I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
pu
t
er
E
ng
ineering
(
I
J
E
CE
)
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
25
,
p
p
.
3494
~
3
5
0
5
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
15
i
3
.
pp
3
4
9
4
-
3
5
0
5
3494
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
Butter
fly o
ptimiz
a
tion
-
ba
sed ense
mble lea
rning
s
tr
a
tegy
f
o
r
a
dv
a
nced int
rusio
n det
e
ction in
int
e
rnet
o
f
t
hing
s
net
wo
rks
M
o
ua
d Cho
uk
ha
iri,
Sa
ra
T
a
hiri,
O
ua
il Cho
uk
ha
iri,
Yo
us
s
ef
F
a
k
hri,
M
o
ha
m
ed
Am
n
a
i
LA
R
I
,
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
,
F
a
c
u
l
t
y
o
f
S
c
i
e
n
c
e
s,
I
b
n
T
o
f
a
i
l
U
n
i
v
e
r
s
i
t
y
,
K
e
n
i
t
r
a
,
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
Au
g
2
0
,
2
0
2
4
R
ev
is
ed
Ma
r
3
,
2
0
2
5
Acc
ep
ted
Ma
r
2
0
,
2
0
2
5
Th
e
m
a
ss
iv
e
g
ro
wt
h
i
n
i
n
tern
e
t
o
f
th
i
n
g
s
(I
o
T)
d
e
v
ice
s
h
a
s
led
to
e
n
h
a
n
c
e
d
fu
n
c
ti
o
n
a
li
ti
e
s
th
r
o
u
g
h
t
h
e
ir
in
t
e
rc
o
n
n
e
c
ti
o
n
s
with
o
t
h
e
r
d
e
v
ic
e
s,
sm
a
rt
in
fra
stru
c
tu
re
s,
a
n
d
n
e
two
r
k
s.
Ho
we
v
e
r,
i
n
c
re
a
se
d
c
o
n
n
e
c
t
iv
it
y
a
lso
in
c
re
a
se
s
th
e
risk
o
f
c
y
b
e
ra
tt
a
c
k
s.
To
p
r
o
tec
t
Io
T
sy
ste
m
s
fr
o
m
t
h
e
se
th
re
a
ts,
in
tru
si
o
n
d
e
tec
ti
o
n
sy
ste
m
s
(I
DS)
e
m
p
lo
y
in
g
m
a
c
h
in
e
lea
rn
in
g
(
M
L)
tec
h
n
iq
u
e
s
h
a
v
e
b
e
e
n
d
e
v
e
l
o
p
e
d
t
o
id
e
n
t
ify
c
y
b
e
rse
c
u
rit
y
t
h
re
a
ts.
Th
is
p
a
p
e
r
in
tro
d
u
c
e
s
a
n
o
v
e
l
e
n
se
m
b
le
IDS
fra
m
e
wo
rk
c
a
ll
e
d
b
u
tt
e
rfl
y
o
p
t
imiz
a
ti
o
n
-
ba
se
d
e
n
se
m
b
le
lea
rn
in
g
(BOEL
).
Th
is
fra
m
e
wo
rk
in
teg
ra
tes
th
e
b
u
tt
e
rfl
y
o
p
ti
m
iza
ti
o
n
a
l
g
o
ri
th
m
(BOA
)
with
e
n
se
m
b
le
lea
rn
in
g
tec
h
n
iq
u
e
s
to
imp
ro
v
e
IDS
d
e
tec
ti
o
n
p
e
rfo
rm
a
n
c
e
in
Io
T
n
e
two
rk
s.
BOEL
is
d
e
sig
n
e
d
to
a
c
c
u
ra
tely
d
e
tec
t
v
a
rio
u
s
ty
p
e
s
o
f
a
tt
a
c
k
s
in
Io
T
n
e
two
r
k
s
b
y
d
y
n
a
m
ica
ll
y
o
p
ti
m
izi
n
g
t
h
e
we
ig
h
ts o
f
b
a
se
le
a
rn
e
rs,
wh
ich
a
re
th
e
fo
u
r
so
p
h
ist
ica
ted
M
L
g
ra
d
ien
t
-
b
o
o
stin
g
a
lg
o
rit
h
m
s
(GBM,
Ca
tBo
o
st
,
XG
Bo
o
st,
a
n
d
L
ig
h
tG
BM
)
fo
r
e
a
c
h
a
tt
a
c
k
c
a
teg
o
ry
,
a
n
d
i
d
e
n
ti
fy
i
n
g
t
h
e
b
e
st
we
ig
h
t
c
o
m
b
i
n
a
ti
o
n
fo
r
e
n
se
m
b
le
m
o
d
e
ls.
E
x
p
e
rime
n
ts
c
o
n
d
u
c
ted
o
n
two
p
u
b
li
c
I
o
T
se
c
u
rit
y
d
a
tas
e
ts,
CICIDS2
0
1
7
a
n
d
B
o
t
-
Io
T,
d
e
m
o
n
stra
te
th
e
r
o
b
u
stn
e
ss
o
f
th
e
p
ro
p
o
se
d
BOEL
in
i
n
tru
si
o
n
d
e
tec
ti
o
n
a
c
ro
ss
d
iv
e
rse
I
o
T
e
n
v
i
ro
n
m
e
n
ts
,
a
c
h
iev
in
g
9
9
.
7
9
5
%
a
c
c
u
ra
c
y
o
n
CICIDS2
0
1
7
a
n
d
9
9
.
9
6
6
%
a
c
c
u
ra
c
y
o
n
Bo
t
-
Io
T
.
Th
e
se
re
su
lt
s
o
u
tl
in
e
t
h
e
su
c
c
e
ss
fu
l
a
p
p
li
c
a
ti
o
n
o
f
d
iv
e
rs
e
lea
rn
in
g
a
p
p
ro
a
c
h
e
s
a
n
d
h
ig
h
li
g
h
t
t
h
e
fra
m
e
wo
rk
’s
p
o
ten
ti
a
l
t
o
e
n
h
a
n
c
e
IDS
in
a
d
d
re
ss
in
g
I
o
T
c
y
b
e
r
t
h
re
a
ts
.
K
ey
w
o
r
d
s
:
B
u
tter
f
ly
o
p
tim
izatio
n
alg
o
r
ith
m
C
y
b
er
s
ec
u
r
ity
E
n
s
em
b
le
lear
n
in
g
Gr
ad
ien
t
-
b
o
o
s
tin
g
I
n
ter
n
et
o
f
th
in
g
s
I
n
tr
u
s
io
n
d
etec
tio
n
s
y
s
tem
Ma
ch
in
e
lear
n
in
g
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
:
Mo
u
ad
C
h
o
u
k
h
air
i
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
,
Facu
lty
o
f
Scien
ce
s
,
I
b
n
T
o
f
ail
Un
iv
er
s
ity
B
.
P 1
3
3
,
Un
iv
er
s
ity
C
am
p
u
s
,
Ken
itra
,
Mo
r
o
cc
o
E
m
ail:
m
o
u
ad
.
c
h
o
u
k
h
air
i@
u
it.a
c.
m
a
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
ex
p
o
n
en
tial
g
r
o
wth
o
f
in
t
er
n
et
o
f
th
in
g
s
(
I
o
T
)
d
ev
ices
a
n
d
n
etwo
r
k
s
h
as
u
s
h
er
ed
in
a
n
ew
er
a
o
f
co
n
n
ec
tiv
ity
,
b
u
t
it
h
as
also
l
ed
to
a
n
alar
m
i
n
g
i
n
cr
ea
s
e
in
cy
b
er
th
r
ea
ts
an
d
s
o
p
h
is
ticated
attac
k
s
[
1
]
.
I
o
T
n
etwo
r
k
s
p
r
esen
t
u
n
iq
u
e
ch
allen
g
es
f
o
r
s
ec
u
r
ity
im
p
lem
en
ta
tio
n
s
d
u
e
to
th
eir
h
eter
o
g
e
n
eo
u
s
n
atu
r
e,
r
eso
u
r
c
e
co
n
s
tr
ain
ts
,
an
d
lar
g
e
-
s
ca
le
d
e
p
lo
y
m
en
ts
,
wh
ich
co
llectiv
ely
cr
ea
te
a
v
ast
attac
k
s
u
r
f
ac
e
f
o
r
m
alicio
u
s
ac
to
r
s
to
ex
p
lo
it
[
2
]
.
T
r
a
d
itio
n
al
a
p
p
r
o
ac
h
es
to
in
t
r
u
s
io
n
d
etec
tio
n
s
y
s
tem
s
(
I
DS)
d
esig
n
e
d
f
o
r
co
n
v
en
tio
n
al
co
m
p
u
ter
n
etwo
r
k
s
ar
e
o
f
ten
i
n
s
u
f
f
icien
t
to
m
ee
t
th
e
d
is
tin
ct
ch
allen
g
es
p
o
s
ed
b
y
I
o
T
ec
o
s
y
s
tem
s
[
3
]
.
T
h
es
e
ch
allen
g
es
in
clu
d
e
th
e
n
ee
d
f
o
r
ac
cu
r
ate
d
etec
tio
n
,
h
a
n
d
lin
g
o
f
d
iv
er
s
e
d
ata
ty
p
es,
an
d
a
d
ap
tatio
n
to
r
ap
i
d
ly
ev
o
lv
in
g
th
r
ea
t
lan
d
s
ca
p
es
wh
ile
o
p
er
atin
g
with
in
th
e
co
n
s
tr
ain
ts
o
f
lim
ited
co
m
p
u
tatio
n
al
r
eso
u
r
ce
s
ty
p
ical
o
f
I
o
T
d
e
v
ices
[
4
]
.
C
o
n
s
eq
u
e
n
tly
,
ad
v
an
ce
d
,
ad
ap
tiv
e
,
an
d
ef
f
icien
t
I
DS
ar
e
cr
u
cial
s
ec
u
r
ity
p
r
ec
au
tio
n
s
f
o
r
th
is
in
cr
ea
s
e
in
cy
b
er
th
r
ea
ts
i
n
I
o
T
en
v
ir
o
n
m
en
ts
[
5
]
.
E
x
ten
s
iv
e
r
esear
ch
h
as
f
o
c
u
s
ed
o
n
e
n
h
an
cin
g
in
tr
u
s
io
n
d
etec
tio
n
ca
p
a
b
ilit
ies
in
I
o
T
n
etwo
r
k
s
,
with
m
ac
h
in
e
lea
r
n
i
n
g
(
ML
)
a
p
p
r
o
ac
h
es
g
ain
i
n
g
s
ig
n
if
ican
t
tr
ac
tio
n
.
T
h
ese
tech
n
iq
u
es
le
v
er
ag
e
th
e
ir
ab
ilit
y
to
a
n
aly
ze
la
r
g
e
d
atasets
an
d
id
en
tify
p
atter
n
s
,
s
ig
n
if
ican
tly
im
p
r
o
v
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
B
u
tter
fly
o
p
timiz
a
tio
n
-
b
a
s
ed
e
n
s
emb
le
lea
r
n
in
g
s
tr
a
teg
y
fo
r
a
d
va
n
ce
d
…
(
Mo
u
a
d
C
h
o
u
kh
a
ir
i
3495
I
o
T
n
etwo
r
k
d
ef
e
n
s
es.
T
h
u
s
,
th
ey
im
p
r
o
v
e
I
DS
b
y
r
ed
u
cin
g
f
alse
p
o
s
itiv
es
an
d
i
n
cr
ea
s
in
g
ac
cu
r
ac
y
.
C
h
u
r
ch
er
et
a
l.
[
6
]
co
n
d
u
cted
a
co
m
p
ar
is
o
n
o
f
s
ev
en
m
ac
h
in
e
lear
n
in
g
alg
o
r
ith
m
s
f
o
r
in
tr
u
s
io
n
d
etec
tio
n
i
n
I
o
T
n
etwo
r
k
s
u
s
in
g
t
h
e
B
o
t
-
I
o
T
d
ataset.
T
h
ey
f
o
u
n
d
th
at
r
an
d
o
m
f
o
r
est
(
R
F)
p
er
f
o
r
m
ed
b
est
in
b
in
ar
y
class
if
icatio
n
,
wh
ile
k
-
n
ea
r
est
n
eig
h
b
o
r
s
(
KNN)
ex
ce
lled
i
n
m
u
lti
-
class
class
if
icatio
n
with
9
9
%
ac
cu
r
ac
y
.
C
h
o
u
k
h
air
i
et
a
l.
[
7
]
p
r
o
p
o
s
ed
a
tr
ee
-
s
tr
u
ctu
r
e
d
ML
-
b
ased
I
DS
f
o
r
en
h
a
n
ce
d
s
ec
u
r
ity
in
I
o
T
n
etwo
r
k
s
,
ef
f
ec
tiv
ely
d
etec
tin
g
d
iv
er
s
e
cy
b
er
attac
k
s
with
h
ig
h
ac
cu
r
a
cy
an
d
lo
w
co
m
p
u
tatio
n
al
co
s
ts
o
n
th
e
UNS
W
-
NB
1
5
d
ataset.
Z
h
an
g
et
a
l.
[
8
]
d
ev
elo
p
ed
a
two
-
s
tag
e
in
tr
u
s
io
n
d
etec
tio
n
m
o
d
el
f
o
r
I
o
T
n
etwo
r
k
s
,
u
s
in
g
a
lig
h
t
g
r
ad
ien
t
-
b
o
o
s
tin
g
m
ac
h
in
e
(
L
ig
h
tGB
M)
f
o
r
in
itial
t
r
af
f
ic
class
if
icatio
n
an
d
a
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
(
C
NN)
f
o
r
d
etailed
attac
k
id
en
tific
atio
n
.
T
h
eir
m
o
d
el,
test
ed
o
n
th
e
C
S
E
-
C
I
C
-
I
DS
2
0
1
8
d
ataset,
s
h
o
wed
s
u
p
er
i
o
r
p
er
f
o
r
m
an
c
e
in
h
an
d
lin
g
im
b
alan
ce
d
,
l
ar
g
e
-
s
ca
le
n
etwo
r
k
d
ata
co
m
p
ar
ed
to
e
x
is
tin
g
s
y
s
tem
s
.
Z
h
ao
et
a
l.
[
9
]
d
ev
el
o
p
ed
a
lig
h
tweig
h
t
n
etwo
r
k
in
tr
u
s
io
n
d
etec
tio
n
m
eth
o
d
f
o
r
I
o
T
u
s
in
g
p
r
in
cip
al
co
m
p
o
n
en
t
a
n
aly
s
is
(
PC
A)
f
o
r
d
im
en
s
io
n
ality
r
ed
u
ctio
n
a
n
d
a
cu
s
to
m
n
e
u
r
al
n
etwo
r
k
ar
ch
itectu
r
e.
T
h
eir
ap
p
r
o
ac
h
ac
h
iev
ed
h
ig
h
class
i
f
icatio
n
p
er
f
o
r
m
an
ce
with
lo
w
co
m
p
u
tatio
n
al
co
m
p
lex
ity
,
en
ab
lin
g
it to
b
e
u
s
ed
o
n
r
eso
u
r
ce
-
lim
ited
I
o
T
d
e
v
i
ce
s
.
