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
.
4
,
A
u
g
u
s
t
20
25
,
p
p
.
3
8
5
1
~
3
8
6
6
I
SS
N:
2088
-
8
7
0
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijece.
v
15
i
4
.
pp
3
8
5
1
-
3
8
6
6
3851
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
M
a
chine learning
appro
a
ches to cy
bersecuri
ty in
the
industrial
interne
t
o
f
things
:
a revi
ew
M
ela
nie H
eier
1
,
P
ena
t
iy
a
na
W.
Cha
nd
a
na
P
ra
s
a
d
2
,
M
d
Sh
o
hel Sa
y
ee
d
3
1
S
c
h
o
o
l
o
f
C
o
m
p
u
t
i
n
g
,
M
a
t
h
e
mat
i
c
s
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
C
h
a
r
l
e
s
S
t
u
r
t
U
n
i
v
e
r
s
i
t
y
,
B
a
t
h
u
r
st
,
A
u
st
r
a
l
i
a
2
I
n
t
e
r
n
a
t
i
o
n
a
l
S
c
h
o
o
l
,
D
u
y
T
a
n
U
n
i
v
e
r
si
t
y
,
D
a
N
a
n
g
,
V
i
e
t
n
a
m
3
C
e
n
t
r
e
f
o
r
I
n
t
e
l
l
i
g
e
n
t
C
l
o
u
d
C
o
m
p
u
t
i
n
g
,
C
o
E
f
o
r
A
d
v
a
n
c
e
C
l
o
u
d
,
F
a
c
u
l
t
y
o
f
I
n
f
o
r
ma
t
i
o
n
S
c
i
e
n
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
M
u
l
t
i
me
d
i
a
U
n
i
v
e
r
s
i
t
y
,
M
e
l
a
k
a
,
M
a
l
a
y
si
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ma
y
5
,
2
0
2
4
R
ev
is
ed
Ap
r
1
2
,
2
0
2
5
Acc
ep
ted
Ma
y
2
3
,
2
0
2
5
Th
e
in
d
u
strial
in
tern
e
t
of
th
in
g
s
(IIo
T
)
is
in
c
re
a
sin
g
l
y
u
se
d
wit
h
i
n
v
a
rio
u
s
se
c
to
rs
to
p
r
o
v
i
d
e
in
n
o
v
a
ti
v
e
b
u
si
n
e
ss
so
lu
ti
o
n
s.
Th
e
se
tec
h
n
o
l
o
g
ica
l
in
n
o
v
a
ti
o
n
s
c
o
m
e
wit
h
a
d
d
it
i
o
n
a
l
c
y
b
e
rse
c
u
rit
y
r
isk
s,
a
n
d
m
a
c
h
in
e
lea
rn
in
g
(M
L)
is
an
e
m
e
rg
i
n
g
tec
h
n
o
lo
g
y
th
a
t
h
a
s
b
e
e
n
stu
d
ied
as
a
so
l
u
ti
o
n
to
th
e
se
c
o
m
p
lex
se
c
u
rit
y
c
h
a
ll
e
n
g
e
s.
At
t
ime
of
writi
n
g
,
to
t
h
e
a
u
t
h
o
r’s
k
n
o
wle
d
g
e
,
a
re
v
iew
of
re
c
e
n
t
stu
d
ies
on
t
h
is
to
p
ic
h
a
d
not
b
e
e
n
u
n
d
e
rtak
e
n
.
T
h
is
re
v
iew
th
e
re
fo
re
a
ims
to
p
ro
v
i
d
e
a
c
o
m
p
re
h
e
n
siv
e
p
ictu
re
of
th
e
c
u
rre
n
t
st
a
te
of
ML
so
lu
ti
o
n
s
fo
r
IIo
T
c
y
b
e
rse
c
u
rit
y
with
in
si
g
h
ts
in
t
o
w
h
a
t
wo
r
k
s
to
i
n
fo
rm
fu
tu
re
re
se
a
rc
h
or
re
a
l
-
wo
rld
s
o
lu
ti
o
n
s.
A
l
it
e
ra
ry
se
a
rc
h
fo
u
n
d
twe
lv
e
p
a
p
e
rs
to
re
v
iew
p
u
b
li
s
h
e
d
in
2
0
2
1
or
late
r
th
a
t
p
ro
p
o
se
d
ML
so
l
u
ti
o
n
s
to
IIo
T
c
y
b
e
rse
c
u
rit
y
c
o
n
c
e
rn
s.
