
\
x
3
d
h
ttp
s
://car
r
o
r
o
s
ab
r
ea
s
.
b
lo
g
s
p
o
t.c
o
m
/
\
x
2
6
v
t
\
x
3
d
2
1
7
7
3
3
3
6
6
7
4
6
8
6
5
2
9
1
4
.
T
h
e
av
er
ag
e
wo
r
d
len
g
th
:
Fe
atu
r
e
6
5
F
ca
lcu
lates
th
e
av
er
ag
e
wo
r
d
len
g
th
i
n
th
e
h
o
s
tn
am
e
p
ar
t
o
f
UR
L
s
.
T
h
is
is
a
v
alu
ab
le
f
ea
tu
r
e
f
o
r
d
etec
tin
g
p
h
is
h
in
g
at
tack
s
.
I
t
h
elp
s
d
etec
t
an
o
m
ali
es
ass
o
ciate
d
with
p
h
is
h
in
g
web
s
ites
.
Stan
d
ar
d
d
ev
iatio
n
(
STD)
:
Me
asu
r
in
g
t
h
e
s
tan
d
ar
d
d
e
v
iatio
n
o
f
v
ar
ia
b
ilit
y
in
d
icate
s
h
o
w
f
ar
w
o
r
d
len
g
th
s
in
UR
L
s
d
e
v
iate
f
r
o
m
th
e
av
er
ag
e
.
I
n
o
u
r
r
esear
ch
,
we
ex
tr
ac
te
d
th
e
s
tan
d
ar
d
d
ev
iatio
n
in
two
way
s
.
T
h
e
f
ir
s
t
i
s
th
r
o
u
g
h
f
ea
tu
r
e
6
6
F,
wh
ich
m
ea
s
u
r
es
th
e
STD
o
f
wo
r
d
len
g
th
s
s
ep
ar
ated
b
y
d
o
ts
,
s
u
ch
as
d
o
m
ain
a
n
d
f
ile
n
am
es.
T
h
e
s
ec
o
n
d
f
ea
t
u
r
e
is
6
7
F,
wh
ic
h
m
ea
s
u
r
es
th
e
STD
o
f
wo
r
d
l
en
g
th
s
s
ep
ar
ated
b
y
th
e
s
lash
'
/
'
s
y
m
b
o
l,
s
u
ch
as
wo
r
d
s
in
th
e
p
ath
.
As
a
r
esu
lt
,
b
o
th
f
ea
tu
r
es
r
e
v
ea
led
u
n
u
s
u
al
n
u
m
er
ical
v
al
u
es
an
d
a
lar
g
e
d
is
p
er
s
io
n
in
p
h
is
h
in
g
UR
L
s
co
m
p
ar
ed
t
o
th
eir
le
g
itima
te
co
u
n
ter
p
ar
ts
.
UR
L
e
n
tr
o
p
y
:
T
h
e
f
ea
tu
r
e
7
0
F
m
ea
s
u
r
es
th
e
lev
el
o
f
r
a
n
d
o
m
n
ess
in
UR
L
co
n
ten
t.
T
h
is
f
ea
tu
r
e
is
im
p
o
r
tan
t
b
ec
au
s
e
m
an
y
p
h
is
h
in
g
UR
L
s
co
n
tain
o
b
f
u
s
ca
ted
ch
ar
ac
ter
s
to
d
ec
eiv
e
u
s
er
s
.
Dig
its
in
a
UR
L
:
Featu
r
e
6
8
F
ca
lcu
lates
th
e
r
atio
o
f
n
u
m
b
er
s
to
letter
s
in
a
h
o
s
tn
am
e,
wh
ich
h
el
p
s
d
etec
t
p
h
is
h
in
g
.
Ad
d
itio
n
ally
,
f
ea
tu
r
e
3
6
F
m
e
asu
r
es
th
e
n
u
m
er
ic
co
n
ten
t
in
UR
L
s
,
a
s
a
h
ig
h
n
u
m
b
er
o
f
n
u
m
b
er
s
in
UR
L
s
in
d
icate
s
a
p
h
is
h
in
g
attem
p
t.