Sh
ith
ar
th
et
a
l.
[
1
0
]
d
e
v
elo
p
ed
a
n
o
v
el
clu
s
ter
in
g
-
b
ased
class
if
icatio
n
m
eth
o
d
f
o
r
n
etwo
r
k
I
DS
u
s
in
g
NSL
-
KDD,
C
I
C
I
DS2
0
1
7
,
a
n
d
B
o
t
-
I
o
T
d
atasets
.
T
h
ey
co
m
b
in
ed
an
ticip
ated
d
is
tan
ce
-
b
ased
clu
s
ter
in
g
(
A
DC
)
with
d
en
s
ity
-
b
ased
s
p
atial
clu
s
ter
in
g
o
f
ap
p
licatio
n
s
with
n
o
is
e
(
DB
Scan
)
f
o
r
d
ata
g
r
o
u
p
i
n
g
,
o
p
tim
ized
p
ar
am
eter
s
u
s
in
g
p
e
r
p
etu
a
l
p
ig
eo
n
g
alv
a
n
ized
o
p
tim
iz
atio
n
(
PP
GO)
,
an
d
em
p
lo
y
ed
lik
elih
o
o
d
n
aïv
e
B
ay
es
(
L
NB
)
f
o
r
f
in
al
c
lass
if
icatio
n
.
T
h
eir
ADC
-
DB
Scan
-
L
NB
m
o
d
el
o
u
tp
er
f
o
r
m
ed
o
t
h
er
tec
h
n
iq
u
e
s
in
p
e
r
f
o
r
m
an
ce
ev
alu
atio
n
s
.
Ho
wev
er
,
e
n
s
em
b
le
lear
n
in
g
m
eth
o
d
s
,
w
h
ich
co
m
b
in
e
m
u
ltip
le
class
if
ier
s
to
lev
er
a
g
e
th
eir
co
llectiv
e
s
tr
en
g
th
s
,
h
av
e
s
h
o
wn
s
u
p
er
i
o
r
p
er
f
o
r
m
an
ce
in
h
an
d
lin
g
co
m
p
le
x
an
d
d
iv
e
r
s
e
d
ata
ty
p
es
an
d
attac
k
p
a
tter
n
s
o
f
ten
e
n
co
u
n
ter
ed
in
I
o
T
en
v
ir
o
n
m
en
ts
[
1
1
]
,
[
1
2
]
.
Ad
d
itio
n
ally
,
r
ec
en
t
s
tu
d
ies
o
n
en
s
em
b
le
lear
n
in
g
m
eth
o
d
s
f
o
r
I
DS
in
I
o
T
n
et
wo
r
k
s
h
av
e
f
o
cu
s
e
d
o
n
en
h
an
cin
g
an
o
m
aly
d
ete
ctio
n
ca
p
ab
ilit
ies.
B
y
u
tili
zin
g
en
s
em
b
le
tec
h
n
iq
u
es
lik
e
ex
tr
em
e
g
r
ad
ien
t
-
b
o
o
s
tin
g
(
XGBo
o
s
t)
,
L
i
g
h
tGB
M,
an
d
s
u
p
er
lear
n
er
,
th
ese
s
t
u
d
ies
aim
to
im
p
r
o
v
e
th
e
ac
c
u
r
ac
y
an
d
ef
f
icien
c
y
o
f
a
n
o
m
aly
d
etec
tio
n
[
1
2
]
–
[
1
4
]
.
So
n
i
et
a
l.
[
1
3
]
em
p
lo
y
ed
e
n
s
em
b
le
lear
n
in
g
tech
n
iq
u
es,
p
ar
ticu
lar
ly
XGBo
o
s
t
an
d
L
ig
h
tGB
M,
to
i
m
p
r
o
v
e
b
in
a
r
y
class
if
icatio
n
in
I
DS
f
o
r
I
o
T
n
etwo
r
k
s
.
T
h
es
e
m
eth
o
d
s
en
h
a
n
ce
an
o
m
aly
d
etec
tio
n
ac
c
u
r
ac
y
an
d
im
p
r
o
v
e
th
e
d
is
tr
ib
u
tio
n
o
f
d
etec
tio
n
ca
p
ab
ilit
ies
ac
r
o
s
s
I
o
T
d
e
v
ices.
B
aleg
a
et
a
l.
[
1
4
]
in
d
icate
d
th
at
XGBo
o
s
t
is
a
s
u
p
er
io
r
m
o
d
el
f
o
r
an
o
m
aly
d
etec
tio
n
i
n
I
o
T
n
etwo
r
k
s
.
I
t
o
u
tp
er
f
o
r
m
ed
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
SVM)
an
d
d
ee
p
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
s
(
D
C
NN)
,
ac
h
iev
in
g
u
p
to
9
9
.
9
8
%
ac
cu
r
ac
y
.
A
d
d
itio
n
ally
,
XGBo
o
s
t
d
em
o
n
s
tr
ated
s
ig
n
if
ican
tly
f
aster
tr
ain
in
g
t
im
es,
b
ein
g
7
1
7
.
7
5
tim
es q
u
ick
er
th
an
SVM.
D
e
s
p
i
t
e
t
h
e
s
e
c
r
u
c
i
a
l
ad
v
a
n
ce
s
,
e
x
i
s
t
i
n
g
a
p
p
r
o
a
c
h
e
s
s
t
i
l
l
f
a
c
e
n
o
t
a
b
le
c
h
a
l
le
n
g
e
s
a
n
d
e
n
c
o
u
n
te
r
i
m
p
o
r
t
a
n
t
l
im
i
t
a
t
io
n
s
,
in
c
l
u
d
in
g
h
i
g
h
c
o
m
p
u
t
a
t
i
o
n
a
l
r
eq
u
ir
e
m
e
n
t
s
,
d
i
f
f
i
c
u
l
ty
i
n
d
e
t
e
c
t
in
g
z
e
r
o
-
d
a
y
a
t
t
a
c
k
s
,
l
i
m
i
t
e
d
ad
a
p
ta
b
i
l
i
ty
a
n
d
s
c
al
a
b
i
l
i
t
y
to
e
v
o
l
v
i
n
g
t
h
r
e
a
t
s
,
a
n
d
co
m
p
l
e
x
i
ty
in
m
an
a
g
in
g
h
e
t
e
r
o
g
en
eo
u
s
I
o
T
d
a
t
a
.
T
h
e
s
e
o
b
s
t
a
c
le
s
i
m
p
e
d
e
th
e
d
ev
e
l
o
p
m
en
t
o
f
r
o
b
u
s
t
an
d
s
c
a
l
ab
l
e
s
ec
u
r
i
ty
s
o
lu
t
i
o
n
s
.
Ad
d
r
e
s
s
i
n
g
t
h
e
s
e
g
a
p
s
r
eq
u
ir
e
s
f
o
cu
s
e
d
i
m
p
r
o
v
e
m
e
n
t
s
in
t
w
o
k
ey
a
r
ea
s
:
i
)
r
e
f
in
i
n
g
d
e
t
e
c
t
io
n
p
e
r
f
o
r
m
an
c
e
t
o
e
f
f
e
c
t
iv
e
ly
h
a
n
d
l
e
d
i
v
e
r
s
e
I
o
T
s
i
t
u
a
t
io
n
s
a
n
d
em
e
r
g
in
g
t
h
r
e
a
t
s
,
in
c
lu
d
in
g
z
e
r
o
-
d
a
y
a
t
t
a
ck
s
,
a
n
d
ii
)
en
h
an
c
in
g
s
c
a
l
ab
i
l
i
t
y
a
n
d
r
e
s
o
u
r
c
e
e
f
f
i
c
i
en
c
y
to
m
ee
t
co
m
p
u
t
i
n
g
d
em
a
n
d
s
an
d
a
d
a
p
t
to
th
e
r
ap
i
d
e
x
p
an
s
i
o
n
o
f
th
e
I
o
T
l
a
n
d
s
c
ap
e
.
T
o
ad
d
r
ess
th
ese
is
s
u
es
an
d
o
v
er
co
m
e
th
e
s
ec
u
r
ity
lim
itatio
n
s
o
f
I
o
T
n
etwo
r
k
s
,
th
is
p
ap
e
r
p
r
o
p
o
s
es
a
n
o
v
el
b
u
tter
f
l
y
o
p
tim
izatio
n
-
b
ased
en
s
em
b
le
lear
n
in
g
(
B
OE
L
)
f
r
am
ewo
r
k
f
o
r
I
DS,
in
s
p
ir
ed
b
y
th
e
n
atu
r
al
f
o
r
ag
in
g
b
e
h
av
io
r
o
f
b
u
tter
f
lies
,
s
p
ec
if
ically
th
eir
u
s
e
o
f
s
en
s
o
r
y
m
o
d
alities
an
d
f
r
ag
r
a
n
ce
s
to
lo
ca
te
f
o
o
d
s
o
u
r
ce
s
an
d
c
o
m
m
u
n
icate
with
ea
ch
o
th
e
r
.
T
h
e
B
OE
L
ap
p
r
o
ac
h
lev
er
a
g
es
th
e
b
u
tter
f
ly
o
p
tim
izatio
n
alg
o
r
ith
m
(
B
OA)
to
d
y
n
am
ica
lly
o
p
tim
ize
t
h
e
s
elec
tio
n
a
n
d
weig
h
tin
g
o
f
b
ase
lear
n
er
s
in
th
e
e
n
s
em
b
le,
an
d
to
ac
h
iev
e
o
p
tim
al
ef
f
icie
n
c
y
in
id
e
n
tify
in
g
v
ar
io
u
s
attac
k
ty
p
es
b
y
f
in
d
in
g
th
e
m
o
s
t
ef
f
ec
tiv
e
weig
h
t
co
m
b
in
atio
n
s
f
o
r
en
s
em
b
le
m
o
d
els,
wh
ich
ar
e
f
o
u
r
s
o
p
h
is
ticated
g
r
ad
ien
t
-
b
o
o
s
tin
g
ML
t
ec
h
n
iq
u
es,
s
u
ch
as
XGBo
o
s
t,
ca
teg
o
r
ical
b
o
o
s
tin
g
(
C
atB
o
o
s
t)
,
L
ig
h
tGB
M,
an
d
g
r
ad
ie
n
t
-
b
o
o
s
tin
g
m
ac
h
i
n
es
(
GB
M)
f
o
r
ea
c
h
attac
k
ty
p
e
o
r
s
p
ec
if
ic
class
,
th
er
eb
y
e
n
h
an
ci
n
g
th
e
en
s
em
b
l
e’
s
ad
ap
tab
ilit
y
,
p
e
r
f
o
r
m
an
ce
,
an
d
co
m
p
u
tatio
n
al
ef
f
icien
cy
.
T
h
is
p
ap
er
p
r
im
ar
ily
d
eliv
er
s
th
e
f
o
llo
win
g
co
n
tr
ib
u
tio
n
s
:
i
)
a
n
o
v
el
en
s
em
b
le
lear
n
in
g
f
r
am
ewo
r
k
,
ter
m
e
d
B
OE
L
,
d
e
s
ig
n
ed
f
o
r
e
f
f
ec
tiv
e
in
tr
u
s
io
n
d
etec
tio
n
in
I
o
T
en
v
ir
o
n
m
en
ts
b
y
lev
er
a
g
in
g
th
e
B
OA
f
o
r
d
y
n
am
ic
weig
h
t
o
p
t
im
izatio
n
o
f
b
ase
lear
n
er
s
alo
n
g
s
id
e
g
r
ad
ien
t
-
b
o
o
s
tin
g
ML
ap
p
r
o
ac
h
es
;
ii
)
th
e
p
r
o
p
o
s
ed
B
OE
L
f
r
am
ewo
r
k
is
r
ig
o
r
o
u
s
ly
ev
alu
ated
u
s
in
g
two
wid
ely
r
ec
o
g
n
ized
p
u
b
lic
r
ea
l
-
wo
r
ld
I
o
T
s
ec
u
r
ity
d
atasets
,
B
o
t
-
I
o
T
an
d
C
I
C
I
DS2
0
1
7
;
a
n
d
iii
)
th
e
p
er
f
o
r
m
an
ce
o
f
B
OE
L
is
b
en
ch
m
ar
k
ed
ag
ain
s
t
s
tate
-
of
-
th
e
-
a
r
t I
DS te
ch
n
iq
u
es.
T
h
e
r
est
o
f
th
e
p
ap
e
r
is
s
tr
u
ctu
r
ed
as
f
o
llo
ws:
s
ec
tio
n
2
p
r
o
v
id
es
th
e
ch
o
s
en
m
eth
o
d
o
l
o
g
y
an
d
th
e
p
r
o
p
o
s
ed
B
OE
L
f
r
am
ewo
r
k
d
esig
n
in
d
etail.
Sectio
n
3
ad
d
r
ess
es
th
e
ex
p
er
im
en
tal
r
esu
lts
,
d
is
cu
s
s
io
n
,
an
aly
s
is
,
an
d
ev
alu
atio
n
m
eth
o
d
.
Fin
ally
,
s
ec
tio
n
4
o
u
tlin
es
an
d
co
n
clu
d
es th
e
s
tu
d
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
25
:
3
4
9
4
-
3
5
0
5
3496
2.
P
RO
P
O
SE
D
M
E
T
H
O
D
T
h
is
s
ec
tio
n
d
escr
ib
es
th
e
d
ev
elo
p
m
en
t
o
f
an
ad
v
a
n
ce
d
B
OE
L
I
DS
f
r
am
ewo
r
k
th
at
in
teg
r
ates
s
ev
er
al
ML
m
o
d
els
an
d
o
p
ti
m
izes
th
eir
co
m
b
in
atio
n
u
s
in
g
B
OA,
en
s
u
r
in
g
th
at
th
e
s
y
s
tem
is
b
o
th
ef
f
ec
tiv
e
an
d
ef
f
icien
t
in
id
en
tif
y
in
g
v
ar
io
u
s
cy
b
e
r
th
r
ea
ts
in
I
o
T
n
etwo
r
k
s
,
wh
ich
im
p
lies
a
s
y
s
t
em
atic
ap
p
r
o
ac
h
to
o
v
er
co
m
i
n
g
v
ar
io
u
s
cy
b
er
s
e
cu
r
ity
ch
allen
g
es
b
y
b
o
o
s
ti
n
g
d
etec
tio
n
ca
p
ab
ilit
ies
in
I
o
T
n
etwo
r
k
s
.
As
illu
s
tr
ated
in
Fig
u
r
e
1
,
th
e
s
y
s
tem
o
p
er
ates
in
th
r
ee
p
h
ases
:
d
ata
p
r
ep
r
o
ce
s
s
in
g
p
h
ase,
tr
ain
in
g
p
h
ase,
a
n
d
test
in
g
/p
r
ed
ictio
n
p
h
ase.
I
t
b
e
g
in
s
with
co
m
p
r
eh
en
s
iv
e
d
ata
co
llectio
n
f
r
o
m
estab
lis
h
ed
I
o
T
s
ec
u
r
ity
d
atasets
,
an
d
th
en
,
v
ia
a
s
er
ies
o
f
s
u
cc
ess
iv
e
s
tep
s
,
we
o
b
tain
th
e
f
in
al
p
r
ed
icted
class
to
p
er
f
o
r
m
attac
k
-
b
ased
d
etec
tio
n
.