Th
is
re
v
i
e
w
fo
u
n
d
th
a
t
fe
d
e
ra
t
ed
lea
rn
i
n
g
a
n
d
se
m
i
-
su
p
e
rv
ise
d
lea
rn
i
n
g
in
p
a
rti
c
u
l
a
r
a
re
p
ro
m
isin
g
ML
tec
h
n
iq
u
e
s
b
e
in
g
p
ro
p
o
se
d
to
c
o
m
b
a
t
th
e
c
o
n
c
e
rn
s
a
ro
u
n
d
II
o
T
c
y
b
e
rse
c
u
rit
y
.
Artifi
c
ial
n
e
u
ra
l
n
e
two
rk
a
p
p
ro
a
c
h
e
s
a
re
a
lso
c
o
m
m
o
n
ly
p
r
o
p
o
se
d
in
v
a
ri
o
u
s
c
o
m
b
in
a
ti
o
n
s
with
o
th
e
r
tec
h
n
iq
u
e
s
to
e
n
su
re
fa
st
a
n
d
a
c
c
u
ra
te
c
y
b
e
rse
c
u
rit
y
so
lu
t
io
n
s.
Wh
il
e
t
h
e
re
is
n
o
t
c
u
rre
n
tl
y
a
c
o
n
se
n
su
s
on
th
e
b
e
st
ML
tec
h
n
iq
u
e
s
to
a
p
p
ly
to
IIo
T
c
y
b
e
rse
c
u
rit
y
,
th
e
se
fin
d
in
g
s
o
ffe
r
i
n
sig
h
t
i
n
t
o
th
o
se
a
p
p
r
o
a
c
h
e
s
c
u
rre
n
tl
y
b
e
i
n
g
u
t
il
ize
d
a
l
o
n
g
with
g
a
p
s
w
h
e
re
fu
rt
h
e
r
e
x
a
m
i
n
a
ti
o
n
is
re
q
u
ired
.
K
ey
w
o
r
d
s
:
Ar
tific
ial
n
eu
r
al
n
etwo
r
k
s
C
y
b
er
s
ec
u
r
ity
Fed
er
ated
lear
n
in
g
I
n
d
u
s
tr
ial
in
ter
n
et
o
f
th
in
g
s
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
:
Md
Sh
o
h
el
Say
ee
d
C
en
tr
e
f
o
r
I
n
tellig
en
t Cl
o
u
d
C
o
m
p
u
tin
g
,
C
o
E
f
o
r
Ad
v
a
n
ce
d
C
lo
u
d
,
Facu
lty
o
f
I
n
f
o
r
m
atio
n
Scien
ce
an
d
T
ec
h
n
o
lo
g
y
,
Mu
ltime
d
ia
Un
iv
er
s
ity
J
alan
Ay
er
Ker
o
h
L
a
m
a,
7
5
4
5
0
B
u
k
it B
er
u
an
g
,
Me
la
k
a,
Ma
l
ay
s
ia
E
m
ail: sh
o
h
el.
s
ay
ee
d
@
m
m
u
.
e
d
u
.
m
y
1.
I
NT
RO
D
UCT
I
O
N
I
n
d
u
s
tr
ial
in
ter
n
et
of
th
in
g
s
(
I
I
o
T
)
r
e
f
er
s
to
th
e
n
etwo
r
k
o
f
in
ter
co
n
n
ec
ted
d
ev
ices,
m
a
ch
in
es
an
d
s
en
s
o
r
s
u
tili
ze
d
in
v
a
r
io
u
s
in
d
u
s
tr
ies
f
o
r
ac
tiv
ities
s
u
ch
a
s
au
to
m
atio
n
[
1
]
,
m
o
n
ito
r
in
g
,
co
n
tr
o
l,
an
d
d
ata
co
llectio
n
[
2
]
–
[
4
]
.
T
h
e
p
r
o
c
ess
o
p
tim
izatio
n
an
d
f
le
x
ib
ilit
y
p
r
o
v
id
ed
b
y
I
I
o
T
r
esu
lts
in
r
e
d
u
ce
d
co
s
ts
,
in
cr
ea
s
ed
p
r
o
d
u
ctio
n
,
a
n
d
im
p
r
o
v
e
d
ef
f
icien
cy
f
o
r
b
u
s
in
ess
es
o
r
s
er
v
ices
[
3
]
,
[
5
]
,
[
6
]
.