B
ased
o
n
o
u
r
an
aly
s
is
o
f
th
e
p
h
is
h
in
g
d
ataset,
we
cr
ea
ted
f
ea
tu
r
e
3
3
F
to
co
u
n
t
th
e
n
u
m
b
er
o
f
ze
r
o
s
in
t
h
e
h
o
s
tn
am
e,
wh
ile
f
ea
tu
r
e
3
8
F c
o
u
n
ts
th
e
n
u
m
b
er
s
in
t
h
e
q
u
e
r
y
p
a
r
t.
UR
L
len
g
th
:
L
eg
itima
te
UR
L
s
r
ely
o
n
s
h
o
r
t
h
o
s
tn
am
es
to
m
ak
e
th
em
ea
s
y
to
u
s
e.
I
n
co
n
tr
ast,
attac
k
er
s
u
s
e
lo
n
g
h
o
s
tn
am
es
to
co
n
f
u
s
e
th
eir
tar
g
ets.
T
h
e
r
ef
o
r
e,
we
cr
ea
ted
f
ea
tu
r
e
6
9
F,
wh
ich
ca
lcu
lates
th
e
len
g
th
o
f
th
e
h
o
s
tn
am
e
to
i
d
en
tify
p
h
is
h
in
g
a
d
d
r
ess
es
with
o
u
t
s
et
tin
g
a
th
r
esh
o
ld
.
B
ec
au
s
e
s
p
ec
if
y
in
g
a
th
r
esh
o
ld
m
ay
ca
u
s
e
m
is
class
if
icatio
n
[
3
3
]
.
Als
o
,
u
s
in
g
l
o
n
g
wo
r
d
s
in
t
h
e
p
ath
p
ar
t
is
a
c
o
m
m
o
n
tech
n
iq
u
e
i
n
p
h
is
h
in
g
attac
k
s
f
o
r
o
b
f
u
s
ca
tio
n
p
u
r
p
o
s
es.
Acc
o
r
d
in
g
l
y
,
f
ea
tu
r
e
7
1
F
m
ea
s
u
r
es
th
e
len
g
th
o
f
th
e
lo
n
g
est
wo
r
d
in
th
e
UR
L
'
s
p
ath
.
I
t
is
also
im
p
o
r
tan
t
to
m
ea
s
u
r
e
th
e
le
n
g
th
o
f
th
e
en
tire
UR
L
.
B
ased
o
n
o
u
r
d
ataset
an
aly
s
is
,
we
f
o
u
n
d
th
at
th
e
lo
n
g
est
p
h
is
h
in
g
UR
L
ex
ten
d
s
to
6
0
9
7
c
h
ar
ac
ter
s
.
Sin
ce
we
r
ely
o
n
b
r
o
wser
r
esp
o
n
s
es
to
leg
itima
te
UR
L
s
,
th
e
lo
n
g
est
leg
itima
t
e
UR
L
ex
ten
d
s
to
1
2
7
6
ch
a
r
ac
ter
s
.
T
h
er
ef
o
r
e
,
we
cr
ea
ted
f
ea
t
u
r
e
7
3
F
to
ca
lcu
late
th
e
len
g
th
o
f
UR
L
s
with
o
u
t
s
ettin
g
a
p
r
e
d
ef
in
ed
class
if
icati
o
n
th
r
esh
o
ld
.
C
o
n
tr
a
r
y
to
s
tu
d
ies
[
1
2
]
,
[
3
3
]
,
[
3
4
]
,
an
d
[
3
5
]
,
wh
ich
s
et
d
if
f
er
en
t
th
r
esh
o
ld
v
alu
es
f
o
r
UR
L
len
g
th
s
,
th
e
f
o
llo
win
g
v
alu
es
ar
e
u
s
ed
:
5
4
,
2
0
0
,
7
6
,
an
d
7
4
ch
a
r
ac
ter
s
,
r
esp
ec
tiv
el
y
.