Fig
u
r
e
1
.
T
h
e
f
r
a
m
ewo
r
k
o
f
th
e
p
r
o
p
o
s
ed
B
OE
L
-
b
ased
I
DS
2
.
1
.
Da
t
a
s
et
s
co
llect
io
n
T
h
e
C
I
C
I
DS2
0
1
7
d
ataset
[
1
5
]
,
d
e
v
elo
p
e
d
b
y
th
e
C
an
a
d
ian
I
n
s
titu
te
f
o
r
C
y
b
e
r
s
ec
u
r
ity
,
is
a
co
m
p
r
eh
e
n
s
iv
e
d
ataset
wid
ely
u
s
ed
f
o
r
ev
alu
atin
g
in
tr
u
s
io
n
d
etec
tio
n
s
y
s
tem
s
in
I
o
T
en
v
ir
o
n
m
en
ts
.
I
t
co
n
tain
s
ap
p
r
o
x
im
ately
2
.
8
m
illi
o
n
n
etwo
r
k
f
lo
ws
ca
p
t
u
r
ed
o
v
er
f
iv
e
d
ay
s
,
r
ep
r
esen
tin
g
b
o
th
b
en
ig
n
tr
af
f
ic
an
d
6
ty
p
es
o
f
m
o
d
e
r
n
attac
k
s
(
i.e
.
,
d
en
ial
o
f
s
er
v
ice
(
D
o
S),
b
r
u
te
f
o
r
ce
,
b
o
tn
et,
web
attac
k
s
,
in
f
iltra
tio
n
attac
k
s
,
an
d
s
n
if
f
in
g
)
r
elev
a
n
t
to
I
o
T
n
etwo
r
k
s
.
I
t
m
ee
ts
1
1
cr
u
cial
b
en
ch
m
ar
k
s
f
o
r
I
DS
d
ataset
s
an
d
en
co
m
p
ass
es
o
v
er
2
2
5
,
7
4
5
p
a
ck
ag
es,
in
w
h
ich
ea
ch
f
lo
w
is
d
escr
ib
ed
b
y
m
o
r
e
th
a
n
8
0
f
ea
tu
r
es.
T
h
e
d
ataset’
s
p
o
p
u
lar
ity
s
tem
s
f
r
o
m
its
r
e
alis
tic
n
atu
r
e,
p
r
o
p
o
r
tio
n
ate
r
ep
r
esen
tatio
n
o
f
n
o
r
m
al
an
d
attac
k
tr
af
f
ic,
an
d
in
clu
s
io
n
o
f
d
iv
e
r
s
e
attac
k
s
ce
n
ar
io
s
.
T
h
ese
ch
ar
ac
ter
is
tics
m
ak
e
C
I
C
I
DS2
0
1
7
h
ig
h
ly
s
u
i
tab
le
f
o
r
I
o
T
-
b
ased
r
esear
ch
an
d
a
n
id
ea
l c
h
o
ice
f
o
r
d
ev
el
o
p
in
g
an
d
e
v
alu
atin
g
ML
-
b
ased
I
DS f
o
r
I
o
T
e
n
v
ir
o
n
m
en
ts
.
T
h
e
B
o
t
-
I
o
T
d
ataset,
cr
ea
ted
at
UNSW
C
an
b
er
r
a’
s
C
y
b
er
R
an
g
e
L
ab
,
is
a
s
ig
n
if
ican
t
r
eso
u
r
ce
s
p
ec
if
ically
d
esig
n
ed
to
a
d
d
r
e
s
s
th
e
ch
allen
g
es
o
f
in
tr
u
s
io
n
d
etec
tio
n
in
I
o
T
en
v
ir
o
n
m
en
ts
[
1
6
]
.
T
h
e
d
ataset
is
av
ailab
le
in
two
f
o
r
m
ats:
a
co
m
p
r
eh
en
s
iv
e
v
e
r
s
io
n
with
o
v
e
r
7
2
m
illi
o
n
r
ec
o
r
d
s
an
d
a
c
o
n
d
en
s
ed
5
%
s
am
p
le
co
n
tain
in
g
ap
p
r
o
x
im
ately
3
m
illi
o
n
en
tr
ies
o
f
b
o
th
n
o
r
m
al
a
n
d
m
alicio
u
s
tr
af
f
ic
ca
p
tu
r
e
d
f
r
o
m
a
r
ea
lis
tic
I
o
T
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
B
u
tter
fly
o
p
timiz
a
tio
n
-
b
a
s
ed
e
n
s
emb
le
lea
r
n
in
g
s
tr
a
teg
y
fo
r
a
d
va
n
ce
d
…
(
Mo
u
a
d
C
h
o
u
kh
a
ir
i
3497
n
etwo
r
k
s
etu
p
.
T
h
e
d
ataset
in
c
lu
d
es v
ar
io
u
s
I
o
T
-
s
p
ec
if
ic
atta
ck
s
s
u
ch
as d
is
tr
ib
u
ted
d
en
ial
o
f
s
er
v
ice
(
DDo
S),
OS
an
d
s
er
v
ice
s
ca
n
,
Do
S,
k
ey
lo
g
g
in
g
,
an
d
d
ata
ex
f
iltra
tio
n
,
m
ak
in
g
it
h
i
g
h
ly
r
elev
a
n
t
f
o
r
I
o
T
s
ec
u
r
ity
r
esear
ch
.
T
h
e
B
o
t
-
I
o
T
d
ata
s
et’
s
p
o
p
u
lar
ity
in
I
o
T
s
ec
u
r
ity
ex
p
er
im
e
n
ts
lies
in
its
co
m
p
r
eh
en
s
iv
e
r
ep
r
esen
tatio
n
o
f
I
o
T
-
s
p
ec
if
ic
th
r
ea
ts
,
h
i
g
h
-
q
u
ality
g
r
o
u
n
d
tr
u
th
lab
els,
a
n
d
th
e
i
n
clu
s
io
n
o
f
b
o
th
I
o
T
an
d
in
d
u
s
tr
ial
in
ter
n
et
o
f
th
in
g
s
(
I
I
o
T
)
tr
af
f
ic.
I
t
ad
d
r
ess
es
th
e
lim
itatio
n
s
o
f
p
r
e
v
io
u
s
d
atasets
b
y
in
co
r
p
o
r
atin
g
a
d
iv
er
s
e
r
an
g
e
o
f
I
o
T
p
r
o
to
c
o
ls
an
d
d
e
v
ices,
m
ak
in
g
it m
o
r
e
r
ep
r
esen
tativ
e
o
f
r
ea
l
-
wo
r
ld
I
o
T
ec
o
s
y
s
tem
s
.
2
.
2
.
Da
t
a
prepa
ra
t
i
o
n (
prepro
ce
s
s
ing
)
T
h
e
d
ata
p
r
ep
r
o
ce
s
s
in
g
s
tag
e
is
d
ed
icate
d
to
p
r
e
p
ar
in
g
tr
af
f
ic
d
ata
f
r
o
m
t
h
e
C
I
C
I
DS2
0
1
7
an
d
B
o
t
-
I
o
T
d
atasets
.
I
t
tr
a
n
s
f
o
r
m
s
r
a
w
d
ata
in
to
a
s
tr
u
ctu
r
e
d
tab
le
o
f
p
r
ed
ef
in
e
d
f
ea
tu
r
es
o
p
tim
iz
ed
f
o
r
in
p
u
t
in
to
th
e
ML
m
o
d
els
i
n
th
e
tr
ain
in
g
p
r
o
ce
s
s
s
tag
e.
T
h
is
co
m
p
o
n
en
t
o
p
er
ates
th
r
o
u
g
h
a
s
er
ies
o
f
co
n
s
ec
u
tiv
e
s
tep
s
,
ea
c
h
d
esig
n
ed
to
r
ef
in
e
a
n
d
f
o
r
m
at
th
e
d
ata
f
o
r
o
p
tim
al
an
al
y
s
is
an
d
m
o
d
el
p
er
f
o
r
m
an
ce
,
wh
ich
in
clu
d
e
th
e
f
o
llo
win
g
p
r
o
ce
s
s
es:
2
.
2
.
1
.
Da
t
a
clea
nin
g
I
n
o
u
r
s
tu
d
y
,
we
im
p
lem
en
ted
a
co
m
p
r
eh
en
s
iv
e
d
ata
-
clea
n
i
n
g
p
r
o
ce
d
u
r
e
to
elim
in
ate
in
co
n
s
is
ten
cie
s
an
d
in
ac
c
u
r
ac
ies.
T
h
is
p
r
o
ce
s
s
in
v
o
lv
ed
a
th
o
r
o
u
g
h
ex
a
m
in
atio
n
o
f
th
e
d
ata
to
g
ain
d
ee
p
er
in
s
ig
h
ts
an
d
r
ec
tify
an
y
m
is
in
ter
p
r
etatio
n
s
.
Fo
r
b
o
th
C
I
C
I
DS2
0
1
7
an
d
B
o
t
-
I
o
T
d
atasets
,
m
is
s
in
g
v
alu
e
s
wer
e
h
a
n
d
led
b
y
im
p
u
tin
g
with
m
ed
ian
v
alu
es.
T
h
e
m
e
d
ian
im
p
u
tatio
n
was
s
elec
ted
f
o
r
its
r
esil
ien
ce
ag
a
in
s
t
o
u
tlier
ef
f
ec
ts
co
m
p
ar
ed
to
m
ea
n
-
b
ased
m
e
th
o
d
s
.
Ad
d
itio
n
ally
,
d
u
p
licat
e
an
d
c
o
r
r
u
p
ted
r
ec
o
r
d
s
,
wh
i
ch
co
u
l
d
s
k
ew
th
e
an
aly
s
is
,
wer
e
id
en
tifie
d
b
y
f
ilter
in
g
tech
n
iq
u
es
f
o
r
m
is
s
in
g
v
alu
e
co
lu
m
n
s
an
d
d
u
p
lica
te
r
o
ws
an
d
wer
e
r
em
o
v
ed
to
en
s
u
r
e
th
e
u
n
iq
u
e
n
ess
o
f
ea
ch
d
ata
p
o
in
t.
Fin
ally
,
we
r
ec
tifie
d
attr
ib
u
te
lab
els,
a
co
m
m
o
n
is
s
u
e
in
co
m
m
a
-
s
ep
ar
ated
v
alu
es (
C
SV)
f
o
r
m
atted
d
ata.
T
o
p
r
eser
v
e
d
ata
in
te
g
r
ity
,
we
estab
lis
h
ed
a
s
tan
d
ar
d
ized
d
ata
r
ep
r
esen
tatio
n
,
e
n
s
u
r
in
g
ea
c
h
attr
ib
u
te
m
ai
n
tain
ed
a
s
in
g
u
lar
,
u
n
am
b
ig
u
o
u
s
v
alu
e
p
er
en
t
r
y
.
T
h
is
co
m
p
r
eh
e
n
s
iv
e
r
ef
in
em
e
n
t p
r
o
ce
s
s
was in
s
tr
u
m
en
tal
in
cr
ea
tin
g
a
m
o
r
e
ac
cu
r
ate
a
n
d
r
eliab
l
e
d
ataset,
ess
en
tial
f
o
r
d
e
v
elo
p
in
g
a
r
o
b
u
s
t I
DS.
2
.
2
.
2
.
Da
t
a
no
rm
a
liza
t
io
n
Data
n
o
r
m
aliza
tio
n
is
ess
en
tial
to
en
s
u
r
e
th
at
th
e
f
ea
tu
r
es
in
th
e
d
ataset
ar
e
o
n
a
s
im
ilar
s
ca
le,
wh
ich
h
elp
s
im
p
r
o
v
e
t
h
e
p
e
r
f
o
r
m
an
c
e
o
f
m
an
y
m
ac
h
in
e
lear
n
in
g
a
lg
o
r
ith
m
s
.
W
e
em
p
lo
y
ed
z
-
s
c
o
r
e
n
o
r
m
aliza
tio
n
,
also
k
n
o
wn
as
s
tan
d
ar
d
izatio
n
,
wh
ich
tr
a
n
s
f
o
r
m
s
th
e
d
ata
to
h
av
e
a
m
ea
n
o
f
0
a
n
d
a
s
tan
d
ar
d
d
e
v
iatio
n
o
f
1
,
to
en
s
u
r
e
all
f
ea
tu
r
es
ar
e
o
n
a
s
im
ilar
s
ca
le.
T
h
is
tech
n
iq
u
e
is
p
ar
ticu
lar
ly
ef
f
ec
tiv
e
in
h
an
d
lin
g
d
atasets
with
attr
ib
u
tes
th
at
h
av
e
v
astl
y
d
i
f
f
er
en
t
r
an
g
es,
e
n
s
u
r
in
g
th
at
ea
ch
f
ea
tu
r
e
co
n
tr
ib
u
tes
eq
u
a
lly
to
th
e
m
o
d
el’
s
lear
n
in
g
p
r
o
ce
s
s
.
Fo
r
i
n
s
tan
ce
,
in
th
e
co
n
tex
t
o
f
r
ain
f
all
d
ata
class
if
icatio
n
,
z
-
s
co
r
e
n
o
r
m
aliza
tio
n
h
as
b
ee
n
s
h
o
wn
to
im
p
r
o
v
e
ac
c
u
r
ac
y
,
s
en
s
itiv
ity
,
an
d
s
p
ec
if
icity
r
at
es,
o
u
tp
er
f
o
r
m
in
g
u
n
n
o
r
m
aliz
ed
d
ata
an
d
o
th
e
r
n
o
r
m
aliza
tio
n
tec
h
n
iq
u
es lik
e
m
in
-
m
ax
n
o
r
m
aliza
tio
n
[
1
7
]
.
T
h
e
f
o
r
m
u
la
f
o
r
z
-
s
co
r
e
n
o
r
m
aliza
tio
n
is
:
=
(
−
)
(
1
)
wh
er
e
X
r
ep
r
esen
ts
t
h
e
o
r
ig
in
al
d
ata
p
o
in
t,
µ
d
en
o
tes
th
e
m
ea
n
o
f
t
h
e
f
ea
tu
r
e,
an
d
σ
s
ig
n
if
ies
th
e
s
tan
d
ar
d
d
ev
iatio
n
o
f
th
e
f
ea
t
u
r
e.
2
.
2
.
3
.
Ca
t
eg
o
rica
l e
nco
din
g
Fo
r
ca
teg
o
r
ical
f
ea
tu
r
es,
we
em
p
lo
y
ed
lab
el
e
n
co
d
in
g
,
w
h
ich
is
a
s
tr
aig
h
tf
o
r
war
d
tec
h
n
iq
u
e
f
o
r
co
n
v
er
tin
g
ca
teg
o
r
ical
v
ar
iab
les
in
to
n
u
m
er
ical
v
alu
es
b
y
ass
ig
n
in
g
a
u
n
i
q
u
e
in
te
g
er
t
o
ea
ch
ca
teg
o
r
y
t
o
tr
an
s
f
o
r
m
ca
te
g
o
r
ical
f
ea
tu
r
es p
r
esen
t
in
o
u
r
d
atasets
in
to
n
u
m
er
ical
r
ep
r
esen
tatio
n
s
.