As
tech
n
o
lo
g
y
h
as
im
p
r
o
v
e
d
,
I
I
o
T
h
as b
ec
o
m
e
in
cr
ea
s
in
g
ly
u
tili
ze
d
f
o
r
v
ar
i
o
u
s
b
u
s
in
ess
an
d
in
d
u
s
tr
ial
p
r
o
ce
s
s
es.
T
h
e
I
I
o
T
p
r
o
v
id
es
a
u
n
iq
u
e
an
d
ch
allen
g
in
g
co
n
tex
t
f
o
r
cy
b
e
r
s
ec
u
r
i
ty
[
7
]
.
I
I
o
T
n
etwo
r
k
s
c
o
m
p
r
is
e
a
lar
g
e
n
u
m
b
e
r
o
f
in
ter
co
n
n
ec
te
d
d
ev
ices
with
g
r
ea
ter
life
s
p
an
s
th
an
co
n
s
u
m
er
d
ev
ices
[
4
]
,
[
8
]
.
T
h
ese
d
ev
ices
m
ay
n
ee
d
to
in
ter
ac
t
with
leg
ac
y
s
y
s
tem
s
,
p
u
ttin
g
th
em
at
r
is
k
[
9
]
.
T
h
e
y
p
r
o
d
u
ce
lar
g
e
am
o
u
n
ts
o
f
d
ata
[
5
]
an
d
p
er
f
o
r
m
cr
itical
b
u
s
in
ess
task
s
an
d
s
af
ety
f
u
n
ctio
n
s
[
1
0
]
–
[
1
2
]
.
Dev
ices
th
em
s
elv
es
as
well
as
th
eir
s
o
f
twar
e
m
ay
b
e
o
u
td
ated
,
lead
in
g
to
r
is
k
s
ass
o
ciate
d
with
a
lack
o
f
s
ec
u
r
ity
u
p
d
ates
[
8
]
.
I
I
o
T
d
ev
ices
ten
d
to
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
.
4
,
Au
g
u
s
t
20
25
:
3
8
5
1
-
3866
3852
h
av
e
lim
ited
r
eso
u
r
ce
s
in
ter
m
s
o
f
p
o
wer
a
n
d
m
em
o
r
y
,
a
n
d
s
o
cy
b
e
r
s
ec
u
r
ity
s
o
lu
tio
n
s
n
ee
d
to
h
a
v
e
lo
w
p
o
wer
an
d
lo
w
m
e
m
o
r
y
r
eq
u
ir
em
en
ts
[
5
]
,
[
1
3
]
,
[
1
4
]
.
T
h
e
s
e
lim
itatio
n
s
m
ea
n
th
at
s
o
lu
tio
n
s
m
u
s
t
also
b
e
s
ca
lab
le
an
d
ad
ap
tab
le
to
m
ee
t
b
u
s
in
ess
n
ee
d
s
an
d
h
av
e
th
e
ca
p
ac
ity
to
b
e
r
etr
o
f
itted
[
4
]
.
So
lu
tio
n
s
m
u
s
t
als
o
b
e
ab
le
to
p
r
o
ce
s
s
lar
g
e
am
o
u
n
ts
o
f
d
ata
q
u
ick
ly
an
d
ac
cu
r
a
tely
[
1
5
]
.
T
r
ad
itio
n
al
cy
b
er
s
ec
u
r
ity
s
o
lu
tio
n
s
ca
n
h
av
e
d
i
f
f
icu
lties
co
p
in
g
wit
h
th
e
u
n
iq
u
e
ch
allen
g
es
p
r
e
s
en
ted
b
y
I
I
o
T
[
1
6
]
.
T
r
ad
iti
o
n
al
cy
b
er
s
ec
u
r
ity
s
o
lu
tio
n
s
ca
n
also
r
eq
u
i
r
e
m
o
r
e
p
r
o
ce
s
s
in
g
p
o
wer
a
n
d
m
em
o
r
y
th
a
n
I
I
o
T
d
ev
ices
p
o
s
s
ess
,
cr
ea
tin
g
a
ch
all
en
g
in
g
e
n
v
ir
o
n
m
en
t
f
o
r
d
ev
ice
an
d
n
etwo
r
k
p
r
o
tectio
n
[
1
4
]
.