E
x
ce
e
d
in
g
th
ese
v
alu
es
is
co
n
s
id
er
ed
p
h
is
h
in
g
.
T
h
ese
d
if
f
er
en
ce
s
in
cr
iter
ia
h
ig
h
lig
h
t
th
at
th
er
e
is
n
o
u
n
if
o
r
m
s
tan
d
ar
d
f
o
r
ev
al
u
atin
g
UR
L
len
g
th
s
to
d
eter
m
in
e
wh
eth
er
th
ey
ar
e
p
h
is
h
in
g
.
Ou
r
ap
p
r
o
ac
h
h
elp
s
d
is
tin
g
u
is
h
b
etwe
en
le
g
itima
te
an
d
p
h
is
h
in
g
UR
L
s
b
ased
o
n
th
eir
d
if
f
e
r
en
t
len
g
th
s
,
wh
ich
ar
e
class
if
ied
ac
co
r
d
in
g
to
th
e
m
o
d
el'
s
p
r
ed
ictio
n
.
I
PF
S:
T
h
e
I
PF
S
p
r
o
to
co
l
is
u
s
ed
f
o
r
d
ec
en
tr
alize
d
f
ile
u
p
lo
a
d
in
g
,
s
h
ar
i
n
g
,
a
n
d
ac
ce
s
s
[
3
6
]
.
Ph
is
h
in
g
web
s
ites
h
o
s
ted
o
n
I
PF
S
ar
e
s
im
ilar
to
tr
ad
itio
n
al
p
h
is
h
in
g
web
s
ites
ex
ce
p
t
th
at
th
ey
u
s
e
a
h
ash
v
alu
e
f
o
r
th
e
co
n
ten
t
in
s
tead
o
f
a
d
o
m
ain
n
a
m
e,
m
ak
in
g
it d
if
f
icu
lt to
tr
ac
e
th
e
s
o
u
r
ce
o
r
v
e
r
if
y
th
e
p
u
b
lis
h
er
'
s
id
en
tity
.
Ou
r
an
aly
s
is
o
f
o
u
r
d
ataset
r
e
v
ea
led
th
at
attac
k
er
s
u
s
e
d
o
u
b
le
e
n
co
d
in
g
o
f
I
PF
S
-
h
o
s
ted
web
p
a
g
e
UR
L
s
to
b
y
p
ass
d
etec
tio
n
s
y
s
tem
s
,
as
illu
s
tr
ated
in
E
x
am
p
le
7
.
T
h
er
ef
o
r
e,
f
ea
tu
r
e
4
7
F
d
etec
ts
th
e
p
r
es
en
ce
o
f
a
web
s
ite
h
o
s
ted
o
n
I
PF
S
an
d
a
n
aly
s
es
its
UR
L
en
co
d
in
g
lev
el.
Do
u
b
le
-
en
co
d
e
d
UR
L
s
ar
e
class
if
ie
d
as
p
h
is
h
in
g
,
wh
ile
u
n
en
co
d
ed
o
n
es a
r
e
class
if
ied
as su
s
p
icio
u
s
; o
th
er
wis
e,
th
ey
ar
e
class
if
ied
as leg
itima
te.
T
h
e
liter
atu
r
e
h
as n
o
t
ex
ten
s
iv
ely
an
aly
ze
d
th
is
f
ea
t
u
r
e
f
o
r
d
etec
tin
g
p
h
is
h
in
g
atta
ck
s
,
a
cr
u
cial
asp
ec
t
th
at
wa
r
r
an
ts
f
u
r
th
er
r
esear
ch
to
m
itig
ate
p
h
is
h
in
g
t
h
r
ea
ts
.