T
h
is
m
eth
o
d
was
c
h
o
s
en
d
u
e
to
its
s
im
p
licity
an
d
ef
f
ec
tiv
en
ess
,
p
ar
ticu
lar
ly
f
o
r
o
r
d
in
al
f
ea
tu
r
es
wh
er
e
th
e
o
r
d
er
o
f
ca
teg
o
r
ies
is
m
ea
n
in
g
f
u
l.
W
e
en
s
u
r
ed
th
at
t
h
e
en
co
d
in
g
m
et
h
o
d
alig
n
ed
w
ell
with
th
e
m
o
d
el
r
e
q
u
ir
em
e
n
ts
.
2
.
2
.
4
.
Da
t
a
ba
la
ncing
T
o
ad
d
r
ess
th
e
is
s
u
e
o
f
cla
s
s
im
b
alan
ce
in
I
o
T
d
ataset
s
,
we
ap
p
lied
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
(
SM
OT
E
)
[
1
8
]
.
T
h
is
tech
n
iq
u
e
g
e
n
er
ates
s
y
n
th
etic
s
am
p
les
f
o
r
th
e
m
in
o
r
ity
class
es
b
y
in
te
r
p
o
latin
g
b
etwe
en
ex
is
tin
g
m
in
o
r
ity
s
am
p
les,
th
er
e
b
y
b
alan
cin
g
th
e
d
ataset.
B
y
ap
p
ly
in
g
SMOT
E
,
we
b
alan
ce
d
th
e
d
atasets
m
o
r
e
ef
f
ec
tiv
ely
,
en
s
u
r
in
g
th
at
th
e
I
D
S
m
o
d
el
co
u
ld
lear
n
e
q
u
ally
f
r
o
m
all
class
es
an
d
th
u
s
p
er
f
o
r
m
b
etter
i
n
d
etec
tin
g
r
ar
e
attac
k
t
y
p
es.
2
.
2
.
5
.
Da
t
a
pa
rt
it
io
nin
g
T
h
e
d
ata
p
a
r
titi
o
n
in
g
p
h
ase
p
lay
s
a
c
r
itical
r
o
le
in
ML
wo
r
k
f
lo
w.
E
ac
h
p
r
e
-
p
r
o
ce
s
s
ed
d
ata
was
r
an
d
o
m
l
y
p
a
r
titi
o
n
ed
in
t
o
tr
ai
n
in
g
a
n
d
test
in
g
s
ets
u
s
in
g
a
n
8
0
/2
0
s
p
lit,
wh
ich
is
a
co
m
m
o
n
p
r
ac
tice
in
ML
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
25
:
3
4
9
4
-
3
5
0
5
3498
f
o
r
m
o
d
el
ev
alu
atio
n
.
I
n
ad
d
i
tio
n
,
a
5
-
f
o
l
d
cr
o
s
s
-
v
alid
atio
n
is
in
teg
r
ated
to
ev
alu
ate
th
e
p
er
f
o
r
m
a
n
ce
an
d
g
en
er
aliza
b
ilit
y
o
f
a
m
o
d
el
b
y
p
ar
titi
o
n
in
g
th
e
d
ataset
in
to
f
i
v
e
s
u
b
s
ets,
o
r
f
o
ld
s
.
E
ac
h
f
o
l
d
is
u
s
ed
as a
te
s
t se
t
o
n
ce
,
wh
ile
th
e
r
em
ain
in
g
f
o
u
r
f
o
ld
s
ar
e
u
s
ed
f
o
r
tr
ain
in
g
,
en
s
u
r
in
g
th
at
ev
er
y
d
ata
p
o
i
n
t
is
u
s
ed
f
o
r
b
o
th
tr
ain
in
g
an
d
v
alid
atio
n
.
T
h
is
m
eth
o
d
h
el
p
s
m
itig
ate
is
s
u
es
r
elate
d
to
o
v
e
r
f
itti
n
g
an
d
p
r
o
v
id
es
a
m
o
r
e
r
eliab
le
esti
m
ate
o
f
m
o
d
el
p
e
r
f
o
r
m
an
c
e
th
an
a
s
in
g
le
tr
ain
-
test
s
p
lit.
2
.
3
.
B
a
s
e
ma
chine le
a
rning
m
o
dels
a
nd
t
ra
ini
ng
pro
ce
s
s
Dec
is
io
n
tr
ee
s
(
DT
)
a
r
e
a
s
u
p
er
v
is
ed
ML
tech
n
iq
u
e
co
m
m
o
n
ly
a
p
p
lied
to
b
o
th
class
if
icatio
n
an
d
r
eg
r
ess
io
n
p
r
o
b
lem
s
,
r
en
o
wn
e
d
f
o
r
th
eir
s
im
p
licity
an
d
in
te
r
p
r
etab
ilit
y
.
DT
s
f
u
n
ctio
n
b
y
iter
ativ
ely
s
p
litt
in
g
th
e
d
ataset
in
to
s
m
aller
s
u
b
g
r
o
u
p
s
b
ased
o
n
th
e
v
alu
es
o
f
th
e
in
p
u
t
f
ea
tu
r
es.
T
h
is
r
ec
u
r
s
iv
e
p
ar
titi
o
n
in
g
co
n
s
tr
u
cts
a
h
ier
a
r
ch
ical,
tr
e
e
-
lik
e
m
o
d
el
o
f
d
ec
is
io
n
r
u
les.
E
ac
h
in
ter
n
al
n
o
d
e
in
th
e
tr
ee
s
tr
u
ctu
r
e
r
ep
r
esen
ts
a
”test”
co
n
d
u
cted
o
n
an
attr
ib
u
te,
th
e
b
r
a
n
ch
es
em
a
n
atin
g
f
r
o
m
th
e
n
o
d
e
c
o
r
r
esp
o
n
d
to
t
h
e
o
u
tco
m
es
o
f
th
ese
test
s
,
an
d
th
e
leaf
n
o
d
es d
en
o
t
e
th
e
f
in
al
class
lab
els
o
r
a
co
n
tin
u
o
u
s
v
alu
e
(
in
th
e
ca
s
e
o
f
r
eg
r
ess
io
n
)
ass
ig
n
ed
to
th
e
d
ata
in
s
tan
ce
s
[
1
9
]
.
DT
s
ar
e
f
u
n
d
am
en
tal
co
m
p
o
n
e
n
t
s
o
f
g
r
ad
ien
t
-
b
o
o
s
tin
g
d
ec
is
io
n
tr
ee
s
(
GB
DT
)
,
a
p
o
wer
f
u
l
en
s
em
b
le
lear
n
in
g
t
ec
h
n
iq
u
e
th
at
co
m
b
in
es
th
e
p
r
ed
ictio
n
s
o
f
m
u
ltip
le
DT
s
to
im
p
r
o
v
e
ac
c
u
r
ac
y
an
d
r
o
b
u
s
tn
ess
.
GB
DT
co
n
s
tr
u
cts
a
m
o
d
el
in
a
s
tag
e
-
wis
e
f
ash
io
n
b
y
s
eq
u
en
tially
a
d
d
in
g
DT
s
,
wh
e
r
e
ea
ch
n
ew
tr
ee
c
o
r
r
ec
ts
th
e
er
r
o
r
s
m
ad
e
b
y
th
e
p
r
e
v
io
u
s
o
n
es.
T
h
is
is
ac
h
iev
ed
b
y
f
itti
n
g
th
e
n
ew
tr
ee
to
th
e
r
esid
u
al
er
r
o
r
s
o
f
th
e
co
m
b
in
ed
en
s
e
m
b
le
o
f
tr
ee
s
,
e
f
f
ec
tiv
ely
u
s
in
g
g
r
ad
ien
t
d
escen
t
t
o
m
in
im
ize
th
e
lo
s
s
f
u
n
ctio
n
[
2
0
]
.
GB
DT
s
h
av
e
ev
o
lv
e
d
s
i
g
n
if
ican
tly
,
r
esu
ltin
g
in
s
ev
er
al
s
tate
-
of
-
th
e
-
ar
t
al
g
o
r
ith
m
s
t
h
at
ar
e
wid
ely
u
s
ed
in
ML
.
T
h
e
p
r
im
ar
y
ty
p
es
o
f
g
r
ad
ien
t
-
b
o
o
s
tin
g
alg
o
r
ith
m
s
i
n
clu
d
e
th
e
o
r
ig
in
al
GB
M,
XGBo
o
s
t,
L
ig
h
tGB
M,
an
d
C
atB
o
o
s
t.
GB
M
is
th
e
f
o
u
n
d
atio
n
al
alg
o
r
ith
m
th
at
in
tr
o
d
u
ce
d
th
e
c
o
n
c
ep
t
o
f
b
o
o
s
tin
g
wea
k
lear
n
e
r
s
to
f
o
r
m
a
s
tr
o
n
g
p
r
e
d
ictiv
e
m
o
d
el
[
2
1
]
.
I
t
im
p
r
o
v
es
ac
cu
r
ac
y
b
y
s
eq
u
en
tially
co
m
b
in
in
g
DT
s
,
wh
er
e
ea
ch
s
u
b
s
eq
u
e
n
t
tr
ee
f
o
cu
s
es
o
n
co
r
r
ec
tin
g
th
e
er
r
o
r
s
o
f
th
e
p
r
ev
i
o
u
s
o
n
e.
GB
M’
s
u
tili
ty
ex
ten
d
s
to
n
atu
r
al
d
is
aster
p
r
ed
ictio
n
,
s
u
ch
as lan
d
s
lid
e
d
etec
tio
n
,
wh
er
e
it d
em
o
n
s
tr
ates h
ig
h
p
r
ed
ictiv
e
p
r
ec
is
io
n
an
d
ef
f
icien
c
y
in
p
r
o
ce
s
s
in
g
lar
g
e
d
atasets
,
o
u
tp
er
f
o
r
m
in
g
o
th
er
m
o
d
els lik
e
L
ig
h
tGB
M
in
ce
r
t
ain
s
ce
n
ar
io
s
[
2
2
]
.
XGBo
o
s
t
i
s
a
s
ca
lab
le
an
d
ef
f
icien
t
GB
DT
th
at
h
as
b
ec
o
m
e
a
p
o
p
u
lar
ch
o
ice
f
o
r
its
r
eli
ab
ilit
y
an
d
p
er
f
o
r
m
an
ce
in
ML
co
m
p
eti
tio
n
s
[
2
3
]
.
I
t
o
p
tim
izes
b
o
th
th
e
lo
s
s
f
u
n
ctio
n
an
d
a
r
eg
u
lar
izatio
n
ter
m
,
in
co
r
p
o
r
atin
g
L
1
a
n
d
L
2
r
e
g
u
l
ar
izatio
n
to
p
r
ev
e
n
t
o
v
er
f
itti
n
g
.
XGBo
o
s
t
f
ea
tu
r
es
b
u
ilt
-
in
h
an
d
lin
g
o
f
m
is
s
in
g
v
alu
es,
tr
ee
p
r
u
n
in
g
b
ased
o
n
’
ma
x_
d
ep
th
’
an
d
lo
s
s
r
ed
u
ctio
n
,
f
ea
tu
r
e
im
p
o
r
tan
ce
s
co
r
i
n
g
,
an
d
s
u
p
p
o
r
t
f
o
r
p
ar
allel
an
d
d
is
tr
ib
u
ted
co
m
p
u
tin
g
.
I
ts
co
m
p
u
tatio
n
al
co
m
p
lex
ity
is
lo
w
an
d
it
is
O(
d
||
x|
|Tlo
g
(
n
)
)
,
wh
e
r
e
d
is
th
e
m
ax
im
u
m
tr
ee
h
eig
h
t,
x
i
s
th
e
n
u
m
b
er
o
f
n
o
n
-
ze
r
o
s
am
p
les,
T
is
th
e
n
u
m
b
er
o
f
tr
e
es,
an
d
n
is
th
e
d
ata
len
g
th
.
XGBo
o
s
t
ca
n
b
e
ex
ec
u
ted
s
er
ially
o
n
a
s
in
g
le
th
r
ea
d
,
in
p
ar
allel
u
s
in
g
m
u
lti
-
th
r
ea
d
in
g
o
n
a
s
in
g
le
m
ac
h
in
e
o
r
d
is
tr
ib
u
ted
ac
r
o
s
s
m
u
ltip
le
m
ac
h
in
es
u
s
in
g
f
r
a
m
ewo
r
k
s
lik
e
Sp
ar
k
.
T
h
ese
c
h
ar
ac
ter
is
tics
,
alo
n
g
with
its
s
p
ee
d
an
d
p
er
f
o
r
m
a
n
c
e,
m
ak
e
XGBo
o
s
t p
ar
ticu
lar
ly
ef
f
ec
tiv
e
f
o
r
s
tr
u
ctu
r
ed
d
ata
p
r
o
b
lem
s
.
L
ig
h
tGB
M
is
an
ef
f
icien
t
a
n
d
f
ast
g
r
ad
ien
t
-
b
o
o
s
tin
g
f
r
a
m
ewo
r
k
th
at
u
s
es
tr
ee
-
b
ased
lear
n
in
g
alg
o
r
ith
m
s
[
2
4
]
.
I
t
em
p
l
o
y
s
a
n
o
v
el
tech
n
iq
u
e
ca
lled
g
r
ad
i
en
t
-
b
ased
o
n
e
-
s
id
e
s
am
p
lin
g
(
GOSS)
to
f
ilter
o
u
t
d
ata
in
s
tan
ce
s
with
s
m
all
g
r
a
d
ien
ts
an
d
e
x
clu
s
iv
e
f
ea
tu
r
e
b
u
n
d
lin
g
(
E
FB
)
to
r
e
d
u
ce
t
h
e
n
u
m
b
er
o
f
f
ea
tu
r
es.
T
h
ese
s
tr
ateg
ies
allo
w
L
ig
h
tG
B
M
to
ac
h
iev
e
f
aster
tr
ain
in
g
s
p
ee
d
an
d
h
ig
h
er
e
f
f
icien
cy
w
ith
lo
wer
m
e
m
o
r
y
u
s
ag
e.
T
h
e
alg
o
r
ith
m
g
r
o
ws
tr
ee
s
leaf
-
wis
e
(
b
est
-
f
ir
s
t)
r
ath
er
th
a
n
lev
el
-
wis
e,
w
h
ich
ca
n
lead
to
b
etter
ac
cu
r
ac
y
.
L
i
g
h
tGB
M
s
u
p
p
o
r
t
s
p
ar
allel
an
d
g
r
a
p
h
ics
p
r
o
ce
s
s
in
g
u
n
it
(
GPU)
lear
n
in
g
an
d
h
an
d
les
lar
g
e
-
s
ca
le
d
ata
ef
f
ec
tiv
el
y
.
B
y
im
p
lem
e
n
tin
g
GOSS,
th
e
alg
o
r
ith
m
r
e
d
u
ce
s
th
e
ef
f
ec
tiv
e
s
am
p
le
s
iz
e
to
N
r
,
wh
ile
E
FB
co
n
d
en
s
es
th
e
f
ea
tu
r
e
s
p
ac
e
to
F
b
.