Ma
ch
in
e
lear
n
in
g
(
ML
)
is
o
n
e
o
f
th
e
em
er
g
in
g
tec
h
n
o
lo
g
ies
b
ein
g
u
tili
ze
d
to
s
o
lv
e
th
ese
cy
b
er
s
ec
u
r
ity
ch
allen
g
es.
As
tech
n
o
lo
g
y
h
as
ev
o
lv
e
d
,
c
y
b
er
-
attac
k
s
h
av
e
b
ec
o
m
e
p
r
o
g
r
ess
iv
ely
m
o
r
e
ef
f
icien
t
an
d
in
c
r
ea
s
in
g
ly
ch
a
llen
g
in
g
to
d
etec
t
[
1
7
]
.
ML
t
ec
h
n
iq
u
es
ca
n
p
r
o
v
id
e
in
n
o
v
a
tiv
e
,
ef
f
icien
t,
an
d
tim
ely
m
eth
o
d
s
to
d
etec
t
an
d
p
r
ev
en
t
attac
k
s
[
2
]
,
[
1
7
]
.
T
h
es
e
tech
n
iq
u
es
ca
n
b
e
u
tili
ze
d
in
a
v
ar
iety
o
f
way
s
to
p
r
o
v
id
e
s
ec
u
r
ity
to
I
I
o
T
s
y
s
tem
s
,
in
clu
d
in
g
an
o
m
al
y
d
et
ec
tio
n
,
f
ea
tu
r
e
s
elec
tio
n
,
an
al
y
s
is
o
f
n
etwo
r
k
s
,
o
r
r
is
k
ass
es
s
m
en
t
[
1
8
]
.
ML
m
o
d
els
ca
n
p
r
o
v
i
d
e
cy
b
er
s
ec
u
r
ity
s
y
s
tem
s
with
in
cr
ea
s
ed
ef
f
icien
cy
,
ac
cu
r
ac
y
an
d
au
to
m
atio
n
[
1
8
]
im
p
o
r
ta
n
t f
ac
to
r
s
in
in
d
u
s
tr
y
a
p
p
licatio
n
s
.
Ma
n
y
in
d
u
s
tr
ies
u
tili
ze
I
I
o
T
in
clu
d
i
n
g
s
m
ar
t
cities
[
1
9
]
,
ag
r
ic
u
ltu
r
e,
h
ea
lth
ca
r
e,
p
o
wer
,
tr
an
s
p
o
r
tatio
n
[
1
0
]
,
[
2
0
]
an
d
m
an
u
f
ac
tu
r
in
g
[
2
1
]
.
T
h
e
r
is
k
t
o
th
ese
in
d
u
s
tr
ies
f
r
o
m
cy
b
er
-
attac
k
th
r
o
u
g
h
I
I
o
T
d
ev
ices
an
d
n
etwo
r
k
s
c
o
u
ld
b
e
ca
tast
r
o
p
h
ic.
Du
e
t
o
th
e
n
atu
r
e
o
f
I
I
o
T
,
attac
k
s
m
a
y
af
f
ec
t
e
q
u
ip
m
e
n
t,
p
r
esen
tin
g
a
s
er
io
u
s
r
is
k
to
p
er
s
o
n
n
el
s
af
ety
an
d
s
er
v
ice
p
r
o
v
is
io
n
[
2
2
]
.
Attack
s
m
a
y
r
esu
lt
in
f
in
an
cial
an
d
r
ep
u
tatio
n
al
l
o
s
s
es
ass
o
ciate
d
with
d
is
r
u
p
tio
n
s
t
o
s
er
v
ice,
in
ter
f
er
en
ce
with
p
r
o
d
u
ctio
n
o
r
d
ata
b
r
ea
c
h
es
[
5
]
,
[
1
0
]
,
[
2
0
]
.
So
m
e
attac
k
s
o
f
co
n
ce
r
n
f
o
r
I
I
o
T
in
clu
d
e
m
an
-
in
-
th
e
-
m
id
d
le
[
2
3
]
,
p
h
y
s
ical,
im
p
er
s
o
n
atio
n
,
r
o
u
tin
g
,
m
alicio
u
s
c
o
d
e
in
jec
tio
n
an
d
d
ata
leak
ag
e
[
4
]
as
we
ll
as
d
en
ial
-
of
-
s
er
v
ice,
r
ep
lay
an
d
d
ec
e
p
tio
n
attac
k
s
[
2
4
]
.