E
x
am
p
le
7
:
h
ttp
s
://b
z9
0
7
.
in
f
u
r
a
-
ip
f
s
.
io
/ip
f
s
/Qm
W
S4
Mu
PNGm
KFeVh
4
v
Gn
4
d
E
SX3
B
7
b
zm
P
HwM
b
t1
Mh
Sh
G6
p
Q?g
ab
r
ie%
2
5
0
0
%2
5
1
Dv
%2
5
0
0
%2
5
1
0
%2
5
2
6
%2
5
2
3
1
3
1
%2
5
3
B
%2
5
1
0
Q%2
5
2
6
%2
5
2
3
1
3
5
%2
5
3
B
%2
5
2
0
%2
5
2
6
%2
5
2
3
1
4
6
%2
5
3
B
%2
5
2
6
%2
5
2
3
1
3
9
%
FHD:
Fre
e
s
er
v
ice
p
r
o
v
id
er
s
p
r
o
v
id
e
th
eir
u
s
er
s
with
s
ec
u
r
e
s
o
ck
ets lay
er
(
SS
L
)
ce
r
tific
ates a
n
d
tr
u
s
ted
T
L
Ds
s
u
ch
as
'
co
m
'
,
wh
ich
en
ab
les
attac
k
er
s
to
b
y
p
ass
d
etec
tio
n
m
ec
h
an
is
m
s
ea
s
ily
.
W
e
d
ev
elo
p
ed
a
f
ea
tu
r
e
5
7
F
t
o
d
etec
t
p
h
is
h
in
g
UR
L
s
h
o
s
ted
o
n
FHDs
an
d
as
s
o
ciate
d
w
ith
s
en
s
itiv
e
wo
r
d
s
.
I
n
o
u
r
r
esear
ch
,
we
co
llected
a
lis
t
o
f
2
6
FHDs
d
er
iv
ed
f
r
o
m
th
e
s
tu
d
y
[
3
7
]
an
d
o
u
r
d
atas
et.
W
e
ca
r
ef
u
lly
ex
tr
ac
ted
FHDs
to
av
o
id
f
alse
p
o
s
itiv
es
b
y
in
clu
d
in
g
th
e
T
L
Ds
th
at
f
r
ee
h
o
s
tin
g
s
er
v
ices
p
r
o
v
id
e
to
th
eir
u
s
er
s
,
in
d
ica
tin
g
th
at
it
is
in
d
ee
d
f
r
ee
h
o
s
tin
g
.
Fo
r
ex
a
m
p
le,
X
1
0
h
o
s
tin
g
o
f
f
er
s
f
r
ee
h
o
s
tin
g
s
er
v
ices,
s
u
ch
as
elem
en
tf
x
.
co
m
,
x
1
0
.
m
x
,
an
d
x
1
0
.
b
z
.
Als
o
,
we
d
etec
ted
1
7
4
s
en
s
itiv
e
k
ey
wo
r
d
s
in
v
ar
i
o
u
s
lan
g
u
ag
es
r
elate
d
to
f
r
ee
h
o
s
tin
g
,
in
clu
d
in
g
Fre
n
ch
,
I
talian
,
E
n
g
lis
h
,
T
u
r
k
is
h
,
I
n
d
o
n
esian
,
Ger
m
an
,
a
n
d
o
th
er
s
.
E
x
am
p
le
8
,
o
b
s
er
v
ed
v
ia
o
u
r
d
ataset
an
aly
s
is
,
s
h
o
ws
a
UR
L
h
id
in
g
th
e
Fre
n
ch
p
h
r
ase
"p
ay
ez
v
o
tr
e
f
ac
tu
r
e"
,
wh
ich
m
ea
n
s
(
p
ay
y
o
u
r
b
ill,
in
E
n
g
lis
h
)
.
Fu
r
th
er
m
o
r
e,
attac
k
er
s
u
s
e
th
e
n
am
es
o
f
g
o
v
e
r
n
m
en
t
en
titi
es
to
tr
ic
k
u
s
er
s
i
n
to
r
e
v
ea
lin
g
th
eir
cr
ed
en
tials
.