C
o
n
s
eq
u
en
tly
,
L
ig
h
tGB
M
ac
h
iev
es
a
s
tr
ea
m
lin
e
d
s
p
ac
e
an
d
tim
e
co
m
p
lex
ity
o
f
O
(
N
r
∗
F
b
)
,
r
ep
r
esen
tin
g
a
s
u
b
s
tan
tial
im
p
r
o
v
e
m
en
t
o
v
er
tr
a
d
itio
n
al
a
p
p
r
o
ac
h
es
an
d
m
ain
tain
in
g
ess
en
tial in
f
o
r
m
atio
n
.
I
t is p
ar
t
icu
lar
ly
well
-
s
u
ited
f
o
r
la
r
g
e
d
atasets
an
d
h
ig
h
-
d
im
e
n
s
io
n
al
f
ea
tu
r
e
s
p
ac
es.
C
atB
o
o
s
t
is
a
p
o
wer
f
u
l
en
s
em
b
le
m
o
d
el
d
esig
n
ed
f
o
r
g
r
a
d
ien
t
b
o
o
s
tin
g
o
n
DT
s
[
2
5
]
.
I
t
is
b
u
ilt
t
o
h
an
d
le
m
an
y
ca
teg
o
r
ical
v
a
r
i
ab
les,
s
u
ch
as
ca
teg
o
r
ical,
te
x
tu
al,
a
n
d
n
u
m
e
r
ical
f
ea
tu
r
es,
ef
f
icien
tly
with
o
u
t
ex
ten
s
iv
e
p
r
e
p
r
o
ce
s
s
in
g
with
th
e
h
el
p
o
f
a
n
ativ
e
f
ea
tu
r
e
s
u
p
p
o
r
t
tech
n
i
q
u
e.
C
atB
o
o
s
t
em
p
lo
y
s
a
n
o
v
el
tech
n
iq
u
e
ca
lled
o
r
d
e
r
ed
b
o
o
s
tin
g
,
wh
ich
av
o
id
s
o
v
er
f
itti
n
g
an
d
r
ed
u
ce
s
p
r
e
d
ictio
n
s
h
if
ts
b
y
u
s
in
g
a
p
er
m
u
tatio
n
-
d
r
iv
e
n
alter
n
ativ
e
to
t
h
e
class
ic
g
r
a
d
ien
t
b
o
o
s
tin
g
s
ch
em
e.
I
t
also
u
s
es
a
s
y
m
m
etr
ic
tr
ee
s
tr
u
ctu
r
e
an
d
im
p
lem
en
ts
th
e
o
b
liv
io
u
s
DT
alg
o
r
ith
m
,
w
h
ich
ca
n
lead
to
f
aster
in
f
er
e
n
ce
ti
m
es
an
d
m
in
im
ize
o
v
er
f
itti
n
g
.
C
atB
o
o
s
t’
s
wea
k
co
m
p
u
tatio
n
al
c
o
m
p
lex
ity
is
g
e
n
er
ally
O
(
PN
DT
)
; in
th
is
co
n
te
x
t,
P
r
ep
r
esen
ts
th
e
n
u
m
b
er
o
f
s
u
b
s
et
p
e
r
m
u
tatio
n
s
,
wh
ile
N
DT
d
e
n
o
tes
th
e
t
o
tal
co
u
n
t
o
f
DT
m
o
d
els
em
p
l
o
y
ed
in
t
h
e
en
s
em
b
le
.
T
h
e
alg
o
r
ith
m
s
u
p
p
o
r
ts
GPU
ac
ce
ler
atio
n
an
d
ca
n
b
e
ex
ec
u
te
d
in
p
ar
allel
o
r
d
is
tr
ib
u
ted
m
o
d
es.
T
h
e
tr
ain
in
g
p
r
o
ce
s
s
f
o
r
b
as
e
lear
n
er
s
in
th
e
B
OE
L
f
r
a
m
ewo
r
k
i
n
v
o
lv
es
i
n
d
ep
e
n
d
en
tly
tr
ain
in
g
m
u
ltip
le
ML
m
o
d
els
o
n
a
tr
ain
in
g
d
ataset.
T
h
is
ap
p
r
o
ac
h
ai
m
s
to
ca
p
tu
r
e
d
iv
er
s
e
p
atter
n
s
an
d
in
s
ig
h
ts
f
r
o
m
th
e
d
ata,
lev
er
ag
in
g
th
e
u
n
iq
u
e
s
tr
en
g
th
s
o
f
v
ar
io
u
s
ML
alg
o
r
ith
m
s
.
B
y
tr
ain
in
g
m
o
d
els
s
u
ch
as
G
B
M,
C
atB
o
o
s
t,
L
ig
h
tG
B
M,
an
d
X
GB
o
o
s
t,
th
e
f
r
am
ewo
r
k
ca
n
lev
er
ag
e
its
ca
p
ab
ilit
ies in
h
an
d
lin
g
d
if
f
er
e
n
t a
s
p
ec
ts
o
f
th
e
d
ata,
in
cl
u
d
in
g
h
ig
h
-
d
i
m
en
s
io
n
al
f
ea
tu
r
es,
n
o
n
-
lin
ea
r
r
elatio
n
s
h
ip
s
,
an
d
im
b
ala
n
ce
d
class
es.
E
ac
h
b
ase
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
B
u
tter
fly
o
p
timiz
a
tio
n
-
b
a
s
ed
e
n
s
emb
le
lea
r
n
in
g
s
tr
a
teg
y
fo
r
a
d
va
n
ce
d
…
(
Mo
u
a
d
C
h
o
u
kh
a
ir
i
3499
lear
n
er
is
tr
ain
ed
to
p
r
ed
ict
th
e
tar
g
et
v
ar
ia
b
le
b
y
m
in
im
izin
g
its
lo
s
s
f
u
n
ctio
n
.
Su
b
s
eq
u
en
tly
,
th
e
p
r
ed
ictio
n
s
o
f
th
ese
m
o
d
els ar
e
co
m
b
i
n
ed
u
s
in
g
o
p
tim
ized
weig
h
ts
to
f
o
r
m
a
r
o
b
u
s
t e
n
s
em
b
le
m
o
d
el.
2
.
4
.
T
esting
/predict
io
n pro
ce
s
s
T
h
e
test
in
g
/p
r
ed
ictio
n
p
h
ase
o
f
th
e
B
OE
L
f
r
am
ewo
r
k
in
v
o
lv
es
lev
er
ag
in
g
th
e
o
p
tim
ized
weig
h
t
co
ef
f
icien
ts
d
er
iv
ed
f
r
o
m
B
O
A
to
ef
f
ec
tiv
ely
co
m
b
in
e
th
e
p
r
ed
ictio
n
s
o
f
th
e
in
d
iv
id
u
al
b
ase
lear
n
er
s
.
T
h
e
g
o
al
is
to
m
ax
i
m
ize
th
e
o
v
er
all
p
r
ed
ictiv
e
p
e
r
f
o
r
m
an
ce
o
f
th
e
en
s
em
b
le
b
y
h
ar
n
ess
in
g
th
e
d
is
tin
ctiv
e
s
tr
en
g
th
s
o
f
th
e
co
n
s
titu
en
t
m
o
d
els.
I
n
th
is
p
r
o
ce
s
s
,
th
e
p
r
ed
ictio
n
s
m
ad
e
b
y
ea
c
h
b
ase
lear
n
er
o
n
t
h
e
test
d
ata
ar
e
weig
h
ted
ac
co
r
d
in
g
to
th
e
o
p
tim
ized
co
ef
f
icien
ts
an
d
a
g
g
r
eg
ated
t
o
f
o
r
m
a
u
n
if
ied
e
n
s
em
b
le
p
r
ed
ictio
n
.
T
h
e
ag
g
r
e
g
ated
p
r
e
d
ictio
n
is
th
en
n
o
r
m
alize
d
to
en
s
u
r
e
it
r
e
p
r
esen
ts
a
v
alid
p
r
o
b
ab
ilit
y
d
i
s
tr
ib
u
tio
n
.
T
h
e
f
in
al
class
if
icatio
n
d
ec
is
io
n
is
m
ad
e
b
ased
o
n
th
e
class
with
th
e
h
ig
h
est
p
r
o
b
a
b
ilit
y
.
T
h
e
o
p
ti
m
ized
en
s
em
b
le’
s
p
er
f
o
r
m
an
ce
is
ass
es
s
ed
u
s
in
g
k
ey
m
etr
ics
s
u
ch
as
p
r
ec
is
io
n
,
F1
-
s
co
r
e,
ac
cu
r
ac
y
,
an
d
r
ec
all,
p
r
o
v
id
i
n
g
a
co
m
p
r
eh
e
n
s
iv
e
ass
ess
m
en
t
o
f
its
ef
f
ec
tiv
en
ess
in
d
etec
tin
g
v
ar
io
u
s
ty
p
es
o
f
cy
b
e
r
attac
k
s
in
th
e
I
o
T
n
etwo
r
k
en
v
ir
o
n
m
en
t.
2
.
4
.
1
.
B
utt
er
f
ly
o
ptim
iz
a
t
io
n
a
lg
o
rit
hm
T
h
e
B
OA
is
a
n
atu
r
e
-
in
s
p
ir
ed
m
eta
-
h
eu
r
is
tic
ap
p
r
o
ac
h
th
at
m
o
d
els
th
e
m
atin
g
an
d
f
o
r
ag
in
g
ac
tio
n
s
/b
eh
av
io
r
s
o
f
b
u
tter
f
lies
[
2
6
]
.
T
h
e
m
o
v
em
en
t
b
e
h
av
io
r
o
f
b
u
tter
f
lies
ca
n
b
e
ex
p
r
ess
ed
as
an
o
p
tim
izatio
n
p
r
o
ce
d
u
r
e,
wh
e
r
e
b
u
tter
f
lies
in
d
icate
s
ee
k
in
g
en
titi
es
an
d
g
en
e
r
ated
f
r
ag
r
a
n
ce
s
co
r
r
esp
o
n
d
to
f
itn
ess
v
alu
es.
I
n
th
e
B
OA,
b
u
tter
f
lies
(
i.e
.
,
s
ee
k
in
g
en
titi
es)
ca
n
p
r
o
d
u
ce
f
r
ag
r
an
ce
(
i.e
.
,
f
itn
ess
)
v
alu
es
with
d
is
tin
g
u
is
h
in
g
q
u
ality
o
v
er
o
th
er
f
r
ag
r
an
ce
s
,
wh
ich
is
r
ep
r
esen
ted
as
(
2
)
:
=
(
2
)
wh
er
e
is
th
e
f
r
ag
r
an
ce
co
n
ce
n
tr
atio
n
o
f
th
e
k
th
b
u
tter
f
ly
,
c
is
a
co
n
s
tan
t
s
ca
lin
g
f
ac
to
r
(
i.e
.
,
th
e
s
en
s
o
r
y
m
o
d
ality
)
,
I
k
is
th
e
s
tim
u
lu
s
i
n
ten
s
ity
,
an
d
is
th
e
f
itn
ess
v
alu
e
o
f
th
e
k
th
b
u
tter
f
ly
,
a
n
d
a
is
an
ex
p
o
n
en
tial
f
ac
to
r
th
at
d
eter
m
in
es th
e
s
h
ap
e
o
f
th
e
f
r
ag
r
a
n
ce
d
is
tr
ib
u
tio
n
d
ep
en
d
i
n
g
o
n
m
o
d
ality
.
T
h
is
b
eh
av
io
r
ca
n
ass
is
t
o
th
er
s
ea
r
ch
en
titi
es
in
u
p
d
atin
g
th
eir
p
o
s
itio
n
s
with
in
th
e
s
ea
r
ch
s
p
ac
e.
W
h
en
th
e
b
u
tter
f
ly
th
at
lo
ca
tes
th
e
o
p
tim
u
m
n
ec
tar
s
o
u
r
ce
in
th
e
s
ea
r
ch
ar
ea
r
elea
s
es
a
f
r
ag
r
an
ce
,
al
l
n
eig
h
b
o
r
in
g
b
u
tter
f
lies
will
f
l
y
to
th
at
b
u
tter
f
ly
’
s
p
o
s
itio
n
.
T
h
is
u
p
d
ate
p
r
o
ce
s
s
is
r
ef
er
r
e
d
to
as
g
lo
b
al
s
ea
r
ch
in
B
OA.
C
o
n
v
er
s
ely
,
b
u
tter
f
lies
will
r
an
d
o
m
ly
n
a
v
ig
at
e
th
e
s
ea
r
ch
s
p
ac
e
wh
en
d
if
f
er
en
t
b
u
tter
f
lies
’
f
r
ag
r
an
ce
s
a
r
e
d
is
co
v
e
r
ed
,
wh
ich
is
k
n
o
wn
as
lo
ca
l
s
ea
r
ch
in
B
OA.
T
h
e
p
o
s
itio
n
o
f
ea
ch
b
u
tter
f
ly
i
n
d
iv
id
u
al
is
r
ep
r
esen
ted
b
y
a
v
ec
t
o
r
o
f
p
ar
am
eter
v
al
u
es
co
r
r
esp
o
n
d
i
n
g
to
t
h
e
p
r
o
b
lem
b
ein
g
o
p
ti
m
ized
.
T
h
is
p
o
s
itio
n
ca
n
b
e
u
p
d
ated
wh
e
n
s
ee
k
in
g
to
f
i
n
d
a
m
o
r
e
o
p
tim
al
p
o
s
itio
n
with
in
t
h
e
s
ea
r
ch
s
p
ac
e
u
s
in
g
th
e
f
o
llo
win
g
m
ath
em
atica
l f
o
r
m
u
la:
+
1
=
+
+
1
(
3
)
wh
er
e
+
1
an
d
r
ep
r
esen
t th
e
ac
tu
al
p
o
s
itio
n
o
f
t
h
e
k
th
b
u
tter
f
l
y
at
iter
atio
n
s
t
+ 1
an
d
t
,
r
esp
ec
tiv
ely
,
an
d
is
th
e
f
r
ag
r
an
ce
u
s
ed
b
y
f
o
r
p
o
s
itio
n
u
p
d
ate
th
r
o
u
g
h
o
u
t t
h
e
iter
atio
n
s
.
As
s
tated
ea
r
lier
,
th
e
u
p
d
ate
p
r
o
ce
s
s
in
B
OA
is
g
o
v
er
n
e
d
b
y
two
k
ey
m
ec
h
a
n
is
m
s
:
lo
ca
l
an
d
g
lo
b
al
s
ea
r
ch
.
I
n
th
e
g
lo
b
al
m
o
d
e,
b
u
tter
f
lies
ar
e
attr
ac
ted
t
o
war
d
s
th
e
to
p
-
p
e
r
f
o
r
m
in
g
b
u
tt
er
f
ly
∗
,
wh
ich
is
m
o
d
eled
b
y
(
4
)
:
+
1
=
×
(
2
×
∗
−
)
(
4
)
with
a
n
u
m
er
ical
r
an
d
o
m
f
a
cto
r
an
d
∈
[
0
,
1
]
.