Oth
er
attac
k
s
m
o
r
e
s
p
ec
if
ic
t
o
th
e
I
I
o
T
m
ay
in
clu
d
e
tam
p
er
in
g
with
p
r
o
d
u
ct
s
,
s
p
ea
r
p
h
is
h
in
g
o
r
th
e
th
ef
t
o
f
in
tellectu
al
p
r
o
p
e
r
ty
[
2
5
]
.
Netwo
r
k
m
o
n
ito
r
i
n
g
an
d
i
n
tr
u
s
io
n
d
etec
tio
n
a
r
e
p
o
s
s
ib
le
s
o
lu
tio
n
s
to
th
ese
cy
b
er
s
ec
u
r
ity
th
r
ea
ts
to
I
I
o
T
,
an
d
th
is
is
an
ar
ea
wh
e
r
e
ML
ap
p
r
o
ac
h
es h
av
e
b
ee
n
p
r
o
p
o
s
ed
.
T
h
e
f
ield
o
f
ML
is
ev
er
g
r
o
win
g
an
d
I
I
o
T
h
as
b
ec
o
m
e
i
n
cr
ea
s
in
g
ly
p
r
ev
alen
t,
p
r
esen
t
in
g
u
n
i
q
u
e
cy
b
er
s
ec
u
r
ity
ch
allen
g
es.
I
t
is
im
p
o
r
tan
t
to
r
ev
iew
r
ec
en
t
d
ev
elo
p
m
e
n
ts
an
d
co
n
s
o
lid
a
te
th
e
in
f
o
r
m
atio
n
av
ailab
le
in
th
ese
ar
ea
s
in
t
h
e
s
ea
r
ch
f
o
r
ap
p
r
o
p
r
iate
s
o
lu
tio
n
s
.
T
h
is
r
e
v
iew
p
a
p
er
ac
h
iev
es
th
is
g
o
al
b
y
co
n
s
o
lid
atin
g
a
n
d
c
o
m
p
ar
in
g
t
h
e
ML
ap
p
r
o
ac
h
es
p
r
o
p
o
s
ed
i
n
twelv
e
r
ec
en
t
p
ap
er
s
,
p
r
o
v
id
in
g
an
o
v
er
v
iew
o
f
th
e
cu
r
r
e
n
t state
o
f
ML
as a
n
a
p
p
r
o
ac
h
to
I
I
o
T
c
y
b
er
s
ec
u
r
ity
.
T
h
er
e
wer
e
two
m
ain
ar
c
h
itectu
r
es
ar
is
in
g
f
r
o
m
th
e
c
u
r
r
en
t
liter
atu
r
e:
an
in
tr
u
s
io
n
o
r
attac
k
d
etec
tio
n
ar
ch
i
tectu
r
e
an
d
a
f
ed
er
ated
lear
n
in
g
ar
ch
itect
u
r
e.
T
h
ese
a
p
p
r
o
ac
h
es
o
f
f
er
a
way
to
d
etec
t
cy
b
er
s
ec
u
r
ity
attac
k
s
o
r
i
n
tr
u
s
io
n
s
an
d
u
tili
ze
ML
ap
p
r
o
a
ch
es
to
p
r
o
ce
s
s
d
ata
an
d
id
e
n
tify
an
o
m
alies.
Of
th
o
s
e
p
ap
er
s
r
ev
iewe
d
,
ten
u
s
ed
o
n
e
o
f
th
ese
ap
p
r
o
ac
h
es.
T
h
e
m
ain
a
r
ch
itectu
r
e
u
tili
ze
d
f
o
r
p
r
o
p
o
s
ed
s
o
lu
tio
n
s
to
I
I
o
T
cy
b
er
s
ec
u
r
ity
was
th
e
attac
k
d
etec
tio
n
ar
ch
itectu
r
e,
as
d
is
p
lay
ed
in
Fig
u
r
e
1
.
I
n
an
attac
k
d
etec
t
io
n
ap
p
r
o
ac
h
,
d
ata
is
f
ir
s
t
co
llected
,
th
en
p
r
e
-
p
r
o
ce
s
s
ed
ac
co
r
d
in
g
to
th
e
m
o
d
el’
s
n
ee
d
s
a
n
d
th
e
c
o
m
p
o
s
i
t
io
n
o
f
th
e
d
ata
[
2
6
]
.