E
x
am
p
le
9
,
f
r
o
m
o
u
r
d
ataset,
s
h
o
ws
a
UR
L
im
p
er
s
o
n
atin
g
th
e
Sau
d
i
Min
is
tr
y
o
f
Fin
a
n
ce
(
m
o
f
.
g
o
v
.
s
a)
o
n
f
r
ee
h
o
s
tin
g
'
f
ir
eb
aseap
p
.
co
m
'
.
E
x
am
p
le
8
:
h
ttp
s
://s
ites
.
g
o
o
g
le.
co
m
/v
ie
w/p
ay
-
ez
-
vo
-
tr
e
-
fa
-
ctu
r
e
.
E
x
am
p
le
9
:
h
ttp
s
://mo
f
-
g
o
v
.
f
ir
eb
aseap
p
.
co
m
/
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
UR
L
-
b
a
s
ed
p
h
is
h
in
g
d
etec
tio
n
u
s
in
g
X
GB
o
o
s
t wi
th
en
g
in
ee
r
ed
fea
tu
r
es
(
Ja
w
a
h
er A
lh
a
r
b
i
)
917
3
.
3
.
Appl
ied m
a
chine le
a
rning
(
M
L
)
a
lg
o
rit
hm
s
T
h
is
s
ec
tio
n
r
ev
iews
th
e
m
o
d
els
ad
o
p
ted
in
th
is
s
tu
d
y
.
As
s
h
o
wn
in
Fig
u
r
e
1
,
we
e
v
alu
a
ted
s
ev
en
ML
class
if
ier
s
.
T
h
e
test
ed
alg
o
r
ith
m
s
in
clu
d
ed
s
tan
d
alo
n
e
class
if
ier
s
,
s
u
ch
as
a
d
ec
is
io
n
tr
ee
[
3
8
]
a
n
d
KNN
[
3
9
]
.
Ad
d
itio
n
ally
,
we
ev
alu
ated
s
ev
er
al
lear
n
in
g
m
eth
o
d
s
,
in
clu
d
i
n
g
en
s
em
b
le
lear
n
in
g
,
s
u
ch
as
RF
[
4
0
]
,
Ad
aBo
o
s
t
[
4
1
]
,
XGBo
o
s
t
[
4
2
]
,
an
d
L
ig
h
tGB
M
[
4
3
]
.
E
n
s
em
b
le
lear
n
in
g
tec
h
n
iq
u
e
s
o
f
ten
o
u
tp
e
r
f
o
r
m
in
d
iv
id
u
al
ML
m
o
d
els
b
y
co
m
b
in
in
g
p
r
e
d
ictio
n
s
f
r
o
m
m
u
ltip
le
m
o
d
els,
th
u
s
a
d
d
r
ess
in
g
th
e
lim
itatio
n
s
o
f
s
in
g
le
ML
m
o
d
els
an
d
le
v
er
a
g
in
g
t
h
eir
s
tr
en
g
th
s
to
im
p
r
o
v
e
class
if
icatio
n
p
er
f
o
r
m
a
n
ce
[
4
4
]
.
W
e
also
tr
ain
e
d
an
ML
P
[
4
5
]
c
o
n
s
is
tin
g
o
f
an
in
p
u
t
la
y
er
,
o
n
e
o
r
m
o
r
e
h
id
d
en
lay
e
r
s
,
an
d
an
o
u
tp
u
t
lay
er
.
Ad
d
itio
n
ally
,
it
r
elies
o
n
n
o
n
-
lin
ea
r
f
u
n
ctio
n
s
,
s
u
ch
as
R
eL
U
an
d
Sig
m
o
id
,
to
p
r
o
ce
s
s
d
ata
an
d
ex
tr
a
ct
p
atter
n
s
.
I
n
o
u
r
ex
p
er
im
en
ts
,
th
e
m
o
d
els
ex
h
ib
it
v
ar
y
in
g
lev
els
o
f
p
r
ed
icti
o
n
ac
cu
r
ac
y
.
B
elo
w,
we
ex
p
lain
th
e
ev
alu
atio
n
m
etr
ics.