T
h
e
lo
ca
l
m
o
d
e
r
ef
in
es
th
e
r
esear
ch
b
y
ex
p
lo
i
tin
g
th
e
n
ea
r
b
y
p
r
o
m
is
in
g
ar
ea
s
an
d
is
m
o
d
eled
b
y
(
5
)
:
+
1
=
×
(
2
×
−
)
(
5
)
wh
er
e
an
d
s
tan
d
f
o
r
ℎ
an
d
ℎ
b
u
tter
f
lies
’
p
o
s
itio
n
s
in
th
e
s
ea
r
ch
s
p
ac
e.
I
n
th
e
co
n
tex
t
o
f
th
e
p
r
o
p
o
s
e
d
f
r
a
m
ewo
r
k
,
th
e
in
itializatio
n
p
h
ase
o
f
B
OA
in
v
o
lv
es
g
e
n
er
atin
g
a
d
iv
er
s
e
p
o
p
u
latio
n
o
f
b
u
tter
f
li
es,
wh
er
e
ea
ch
b
u
tter
f
ly
r
ep
r
e
s
en
ts
a
p
o
ten
tial
weig
h
t
v
ec
t
o
r
f
o
r
t
h
e
en
s
em
b
le’
s
b
ase
lear
n
er
s
.
T
h
e
f
itn
ess
o
f
ea
ch
weig
h
t
v
ec
to
r
is
ev
alu
ated
b
ased
o
n
th
e
en
s
em
b
le
m
o
d
el’
s
F1
-
s
co
r
e
wh
en
u
s
in
g
th
o
s
e
weig
h
ts
.
T
h
e
F1
-
s
co
r
e
m
etr
ic
was
s
elec
ted
d
u
e
to
its
co
m
p
r
eh
en
s
iv
e
p
e
r
f
o
r
m
a
n
ce
ev
alu
atio
n
a
n
d
ef
f
icac
y
in
h
an
d
lin
g
im
b
ala
n
c
ed
d
atasets
.
T
h
e
f
r
a
g
r
an
ce
ca
lcu
latio
n
th
en
tr
an
s
lates
th
ese
f
itn
ess
s
co
r
es
in
to
s
ig
n
als th
at
g
u
id
e
th
e
s
ea
r
ch
p
r
o
ce
s
s
,
d
r
awin
g
b
u
tter
f
lies
to
war
d
m
o
r
e
p
r
o
m
is
in
g
s
o
lu
tio
n
s
.
B
OA’
s
m
o
v
em
en
t
an
d
p
o
s
itio
n
u
p
d
ate
s
tep
s
em
p
lo
y
g
lo
b
al
a
n
d
lo
ca
l sear
c
h
s
tr
ateg
ies,
s
tr
ik
in
g
a
b
alan
ce
b
et
wee
n
ex
p
lo
r
i
n
g
n
e
w
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
25
:
3
4
9
4
-
3
5
0
5
3500
s
o
lu
tio
n
s
an
d
r
ef
in
in
g
k
n
o
wn
g
o
o
d
o
n
es.
T
h
is
iter
ativ
e
p
r
o
ce
s
s
en
s
u
r
es
a
th
o
r
o
u
g
h
s
ea
r
ch
f
o
r
th
e
o
p
tim
al
weig
h
t
co
m
b
in
atio
n
s
.
T
h
e
iter
atio
n
p
h
ase
r
ep
ea
ts
th
ese
s
tep
s
u
n
til
co
n
v
er
g
e
n
ce
cr
iter
ia
ar
e
m
et,
p
r
o
g
r
ess
iv
ely
im
p
r
o
v
in
g
th
e
s
o
lu
tio
n
s
.
T
h
e
f
in
al
o
p
tim
ized
weig
h
t
v
ec
to
r
r
ep
r
esen
ts
th
e
b
est
co
m
b
in
ati
o
n
f
o
u
n
d
,
wh
ich
is
th
en
u
s
ed
to
c
o
m
b
in
e
th
e
b
asi
c
lear
n
er
s
’
p
r
e
d
ictio
n
s
.
2
.
4
.
2
.
O
ptim
ized
ens
em
ble mo
del
T
h
e
en
s
em
b
le
p
r
ed
ictio
n
with
in
th
e
B
OE
L
f
r
am
ewo
r
k
co
m
b
in
es
th
e
o
u
tp
u
ts
o
f
th
e
m
u
ltip
le
b
ase
lear
n
er
s
with
ea
ch
b
ase
lear
n
er
’
s
p
r
ed
ictio
n
,
weig
h
ted
b
y
its
co
r
r
esp
o
n
d
in
g
o
p
tim
ized
weig
h
t
∗
,
d
eter
m
in
ed
th
r
o
u
g
h
B
OA.
T
h
e
co
m
b
in
ed
e
n
s
em
b
le
p
r
ed
i
ctio
n
∗
,
is
th
e
weig
h
ted
s
u
m
o
f
b
ase
lear
n
er
p
r
ed
ictio
n
s
:
∗
=
∑
∗
∙
=
1
(
6
)
wh
er
e
is
th
e
n
u
m
b
er
o
f
b
ase
lear
n
er
s
.
T
h
is
ag
g
r
e
g
ated
p
r
ed
ictio
n
is
th
en
n
o
r
m
alize
d
to
f
o
r
m
a
v
ali
d
p
r
o
b
a
b
ilit
y
d
is
tr
ib
u
tio
n
:
∗
=
∗
∑
∗
=
1
(
7
)
wh
er
e
is
th
e
n
u
m
b
e
r
o
f
class
es.
T
h
e
f
in
al
class
lab
el
is
d
ec
id
ed
b
ased
o
n
th
e
s
elec
tio
n
o
f
th
e
class
h
av
in
g
th
e
h
ig
h
est p
r
o
b
ab
ilit
y
in
th
e
n
o
r
m
alize
d
p
r
ed
ictio
n
:
̂
=
a
r
g
ma
x
∗
(
8
)
T
h
is
p
r
o
ce
s
s
lev
er
ag
es
th
e
s
tr
en
g
th
s
o
f
ea
ch
b
ase
lear
n
er
,
a
s
d
eter
m
in
ed
b
y
t
h
e
o
p
tim
ize
d
weig
h
ts
,
r
esu
ltin
g
in
a
h
ig
h
l
y
ac
cu
r
ate
a
n
d
r
o
b
u
s
t p
r
ed
ictio
n
m
o
d
el.
2
.
4
.
3
.
E
v
a
lua
t
io
n sta
g
e
A
co
m
p
r
eh
e
n
s
iv
e
ev
alu
atio
n
is
cr
u
cial
to
u
n
d
e
r
s
tan
d
in
g
th
e
ef
f
icac
y
o
f
th
e
e
n
s
em
b
le
m
o
d
el,
en
s
u
r
in
g
its
r
eliab
le
d
etec
tio
n
o
f
d
iv
er
s
e
cy
b
er
attac
k
ty
p
es
with
r
o
b
u
s
t
p
er
f
o
r
m
an
ce
m
etr
ics.
T
h
e
ev
alu
atio
n
o
f
th
e
B
OE
L
f
r
am
ewo
r
k
ass
ess
es
th
e
o
p
tim
ized
en
s
em
b
le
m
o
d
el’
s
p
er
f
o
r
m
an
ce
u
s
in
g
v
ar
io
u
s
ass
ess
m
en
t
m
ea
s
u
r
es.
Key
m
etr
ics,
s
u
ch
as
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
a
ll,
an
d
F1
-
s
co
r
e,
co
m
p
r
eh
e
n
s
iv
ely
ev
alu
ate
th
e
m
o
d
el’
s
ef
f
ec
tiv
e
n
ess
in
d
ete
ctin
g
cy
b
e
r
attac
k
s
.
T
h
ese
p
r
i
m
ar
y
ev
al
u
atio
n
m
etr
ics
ar
e
d
etailed
in
T
ab
le
1
.
wh
er
e
r
ep
r
esen
ts
tr
u
e
p
o
s
itiv
es,
r
ep
r
esen
ts
tr
u
e
n
eg
ativ
es,
r
ep
r
esen
ts
f
alse
p
o
s
itiv
es,
an
d
r
ep
r
esen
ts
f
alse n
eg
ativ
es.
T
ab
le
1
.
E
v
alu
atio
n
m
et
r
ics f
o
r
B
OE
L
f
r
am
ewo
r
k
M
e
t
r
i
c
F
o
r
mu
l
a
D
e
scri
p
t
i
o
n
A
c
c
u
r
a
c
y
+
TP
+
TN
+
FP
+
FN
∗
100%
mea
s
u
r
e
d
a
s
t
h
e
p
r
o
p
o
r
t
i
o
n
o
f
c
o
r
r
e
c
t
l
y
c
l
a
ss
i
f
i
e
d
sam
p
l
e
s
t
o
t
h
e
t
o
t
a
l
n
u
mb
e
r
o
f
samp
l
e
s
P
r
e
c
i
s
i
o
n
TP
+
FP
∗
100%
r
e
f
l
e
c
t
s
t
h
e
m
o
d
e
l
’
s
a
b
i
l
i
t
y
t
o
a
c
c
u
r
a
t
e
l
y
i
d
e
n
t
i
f
y
p
o
si
t
i
v
e
i
n
st
a
n
c
e
s
R
e
c
a
l
l
TP
+
FN
∗
100%
a
l
s
o
r
e
f
e
r
r
e
d
t
o
a
s
se
n
si
t
i
v
i
t
y
,
r
e
p
r
e
se
n
t
s t
h
e
m
o
d
e
l
’
s c
a
p
a
c
i
t
y
t
o
d
e
t
e
c
t
a
l
l
p
o
si
t
i
v
e
i
n
s
t
a
n
c
e
s
F1
-
sc
o
r
e
2
∙
Pr
e
c
i
s
i
o
n
·
R
e
c
a
ll
Pr
e
c
i
si
o
n
+
R
e
c
a
ll
∗
100%
i
s t
h
e
h
a
r
mo
n
i
c
m
e
a
n
o
f
p
r
e
c
i
si
o
n
a
n
d
r
e
c
a
l
l
a
n
d
p
r
o
v
i
d
e
s
a
b
a
l
a
n
c
e
d
a
ss
e
ss
men
t
o
f
t
h
e
m
o
d
e
l
’
s
o
v
e
r
a
l
l
p
e
r
f
o
r
ma
n
c
e
.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
is
s
ec
tio
n
p
r
esen
ts
th
e
r
esu
l
ts
f
r
o
m
ex
p
e
r
im
en
ts
co
n
d
u
cte
d
o
n
C
I
C
I
DS2
0
1
7
an
d
B
o
t
-
I
o
T
d
atasets
u
s
in
g
th
e
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
.
T
o
ass
ess
th
e
s
y
s
tem
’
s
ef
f
ec
tiv
en
ess
,
we
co
m
p
ar
ed
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
B
OE
L
f
r
am
ewo
r
k
ag
ain
s
t
f
o
u
r
wid
ely
a
d
o
p
te
d
ML
g
r
ad
ie
n
t
-
b
o
o
s
tin
g
al
g
o
r
ith
m
s
as
b
ase
lear
n
er
s
a
n
d
o
th
er
co
m
m
o
n
l
y
u
s
ed
s
tate
-
of
-
th
e
-
a
r
t
ML
alg
o
r
ith
m
s
.
Du
e
to
th
e
in
h
er
e
n
t
im
b
ala
n
ce
in
I
o
T
n
etwo
r
k
tr
af
f
ic
d
ata,
wh
er
e
attac
k
s
am
p
les
o
f
ten
co
n
s
titu
te
a
s
m
all
f
r
ac
tio
n
o
f
th
e
o
v
er
all
d
ata,
th
e
ev
alu
atio
n
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el’
s
p
er
f
o
r
m
a
n
ce
u
tili
ze
s
f
o
u
r
k
ey
m
etr
ics:
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
s
co
r
e.
Fu
r
th
er
m
o
r
e,
th
e
ex
p
er
im
en
tal
o
u
tco
m
es
wer
e
v
alid
ated
b
y
c
o
m
p
ar
in
g
th
e
m
with
f
in
d
in
g
s
f
r
o
m
r
elev
an
t
r
ec
en
t
s
tu
d
ies.
T
h
e
r
esu
lts
ar
e
p
r
esen
ted
in
a
ta
b
u
l
ar
an
d
f
ig
u
r
ativ
e
f
o
r
m
at
a
n
d
e
v
alu
ated
u
s
in
g
t
h
e
m
etr
ics d
is
cu
s
s
ed
ea
r
lier
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
B
u
tter
fly
o
p
timiz
a
tio
n
-
b
a
s
ed
e
n
s
emb
le
lea
r
n
in
g
s
tr
a
teg
y
fo
r
a
d
va
n
ce
d
…
(
Mo
u
a
d
C
h
o
u
kh
a
ir
i
3501
3
.
1
.
E
x
perim
ent
a
l set
up
T
h
e
ex
p
e
r
im
en
ts
wer
e
co
n
d
u
cted
u
s
in
g
th
e
5
th
iter
atio
n
o
f
th
e
Go
o
g
le
C
o
llab
o
r
at
o
r
y
e
n
v
ir
o
n
m
en
t,
wh
ich
p
r
o
v
id
es
f
r
ee
ac
ce
s
s
to
r
o
b
u
s
t
co
m
p
u
tatio
n
al
r
eso
u
r
c
es,
in
clu
d
in
g
h
ig
h
-
p
er
f
o
r
m
a
n
c
e
GPUs
an
d
ten
s
o
r
p
r
o
ce
s
s
in
g
u
n
its
(
T
PUs
)
,
ess
en
tial
f
o
r
th
e
s
tu
d
y
.
T
h
is
p
latf
o
r
m
o
f
f
er
s
m
em
o
r
y
ca
p
ac
ities
o
f
u
p
to
1
3
GB
o
f
R
AM
an
d
I
n
tel
Xeo
n
C
PU
p
r
o
ce
s
s
o
r
s
with
2
v
ir
tu
al
ce
n
tr
al
p
r
o
ce
s
s
in
g
u
n
its
,
m
ak
in
g
it
r
ep
r
esen
tativ
e
o
f
an
I
o
T
m
ac
h
in
e
in
ter
m
s
o
f
p
r
o
ce
s
s
in
g
ca
p
ab
ilit
ies.
T
o
d
ev
el
o
p
an
d
ev
al
u
ate
th
e
f
r
am
ewo
r
k
,
wid
ely
u
s
ed
Py
th
o
n
lib
r
ar
ies
f
o
r
ML
,
s
u
ch
as
L
ig
h
tGB
M,
XG
B
o
o
s
t,
C
at
B
o
o
s
t,
an
d
Scik
it
-
lear
n
,
wer
e
em
p
lo
y
ed
.
T
h
ese
lib
r
ar
ies
o
f
f
er
ed
th
e
n
ec
ess
ar
y
t
o
o
ls
f
o
r
ef
f
icie
n
t
m
o
d
el
tr
ai
n
in
g
,
test
in
g
,
an
d
co
m
p
r
eh
e
n
s
iv
e
ev
alu
atio
n
.
T
h
e
in
teg
r
atio
n
with
Go
o
g
le
Dr
iv
e
f
ac
ilit
ated
th
e
m
an
ag
em
e
n
t
o
f
d
atasets
an
d
r
esu
lts
,
en
ab
lin
g
a
s
m
o
o
th
an
d
p
r
o
d
u
ctiv
e
r
esear
ch
w
o
r
k
f
l
o
w.