Data
is
th
en
s
p
lit
in
to
test
in
g
o
r
tr
ain
i
n
g
s
eg
m
e
n
ts
an
d
f
e
d
in
to
v
ar
io
u
s
lay
er
s
o
f
m
a
ch
in
e
lear
n
i
n
g
tech
n
iq
u
es
to
p
er
f
o
r
m
th
e
attac
k
d
etec
tio
n
an
d
clas
s
if
icatio
n
[
2
7
]
.
T
h
e
m
o
d
el’
s
p
e
r
f
o
r
m
an
ce
is
th
en
ev
alu
ated
.
T
h
is
ar
ch
itectu
r
e
is
u
tili
ze
d
b
y
s
ev
en
o
f
th
e
twelv
e
p
ap
e
r
s
ex
am
in
ed
in
th
is
s
tu
d
y
.
T
h
is
attac
k
d
etec
tio
n
a
p
p
r
o
ac
h
ca
n
b
e
ap
p
lied
at
th
e
n
etwo
r
k
lev
el
t
o
ad
d
r
ess
I
I
o
T
n
etwo
r
k
v
u
ln
er
a
b
ilit
y
[
3
]
,
[
5
]
,
[
2
8
]
,
[
2
9
]
,
o
r
at
th
e
d
ev
ice
l
ev
el
to
ad
d
r
ess
th
e
v
u
ln
er
ab
i
lity
o
f
p
h
y
s
ical
s
y
s
tem
s
[
3
0
]
,
[
3
1
]
.
T
h
is
ar
ch
itectu
r
e
ca
n
also
b
e
u
tili
ze
d
f
o
r
I
I
o
T
m
o
n
ito
r
in
g
s
y
s
tem
s
[
3
2
]
.
T
h
ese
ty
p
es
o
f
attac
k
d
etec
tio
n
s
y
s
tem
s
ar
e
a
b
le
to
s
u
cc
ess
f
u
lly
u
s
e
v
ar
io
u
s
ML
tech
n
iq
u
es
to
d
etec
t
attac
k
s
an
d
th
e
r
eb
y
p
r
o
tect
I
I
o
T
d
e
v
ices
an
d
s
y
s
tem
s
.
Ho
wev
er
,
th
e
y
d
o
n
o
t
ad
d
r
ess
p
r
iv
ac
y
c
o
n
ce
r
n
s
as
f
ed
er
ated
lear
n
in
g
a
p
p
r
o
a
ch
es
d
o
,
wh
ich
is
an
im
p
o
r
tan
t
p
ar
t
o
f
I
I
o
T
c
y
b
er
s
ec
u
r
it
y
.
Attack
d
et
ec
tio
n
s
o
lu
tio
n
s
u
tili
zin
g
ML
tech
n
i
q
u
es
ca
n
h
elp
to
id
en
tify
d
en
ial
o
f
s
er
v
ice
(
Do
S)
attac
k
s
,
m
alwa
r
e
an
d
o
th
er
cy
b
er
s
ec
u
r
ity
th
r
ea
ts
th
at
m
ay
ca
u
s
e
an
o
m
alies in
d
ata
o
r
n
et
wo
r
k
tr
af
f
ic
[
1
8
]
.
T
h
e
s
ec
o
n
d
m
ai
n
ty
p
e
o
f
ar
c
h
itectu
r
e
p
r
esen
ted
in
th
e
cu
r
r
en
t
liter
atu
r
e
is
d
is
p
lay
ed
b
y
th
e
th
r
ee
m
o
d
els
u
s
in
g
f
ed
e
r
ated
lea
r
n
i
n
g
(
FL)
[
2
]
,
[
3
3
]
,
[
3
4
]
.
T
h
is
ar
ch
itectu
r
e
is
d
is
p
lay
e
d
in
Fig
u
r
e
2
.
I
n
th
is
ty
p
e
o
f
ap
p
r
o
ac
h
I
I
o
T
clien
ts
tr
ain
t
h
eir
o
wn
lo
ca
l
attac
k
o
r
in
t
r
u
s
io
n
d
etec
tio
n
m
o
d
el.