3
.
3
.
1
.
E
v
a
lua
t
io
n pa
ra
m
et
er
s
Af
ter
tr
ain
in
g
th
e
m
o
d
els,
p
er
f
o
r
m
a
n
ce
m
etr
ics
wer
e
ap
p
lied
to
d
eter
m
in
e
th
e
b
est
-
p
er
f
o
r
m
in
g
m
o
d
el.
T
h
e
m
ain
p
er
f
o
r
m
an
c
e
m
etr
ics
in
clu
d
e
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
F1
-
s
co
r
e,
an
d
d
etec
tio
n
tim
e.
T
h
e
ev
alu
atio
n
m
etr
ic
s
ar
e
b
ased
o
n
a
co
n
f
u
s
io
n
m
atr
ix
,
wh
ich
s
u
m
m
ar
izes
co
r
r
ec
tly
a
n
d
in
co
r
r
ec
tly
p
r
ed
icted
ca
s
es.
T
h
e
co
n
f
u
s
io
n
m
atr
ix
h
as f
o
u
r
o
u
tco
m
es:
−
T
r
u
e
p
o
s
itiv
e
(
T
P):
T
h
e
to
tal
n
u
m
b
er
o
f
p
h
is
h
in
g
UR
L
s
co
r
r
e
ctly
class
if
ied
as p
h
is
h
in
g
.
−
T
r
u
e
n
e
g
ativ
e
(
T
N)
: T
h
e
to
tal
n
u
m
b
er
o
f
le
g
itima
te
UR
L
s
co
r
r
ec
tly
class
if
ied
as leg
itima
te.
−
Fals
e
p
o
s
itiv
e
(
FP
)
: T
h
e
to
tal
n
u
m
b
er
o
f
le
g
itima
te
UR
L
s
in
co
r
r
ec
tly
class
if
ied
as p
h
is
h
in
g
.
−
Fals
e
n
eg
ativ
e
(
FN)
: T
h
e
to
tal
n
u
m
b
e
r
o
f
p
h
is
h
in
g
UR
L
s
in
co
r
r
ec
tly
class
if
ied
as leg
itima
te.
Fro
m
th
ese
o
u
tco
m
es,
p
e
r
f
o
r
m
an
ce
m
etr
ics
ca
n
b
e
ca
lcu
la
ted
to
d
eter
m
in
e
t
h
e
b
est
-
p
er
f
o
r
m
in
g
m
o
d
el.
T
h
e
f
o
llo
win
g
ar
e
th
e
m
ain
p
er
f
o
r
m
an
ce
m
etr
ics:
Acc
u
r
ac
y
: Ratio
o
f
co
r
r
ec
t p
r
e
d
ictio
n
s
to
th
e
to
tal
n
u
m
b
er
o
f
p
r
ed
ictio
n
s
,
as in
(
1
)
A
c
c
ura
c
y
=
TP
+
TN
TP
+
TN
+
FN
+
FP
(
1
)
Pre
cisi
o
n
: Ratio
o
f
co
r
r
ec
tly
p
r
ed
icted
p
o
s
itiv
e
o
b
s
er
v
atio
n
s
(
UR
L
s
d
etec
ted
as p
h
is
h
in
g
)
t
o
th
e
to
tal
p
r
ed
icted
p
o
s
itiv
es (
p
h
is
h
in
g
UR
L
s
)
,
as in
(
2
)
Pr
e
c
ision
=
TP
+
FP
(
2
)
R
ec
all: Ra
tio
o
f
co
r
r
ec
tly
p
r
ed
icted
p
h
is
h
in
g
UR
L
s
to
all
ac
tu
al
p
h
is
h
in
g
UR
L
s
,
as in
(
3
)
R
e
c
a
l
l
=
TP
+
FN
(
3
)
F1
-
s
co
r
e:
Har
m
o
n
ic
m
ea
n
o
f
p
r
ec
is
io
n
an
d
r
ec
all
b
alan
cin
g
b
o
th
r
ec
all
an
d
p
r
ec
is
io
n
,
as in
(
4
)
F1
=
2
∗
Pr
ecis
i
o
n
∗
Recal
l
Pr
ecis
i
o
n
+
Recal
l
(
4
)
Dete
ctio
n
tim
e:
T
h
e
tim
e
r
eq
u
ir
ed
to
class
if
y
a
p
h
is
h
in
g
UR
L
is
m
ea
s
u
r
ed
in
m
illi
s
ec
o
n
d
s
o
r
s
ec
o
n
d
s
,
in
d
icatin
g
m
o
d
el
ef
f
icien
cy
.