3
.
2
.
Co
m
pa
ra
t
iv
e
a
na
ly
s
is
a
nd
ev
a
lua
t
io
n
T
h
e
ex
p
er
im
e
n
tal
r
esu
lts
p
r
esen
ted
in
Fig
u
r
es 2
an
d
3
c
o
m
p
a
r
e
th
e
p
er
f
o
r
m
an
ce
o
f
f
o
u
r
b
ase
lear
n
er
s
GB
M,
C
atB
o
o
s
t,
X
GB
o
o
s
t,
an
d
L
ig
h
tGB
M
an
d
th
e
p
r
o
p
o
s
ed
B
OE
L
f
r
am
ewo
r
k
o
n
B
o
t
-
I
o
T
an
d
C
I
C
I
DS2
0
1
7
d
atasets
.
T
h
e
F1
-
s
co
r
e
p
lo
ts
in
Fig
u
r
es
2
an
d
3
illu
s
tr
ate
th
e
v
ar
y
in
g
d
etec
tio
n
ca
p
ab
ilit
ies
o
f
th
e
in
d
iv
id
u
al
b
ase
m
o
d
els
f
o
r
d
if
f
er
en
t
ty
p
es
o
f
attac
k
s
ac
r
o
s
s
th
e
two
d
atasets
.
T
h
ese
r
esu
lts
h
ig
h
lig
h
t
th
e
im
p
o
r
tan
ce
o
f
en
s
em
b
le
o
p
tim
izatio
n
,
s
h
o
win
g
h
o
w
co
m
b
i
n
in
g
m
u
ltip
le
m
o
d
els
ca
n
lead
to
m
o
r
e
r
o
b
u
s
t
an
d
ac
cu
r
ate
in
tr
u
s
io
n
d
etec
tio
n
in
I
o
T
en
v
ir
o
n
m
en
ts
.
Fig
u
r
e
2
.
Mo
d
els p
er
f
o
r
m
an
ce
co
m
p
ar
is
o
n
b
y
ca
te
g
o
r
y
o
n
C
I
C
I
DS2
0
1
7
d
ataset
Fig
u
r
e
3
.
Mo
d
els p
er
f
o
r
m
an
ce
co
m
p
ar
is
o
n
b
y
ca
te
g
o
r
y
o
n
B
o
t
-
I
o
T
d
ataset
T
h
e
ex
p
er
im
en
tal
r
esu
lts
d
em
o
n
s
tr
ate
th
e
v
ar
y
in
g
p
er
f
o
r
m
a
n
ce
o
f
th
e
in
d
iv
id
u
al
b
ase
lea
r
n
er
m
o
d
els
o
n
th
e
C
I
C
I
DS2
0
1
7
an
d
B
o
t
-
I
o
T
d
atasets
.
On
th
e
C
I
C
I
D
S2
0
1
7
d
ataset,
L
ig
h
tGB
M
co
n
s
is
ten
tly
ac
h
iev
ed
th
e
h
ig
h
est
F1
-
s
co
r
e
ac
r
o
s
s
s
ev
er
al
attac
k
ca
teg
o
r
ies,
in
clu
d
in
g
n
o
r
m
al
tr
af
f
ic,
b
r
u
te
-
f
o
r
ce
attac
k
s
,
Do
S,
in
f
iltra
tio
n
,
s
n
if
f
in
g
,
an
d
web
attac
k
s
.
Ho
wev
er
,
XGBo
o
s
t
an
d
C
atB
o
o
s
t
wer
e
m
o
r
e
ef
f
ec
t
iv
e
th
an
L
ig
h
tGB
M
in
d
etec
tin
g
b
o
tn
et
attac
k
s
.
Similar
ly
,
o
n
th
e
B
o
t
-
I
o
T
d
ataset,
L
ig
h
tGB
M
o
u
tp
er
f
o
r
m
ed
th
e
o
th
er
b
ase
lear
n
er
s
,
p
ar
tic
u
lar
ly
in
th
e
DDo
S
an
d
D
o
S
attac
k
class
es,
wh
er
e
it
attain
ed
th
e
h
ig
h
est
F1
-
s
co
r
e
o
f
9
9
.
9
2
%
f
o
r
b
o
th
.
XGBo
o
s
t
also
ex
h
ib
ited
s
tr
o
n
g
p
er
f
o
r
m
a
n
ce
,
with
F1
-
s
co
r
es
o
f
9
9
.
9
0
%
an
d
9
9
.
9
1
%
in
th
ese
two
attac
k
ca
teg
o
r
ies.
W
h
ile
C
atB
o
o
s
t
p
er
f
o
r
m
ed
well
with
n
o
r
m
al
s
am
p
les
an
d
r
ec
o
n
n
aiss
an
ce
attac
k
s
,
s
u
r
p
ass
in
g
XGBo
o
s
t,
it
g
en
er
ally
ac
h
iev
ed
s
lig
h
tly
lo
wer
s
co
r
es
th
an
L
ig
h
tGB
M
an
d
X
GB
o
o
s
t
in
th
e
o
th
e
r
attac
k
class
e
s
,
ex
ce
p
t
f
o
r
th
e
th
ef
t
ca
teg
o
r
y
,
wh
e
r
e
it
attain
ed
a
p
er
f
ec
t
F1
-
s
co
r
e
o
f
1
0
0
%,
m
atch
in
g
th
e
to
p
-
p
er
f
o
r
m
in
g
m
o
d
els.
Ov
er
all,
L
ig
h
tGB
M
an
d
XGBo
o
s
t
em
er
g
ed
as
th
e
lead
in
g
m
o
d
els
i
n
b
o
th
e
x
p
er
im
en
ts
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8
7
0
8
I
n
t J E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
15
,
No
.
3
,
J
u
n
e
20
25
:
3
4
9
4
-
3
5
0
5
3502
d
em
o
n
s
tr
atin
g
s
u
p
e
r
io
r
p
e
r
f
o
r
m
an
ce
ac
r
o
s
s
th
e
m
ajo
r
ity
o
f
t
h
e
ev
alu
ated
attac
k
class
es c
o
m
p
ar
ed
to
C
atB
o
o
s
t
an
d
GB
M.
As
s
h
o
wn
in
Fig
u
r
es
2
an
d
3
,
th
e
p
r
o
p
o
s
ed
B
OE
L
m
o
d
el
ac
h
iev
e
d
th
e
h
ig
h
e
s
t
F1
-
s
co
r
e
f
o
r
ea
ch
ca
teg
o
r
y
,
d
em
o
n
s
tr
atin
g
its
s
u
p
er
io
r
p
er
f
o
r
m
a
n
ce
in
in
tr
u
s
io
n
d
etec
tio
n
ac
r
o
s
s
th
e
ev
al
u
ated
I
o
T
n
etwo
r
k
en
v
ir
o
n
m
en
ts
.
T
ab
le
2
d
em
o
n
s
tr
ates
a
clea
r
p
r
o
g
r
ess
io
n
in
p
er
f
o
r
m
a
n
ce
ac
r
o
s
s
th
e
ev
alu
ate
d
ML
m
o
d
els
o
n
t
h
e
C
I
C
I
DS2
0
1
7
.
T
h
e
s
tate
-
of
-
t
h
e
-
ar
t
m
o
d
els
r
a
n
g
ed
f
r
o
m
t
r
a
d
itio
n
al
ML
tech
n
iq
u
es
lik
e
DT
an
d
Ad
aBo
o
s
t
to
m
o
r
e
ad
v
an
ce
d
m
eth
o
d
s
s
u
ch
as
d
ee
p
b
elief
n
etwo
r
k
s
(
DB
N)
an
d
r
ec
u
r
r
e
n
t
n
eu
r
al
n
etwo
r
k
s
(
R
NN)
,
as
well
as
cu
ttin
g
-
ed
g
e
g
r
ad
ie
n
t
-
b
o
o
s
tin
g
alg
o
r
ith
m
s
lik
e
GB
M,
C
atB
o
o
s
t,
XG
B
o
o
s
t,
an
d
L
ig
h
tGB
M.
Am
o
n
g
th
e
m
o
d
els,
R
NN
ex
h
ib
ited
s
tr
o
n
g
p
e
r
f
o
r
m
an
ce
,
ac
h
iev
in
g
a
n
ac
cu
r
ac
y
o
f
9
8
%
an
d
an
F1
-
s
co
r
e
o
f
9
6
%,
o
u
tp
er
f
o
r
m
in
g
DT
an
d
Ad
a
B
o
o
s
t,
wh
ich
h
ad
l
o
wer
ac
c
u
r
ac
y
a
n
d
p
r
ec
is
io
n
.
No
tab
l
y
,
Ad
aBo
o
s
t
h
ad
a
n
ac
cu
r
ac
y
o
f
o
n
ly
8
1
.
8
3
%
b
u
t
co
m
p
e
n
s
ated
with
a
p
er
f
ec
t
r
ec
all
o
f
1
0
0
%.
DB
N
also
d
eliv
er
e
d
r
o
b
u
s
t
p
er
f
o
r
m
an
ce
,
attain
in
g
an
ac
c
u
r
ac
y
o
f
9
8
.
9
5
%,
alth
o
u
g
h
its
F1
-
s
co
r
e
was
s
lig
h
tly
lo
we
r
th
an
th
at
o
f
R
NN,
in
d
icatin
g
a
tr
ad
e
-
o
f
f
b
etwe
en
p
r
ec
is
io
n
an
d
r
ec
all.
T
h
e
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
h
as
en
h
an
ce
d
p
er
f
o
r
m
a
n
ce
m
etr
ics,
attain
in
g
9
9
.
7
9
5
%
in
ac
cu
r
ac
y
an
d
9
9
.
7
9
2
%
in
F1
-
s
co
r
e,
m
a
k
in
g
it
th
e
t
o
p
-
p
er
f
o
r
m
in
g
ap
p
r
o
ac
h
in
th
e
ex
p
er
im
e
n
t.
Sp
ec
if
ically
,
t
h
e
en
s
em
b
le
m
o
d
el
o
u
t
p
er
f
o
r
m
ed
R
NN
b
y
1
.
7
9
5
%
in
ac
c
u
r
ac
y
an
d
3
.
7
9
2
%
in
F1
-
s
co
r
e,
an
d
DB
N
b
y
0
.
8
4
5
%
in
ac
cu
r
ac
y
an
d
3
.
9
8
2
%
in
F1
-
s
co
r
e.
Mo
r
eo
v
er
,
t
h
e
en
s
e
m
b
le
m
o
d
el
s
lig
h
tly
s
u
r
p
ass
ed
th
e
ad
v
an
ce
d
L
ig
h
t
GB
M
m
eth
o
d
,
ac
h
iev
in
g
a
0
.
0
1
9
% im
p
r
o
v
em
e
n
t in
b
o
th
ac
c
u
r
ac
y
an
d
F1
-
s
co
r
e.
Ad
d
itio
n
ally
,
wh
ile
XGBo
o
s
t
an
d
C
atB
o
o
s
t
also
ex
h
ib
ited
ex
ce
llen
t
p
er
f
o
r
m
a
n
ce
,
with
ac
cu
r
ac
ies
o
f
9
9
.
7
5
7
%
an
d
9
9
.
6
8
3
%,
r
esp
e
ctiv
ely
,
th
e
p
r
o
p
o
s
ed
en
s
em
b
l
e
m
o
d
el
s
till
o
u
tp
er
f
o
r
m
ed
th
em
b
y
0
.
0
3
8
%
an
d
0
.
1
1
2
% in
ac
cu
r
ac
y
an
d
F1
-
s
co
r
e.
T
h
e
ex
p
er
im
e
n
tal
ev
alu
atio
n
,
as
p
r
esen
ted
in
T
ab
le
3
,
o
n
th
e
B
o
t
-
I
o
T
d
ataset,
s
h
o
ws
a
d
is
tin
ct
h
ier
ar
ch
y
in
t
h
e
p
e
r
f
o
r
m
an
ce
o
f
v
ar
io
u
s
ML
m
o
d
els,
g
o
in
g
f
r
o
m
tr
a
d
itio
n
al
m
et
h
o
d
s
li
k
e
SVM
an
d
R
F
to
m
o
r
e
ad
v
a
n
ce
d
ap
p
r
o
ac
h
es
s
u
ch
as
L
S
-
DR
NN,
G
B
M,
C
atB
o
o
s
t,
XGBo
o
s
t,
an
d
L
ig
h
tGB
M.
Am
o
n
g
th
e
tr
ad
itio
n
al
m
eth
o
d
s
,
SVM
ac
h
iev
ed
th
e
lo
west
s
co
r
e
wi
th
an
ac
cu
r
ac
y
o
f
8
9
.
3
5
%
an
d
an
F1
-
s
co
r
e
o
f
8
9
.
3
4
%.
T
h
e
R
F
m
o
d
el,
wh
ile
s
tr
o
n
g
in
ce
r
tain
asp
ec
ts
,
p
ar
ticu
lar
ly
in
p
r
ec
is
io
n
,
lag
g
e
d
with
an
a
cc
u
r
ac
y
o
f
9
8
%
an
d
an
F1
-
s
co
r
e
o
f
9
8
%.
T
h
e
L
S
-
DR
NN
m
o
d
el,
d
esig
n
ed
s
p
e
cif
ically
f
o
r
d
ee
p
lea
r
n
in
g
o
n
tim
e
s
er
ies
d
ata,
d
eliv
er
ed
im
p
r
ess
iv
e
r
esu
lts
with
an
ac
c
u
r
ac
y
o
f
9
9
.
9
3
%
a
n
d
a
n
F1
-
s
co
r
e
o
f
9
8
.
2
2
%,
alt
h
o
u
g
h
its
p
r
ec
is
io
n
was
lo
wer
co
m
p
ar
e
d
to
t
h
e
o
th
er
ad
v
an
ce
d
m
o
d
els.
W
h
en
co
m
p
a
r
ed
t
o
th
e
g
r
ad
ien
t
b
o
o
s
tin
g
m
o
d
els,
L
ig
h
tGB
M
s
to
o
d
o
u
t
with
an
ac
cu
r
ac
y
o
f
9
9
.
9
1
6
%
an
d
an
F
1
-
s
co
r
e
o
f
9
9
.
9
1
6
%,
s
u
r
p
ass
in
g
GB
M,
C
atB
o
o
s
t,
an
d
XGBo
o
s
t,
wh
ich
ac
h
iev
e
d
ac
cu
r
ac
ies
o
f
9
9
.
7
8
2
%,
9
9
.
8
4
9
%,
an
d
9
9
.
8
6
6
%,
r
esp
ec
ti
v
ely
.
Ho
wev
er
,
t
h
e
p
r
o
p
o
s
ed
B
OE
L
f
r
am
ewo
r
k
f
u
r
th
er
elev
ated
p
er
f
o
r
m
a
n
ce
,
ac
h
iev
in
g
an
ac
cu
r
ac
y
o
f
9
9
.
9
6
6
%
an
d
an
F1
-
s
co
r
e
o
f
9
9
.
9
6
6
%.
T
h
is
r
ep
r
esen
ts
an
im
p
r
o
v
e
m
en
t
o
f
0
.
0
5
%
o
v
er
L
ig
h
tGB
M,
0
.
1
0
%
o
v
er
XGBo
o
s
t,
0
.
1
1
7
%
o
v
er
C
atB
o
o
s
t,
an
d
0
.
1
8
4
%
o
v
er
GB
M.
C
o
m
p
ar
e
d
to
t
h
e
L
S
-
DR
NN
m
o
d
el,
B
OE
L
o
u
tp
er
f
o
r
m
ed
it
b
y
0
.
0
3
6
%
in
ac
cu
r
ac
y
an
d
a
s
ig
n
if
ican
t
1
.
7
4
6
%
in
th
e
F1
-
s
co
r
e,
d
em
o
n
s
tr
atin
g
th
e
r
o
b
u
s
tn
ess
an
d
ef
f
ec
tiv
en
ess
o
f
t
h
e
e
n
s
e
m
b
l
e
a
p
p
r
o
a
c
h
.
T
h
e
r
e
s
u
l
t
s
h
i
g
h
l
i
g
h
t
t
h
e
s
u
p
e
r
i
o
r
i
ty
o
f
B
O
E
L
i
n
a
c
h
i
e
v
i
n
g
n
e
a
r
-
p
e
r
f
e
c
t
c
l
a
s
s
i
f
i
c
a
ti
o
n
a
c
c
u
r
a
c
y
a
n
d
b
a
l
a
n
ce
d
p
r
e
c
is
i
o
n
a
n
d
r
e
c
a
ll
,
m
a
k
i
n
g
i
t
a
p
o
w
e
r
f
u
l
t
e
c
h
n
i
q
u
e
f
o
r
I
o
T
in
t
r
u
s
i
o
n
d
e
t
ec
t
i
o
n
,
e
s
p
e
c
i
al
l
y
w
h
e
n
c
o
m
p
a
r
e
d
t
o
t
r
a
d
i
t
i
o
n
a
l
M
L
m
o
d
el
s
a
n
d
m
o
r
e
a
d
v
a
n
c
e
d
g
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
t
ec
h
n
i
q
u
e
s
.
T
ab
le
2
.
E
v
alu
atio
n
o
f
d
if
f
er
e
n
t m
o
d
els’
p
e
r
f
o
r
m
an
ce
u
s
in
g
C
I
C
I
DS2
0
1
7
d
ataset
A
p
p
r
o
a
c
h
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
s
i
o
n
(
%)
R
e
c
a
l
l
(
%)
F1
-
sc
o
r
e
(
%)
D
T
[
2
7
]
9
6
.
6
7
9
7
.
5
85
90
A
d
a
b
o
o
st
[
2
8
]
8
1
.
8
3
8
1
.
8
3
1
0
0
9
0
.
0
1
R
N
N
[
2
9
]
98
96
97
96
D
B
N
[
3
0
]
9
8
.
9
5
9
5
.
8
2
9
5
.
8
1
9
5
.
8
1
G
B
M
9
9
.
6
6
4
9
9
.
6
6
5
9
9
.
6
6
4
9
9
.
6
6
1
C
a
t
B
o
o
s
t
9
9
.
6
8
3
9
9
.
6
8
4
9
9
.
6
8
3
9
9
.
6
8
X
G
B
o
o
st
9
9
.
7
5
7
9
9
.
7
5
8
9
9
.
7
5
7
9
9
.
7
5
5
Li
g
h
t
G
B
M
9
9
.
7
7
6
9
9
.
7
7
7
9
9
.
7
7
6
9
9
.
7
7
3
Pr
o
p
o
sed
B
O
E
L
9
9
.
7
9
5
9
9
.
7
9
6
9
9
.
7
9
5
9
9
.
7
9
2
T
ab
le
3
.
E
v
alu
atio
n
o
f
d
if
f
er
e
n
t m
o
d
els’
p
e
r
f
o
r
m
an
ce
u
s
in
g
B
o
t
-
I
o
T
d
ataset
A
p
p
r
o
a
c
h
A
c
c
u
r
a
c
y
(
%)
P
r
e
c
i
s
i
o
n
(
%)
R
e
c
a
l
l
(
%)
F1
-
sc
o
r
e
(
%)
S
V
M
[
3
1
]
8
9
.
3
5
8
9
.
6
8
9
.
3
5
8
9
.
3
4
R
F
[
3
2
]
98
96
96
98
LS
-
D
R
N
N
[
3
3
]
9
9
.
9
3
9
6
.
8
7
9
9
.
7
5
9
8
.
2
2
G
B
M
9
9
.
7
8
2
9
9
.
7
8
2
9
9
.
7
8
9
9
.
7
8
2
C
a
t
B
o
o
s
t
9
9
.
8
4
9
9
9
.
8
4
9
9
9
.
8
4
9
9
.
8
4
9
X
G
B
o
o
st
9
9
.
8
6
6
9
9
.
8
6
6
9
9
.
8
6
9
9
.
8
6
6
Li
g
h
t
G
B
M
9
9
.
9
1
6
9
9
.
9
1
6
9
9
.
9
1
9
9
.
9
1
6
Pr
o
p
o
sed
B
O
E
L
9
9
.
9
6
6
9
9
.
9
6
6
9
9
.
9
6
9
9
.
9
6
6
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8
7
0
8
B
u
tter
fly
o
p
timiz
a
tio
n
-
b
a
s
ed
e
n
s
emb
le
lea
r
n
in
g
s
tr
a
teg
y
fo
r
a
d
va
n
ce
d
…
(
Mo
u
a
d
C
h
o
u
kh
a
ir
i
3503
4.
CO
NCLU
SI
O
N
T
o
e
n
h
a
n
c
e
I
o
T
n
e
t
w
o
r
k
s
e
c
u
r
i
t
y
,
t
h
is
p
a
p
e
r
i
n
t
r
o
d
u
c
e
s
a
n
o
v
e
l
e
n
s
e
m
b
l
e
le
a
r
n
i
n
g
a
p
p
r
o
a
c
h
c
a
l
l
e
d
t
h
e
B
O
E
L
f
r
a
m
ew
o
r
k
f
o
r
I
D
S
,
w
h
i
c
h
p
r
o
v
i
d
e
s
a
r
o
b
u
s
t
a
n
d
a
d
a
p
t
i
v
e
s
o
l
u
t
i
o
n
t
o
d
e
t
e
ct
v
a
r
i
o
u
s
t
y
p
e
s
o
f
c
y
b
e
r
a
t
t
a
c
k
s
.
T
h
e
p
r
o
p
o
s
e
d
m
o
d
e
l
c
o
m
b
i
n
e
s
M
L
-
b
a
s
e
d
t
e
c
h
n
i
q
u
e
s
,
p
a
r
t
ic
u
l
a
r
l
y
g
r
a
d
ie
n
t
-
b
o
o
s
t
i
n
g
a
l
g
o
r
it
h
m
s
,
i
n
c
l
u
d
i
n
g
GB
M
,
C
a
tB
o
o
s
t
,
L
i
g
h
t
G
B
M
,
a
n
d
X
GB
o
o
s
t
,
wi
t
h
t
h
e
B
OA
m
e
t
h
o
d
t
o
f
o
r
m
a
n
o
p
t
im
i
z
e
d
a
n
d
e
f
f
i
c
ie
n
t
e
n
s
e
m
b
l
e
m
o
d
e
l
.
E
x
p
e
r
i
m
e
n
t
s
c
o
n
d
u
c
t
e
d
o
n
C
I
C
I
DS
2
0
1
7
a
n
d
B
o
t
-
I
o
T
d
a
t
as
e
ts
d
e
m
o
n
s
t
r
ate
d
B
O
E
L
’
s
s
u
p
e
r
i
o
r
c
a
p
a
b
i
l
i
t
y
,
a
c
h
i
e
v
i
n
g
r
e
m
a
r
k
ab
l
e
a
c
c
u
r
a
c
ie
s
o
f
9
9
.
7
9
5
%
an
d
9
9
.
9
6
6
%
,
r
e
s
p
e
c
ti
v
e
l
y
,
s
u
r
p
a
s
s
i
n
g
i
n
d
i
v
i
d
u
a
l
m
o
d
e
l
s
a
n
d
s
t
a
te
-
of
-
t
h
e
-
a
r
t
t
ec
h
n
i
q
u
e
s
.
N
o
t
a
b
l
y
,
B
O
E
L
a
ch
i
e
v
e
d
t
h
e
h
i
g
h
e
s
t
p
e
r
f
o
r
m
a
n
c
e
o
n
t
h
e
B
o
t
-
I
o
T
d
a
t
a
s
et
,
w
it
h
a
p
r
e
ci
s
i
o
n
o
f
9
9
.
9
6
6
%
,
e
n
s
u
r
i
n
g
m
i
n
i
m
al
f
a
ls
e
p
o
s
i
t
i
v
es
;
a
r
e
c
al
l
e
x
c
ee
d
i
n
g
9
9
.
9
%
,
i
n
d
i
c
a
ti
n
g
h
i
g
h
s
e
n
s
it
i
v
it
y
t
o
s
u
b
tl
e
a
tt
a
ck
p
a
t
t
e
r
n
s
;
a
n
d
a
n
F
1
-
s
c
o
r
e
o
f
9
9
.
9
6
6
%
,
h
i
g
h
l
i
g
h
t
i
n
g
b
a
l
a
n
c
e
d
a
n
d
c
o
n
s
i
s
t
e
n
t
d
e
t
e
c
t
i
o
n
ca
p
a
b
i
l
it
i
es
a
c
r
o
s
s
d
i
v
e
r
s
e
at
t
a
c
k
c
a
t
e
g
o
r
i
es
a
n
d
u
n
d
e
r
s
c
o
r
i
n
g
i
ts
e
x
c
e
p
t
i
o
n
a
l
a
b
i
l
i
t
y
t
o
h
a
n
d
l
e
I
o
T
-
s
p
e
c
i
f
i
c
t
h
r
e
a
ts
.
F
u
r
t
h
e
r
m
o
r
e
,
t
h
e
f
r
a
m
e
w
o
r
k
e
x
c
e
l
l
e
d
i
n
i
d
e
n
t
i
f
y
i
n
g
l
o
w
-
f
r
e
q
u
e
n
c
y
at
ta
c
k
t
y
p
e
s
,
s
u
c
h
as
i
n
f
i
l
t
r
a
ti
o
n
a
n
d
s
n
i
f
f
i
n
g
,
w
h
i
ch
a
r
e
o
f
t
e
n
c
h
a
l
l
e
n
g
i
n
g
f
o
r
t
h
e
e
x
i
s
t
i
n
g
m
o
d
e
ls
.
B
O
E
L
’
s
d
y
n
a
m
i
c
w
e
i
g
h
t
i
n
g
m
e
c
h
a
n
i
s
m
f
u
r
t
h
e
r
e
n
h
a
n
c
e
d
i
ts
a
d
a
p
ta
b
i
l
it
y
t
o
v
a
r
y
i
n
g
a
t
t
ac
k
c
o
m
p
l
e
x
it
i
es
,
e
n
s
u
r
in
g
c
o
n
s
i
s
t
e
n
t
h
i
g
h
p
e
r
f
o
r
m
a
n
c
e
a
c
r
o
s
s
d
i
v
e
r
s
e
s
ce
n
a
r
i
o
s
.
T
h
es
e
f
i
n
d
i
n
g
s
u
n
d
e
r
s
c
o
r
e
B
O
E
L
’
s
p
o
t
e
n
ti
a
l
as
a
r
eli
a
b
l
e
,
e
f
f
i
ci
e
n
t
,
a
n
d
s
c
a
l
a
b
l
e
s
o
l
u
t
i
o
n
f
o
r
p
r
o
t
e
c
t
i
n
g
I
o
T
n
e
t
w
o
r
k
s
a
g
a
i
n
s
t
a
n
e
v
e
r
-
e
v
o
l
v
i
n
g
l
a
n
d
s
c
a
p
e
o
f
c
y
b
e
r
t
h
r
e
a
t
s
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
Au
th
o
r
s
s
tate
n
o
f
u
n
d
in
g
in
v
o
lv
ed
.
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
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Mo
u
ad
C
h
o
u
k
h
air
i
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Sar
a
T
ah
ir
i
✓
✓
✓
✓
✓
✓
Ou
ail
C
h
o
u
k
h
air
i
✓
✓
✓
✓
✓
✓
Yo
u
s
s
ef
Fak
h
r
i
✓
✓
✓
✓
✓
✓
✓
Mo
h
am
ed
Am
n
ai
✓
✓
✓
✓
✓
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
n
o
co
n
f
lict o
f
in
t
er
est.
I
NF
O
RM
E
D
CO
NS
E
N
T
T
h
is
s
tu
d
y
d
id
n
o
t in
v
o
lv
e
h
u
m
an
p
ar
ticip
an
ts
,
a
n
d
in
f
o
r
m
e
d
co
n
s
en
t w
as th
er
ef
o
r
e
n
o
t
r
e
q
u
ir
ed
.
E
T
H
I
CAL AP
P
RO
V
AL
T
h
is
r
esear
ch
d
id
n
o
t in
v
o
lv
e
h
u
m
an
o
r
an
im
al
s
u
b
jects a
n
d
d
id
n
o
t
r
eq
u
ir
e
eth
ical
ap
p
r
o
v
a
l.
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
i
s
s
tu
d
y
a
r
e
o
p
en
ly
a
v
ailab
le:
˗
T
h
e
C
I
C
I
DS2
0
1
7
d
ataset
is
av
ailab
le
at:
h
ttp
s
://www.
u
n
b
.
ca
/cic/d
atasets
/id
s
-
2
0
1
7
.
h
tm
l
˗
T
h
e
B
o
t
-
I
o
T
d
ataset
is
av
ailab
le
at:
h
ttp
s
://re
s
ea
r
ch
.
u
n
s
w.
ed
u
.
au
/p
r
o
jects/b
o
t
-
io
t
-
d
ataset
RE
F
E
R
E
NC
E
S
[
1
]
I
.
S
t
e
l
l
i
o
s,
P
.
K
o
t
z
a
n
i
k
o
l
a
o
u
,
M
.
P
sar
a
k
i
s,
C
.
A
l
c
a
r
a
z
,
a
n
d
J.
L
o
p
e
z
,
“
A
su
r
v
e
y
o
f
i
o
t
-
e
n
a
b
l
e
d
c
y
b
e
r
a
t
t
a
c
k
s:
A
sse
ssi
n
g
a
t
t
a
c
k
p
a
t
h
s
t
o
c
r
i
t
i
c
a
l
i
n
f
r
a
st
r
u
c
t
u
r
e
s
a
n
d
s
e
r
v
i
c
e
s
,
”
I
EE
E
C
o
m
m
u
n
i
c
a
t
i
o
n
s
S
u
rv
e
y
s
a
n
d
T
u
t
o
ri
a
l
s
,
v
o
l
.
2
0
,
n
o
.
4
,
p
p
.
3
4
5
3
–
3
4
9
5
,
2
0
1
8
,
d
o
i
:
Evaluation Warning : The document was created with Spire.PDF for Python.