R
ec
eiv
er
o
p
e
r
atin
g
c
h
ar
ac
ter
is
tic
ar
ea
u
n
d
er
th
e
cu
r
v
e
(
R
OC
-
AUC):
T
o
ev
alu
ate
th
e
m
o
d
el'
s
ab
ilit
y
to
d
is
tin
g
u
is
h
leg
itima
te
UR
L
s
f
r
o
m
p
h
is
h
in
g
UR
L
s
ac
r
o
s
s
d
if
f
er
en
t
class
if
icatio
n
th
r
esh
o
ld
s
.
4.
E
XP
E
R
I
M
E
N
T
S
E
T
UP
A
N
D
RE
SU
L
T
S
T
h
is
s
ec
tio
n
p
r
esen
ts
th
e
e
x
p
er
im
en
tal
s
etu
p
an
d
p
er
f
o
r
m
an
ce
r
esu
lts
o
f
all
ev
alu
at
ed
m
o
d
els,
f
o
llo
wed
b
y
an
a
b
latio
n
s
tu
d
y
to
ev
alu
ate
th
e
e
f
f
ec
t o
f
f
ea
tu
r
e
s
elec
tio
n
o
n
th
e
p
r
o
p
o
s
ed
X
GB
o
o
s
t m
o
d
el.
4
.
1
.
E
x
perim
ent
a
l set
up
T
o
o
p
tim
ize
m
o
d
el
p
e
r
f
o
r
m
a
n
ce
,
we
u
s
ed
Gr
id
Sear
ch
C
V
with
3
-
f
o
ld
cr
o
s
s
-
v
alid
atio
n
[
4
6
]
o
n
th
e
tr
ain
in
g
s
et
f
o
r
th
e
tr
ad
itio
n
al
ML
class
if
ier
s
to
id
en
tify
o
p
ti
m
al
p
ar
am
eter
s
.
T
h
e
o
p
tim
al
h
y
p
er
p
ar
am
eter
s
f
o
r
ea
ch
m
o
d
el
wer
e
s
elec
te
d
b
a
s
ed
o
n
th
e
h
ig
h
est
cr
o
s
s
-
v
alid
atio
n
ac
c
u
r
ac
y
.
No
tab
ly
,
w
e
ex
ec
u
ted
s
ev
er
al
ex
p
er
im
en
ts
with
th
e
KNN
m
o
d
el,
wh
e
r
e
its
p
e
r
f
o
r
m
an
ce
d
ep
en
d
s
o
n
t
h
e
n
u
m
b
e
r
o
f
n
ea
r
est
n
eig
h
b
o
r
s
(
k
)
.
Usi
n
g
o
d
d
n
u
m
b
er
s
f
o
r
(
k
)
is
h
elp
f
u
l
f
o
r
b
in
ar
y
class
if
icatio
n
d
atasets
,
as
it
p
r
ev
en
ts
tie
v
o
tes
b
etwe
en
n
eig
h
b
o
r
s
an
d
en
s
u
r
es
th
at
th
e
m
ajo
r
ity
d
ec
is
io
n
is
clea
r
ly
d
ef
in
ed
[
4
7
]
.
T
h
er
ef
o
r
e,
we
te
s
ted
k
=
1
,
3
,
an
d
5
Evaluation Warning : The document was created with Spire.PDF for Python.