T
h
e
r
esu
ltin
g
tr
ain
in
g
in
f
o
r
m
atio
n
is
th
en
s
en
t
to
a
ce
n
tr
al
s
er
v
er
,
wh
ich
u
p
d
at
es
th
e
g
lo
b
al
m
o
d
el
with
th
e
lo
ca
l
d
ata
b
e
f
o
r
e
r
etu
r
n
in
g
th
e
u
p
d
ated
g
l
o
b
al
i
n
f
o
r
m
atio
n
to
ea
ch
clien
t
[
3
5
]
.
T
h
e
clien
ts
th
en
u
p
d
ate
th
eir
o
wn
lo
ca
l
m
o
d
els
in
o
r
d
e
r
to
p
e
r
f
o
r
m
attac
k
d
ete
ctio
n
[
1
]
.
FL
is
lar
g
ely
u
s
ed
to
ad
d
r
ess
p
r
iv
ac
y
co
n
ce
r
n
s
ar
o
u
n
d
d
ata
tr
an
s
m
is
s
io
n
[
2
]
,
[
3
6
]
,
as
r
aw
d
ata
is
n
o
t
s
en
t,
r
ath
er
it
is
th
e
tr
ain
e
d
p
a
r
am
eter
s
th
at
ar
e
tr
an
s
m
itted
to
a
ce
n
tr
al
s
er
v
er
[
1
]
,
[
3
0
]
,
[
3
7
]
–
[
3
9
]
.
FL
ca
n
also
p
r
o
v
id
e
s
ca
lab
ilit
y
an
d
r
ea
l
-
tim
e
d
etec
tio
n
o
f
an
o
m
alies
[
3
9
]
.
L
i
et
a
l.
[
3
3
]
an
d
Ma
k
k
ar
et
a
l.
[
3
4
]
tak
e
FL’
s
p
r
iv
ac
y
a
s
tep
f
u
r
th
er
b
y
also
ad
o
p
tin
g
an
e
n
cr
y
p
tio
n
s
y
s
tem
to
en
s
u
r
e
th
e
in
f
o
r
m
atio
n
b
ein
g
tr
an
s
f
er
r
ed
h
as
an
ex
tr
a
lay
er
o
f
s
ec
u
r
ity
.
B
o
th
m
o
d
els
u
tili
ze
a
Pailli
er
en
cr
y
p
tio
n
s
y
s
tem
,
with
L
i
et
a
l.
[
3
3
]
also
a
d
d
in
g
AE
S
en
cr
y
p
tio
n
.
FL
ar
c
h
itectu
r
e
is
ab
le
to
ad
d
r
ess
r
ea
l
-
wo
r
ld
co
n
ce
r
n
s
s
u
c
h
as
d
ata
p
r
iv
ac
y
an
d
s
ec
u
r
ity
[
3
5
]
,
[
3
8
]
.
Ho
wev
er
,
m
o
d
el
c
o
m
p
lex
ity
a
n
d
th
e
p
r
o
ce
s
s
in
g
c
ap
a
b
ilit
ies
o
f
I
I
o
T
d
ev
ices
m
u
s
t
b
e
co
n
s
id
er
ed
as
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
Ma
ch
in
e
lea
r
n
in
g
a
p
p
r
o
a
c
h
e
s
to
cy
b
ers
ec
u
r
ity
in
th
e
in
d
u
s
tr
ia
l
…
(
Mela
n
ie
Heie
r
)
3853
th
ey
n
ee
d
to
b
e
ab
le
to
p
er
f
o
r
m
th
eir
o
wn
m
o
d
el
tr
ain
in
g
an
d
th
ese
d
ev
ices
m
ay
n
o
t
h
av
e
th
e
r
eq
u
ir
e
d
p
r
o
ce
s
s
in
g
p
o
wer
[
4
0
]
.
As
well
as
p
r
o
ce
s
s
in
g
lo
ad
co
n
s
id
er
atio
n
s
,
f
ed
er
ated
lear
n
in
g
t
ec
h
n
iq
u
es
also
f
ac
e
ch
allen
g
es
o
f
d
ev
ices
r
ec
o
n
n
ec
tin
g
af
ter
b
ein
g
o
f
f
lin
e
an
d
p
r
o
tect
ea
c
h
f
ac
et
o
f
th
e
p
r
o
ce
s
s
,
in
clu
d
in
g
th
e
ce
n
tr
alize
d
d
ata
co
llectio
n
p
o
i
n
t a
n
d
in
f
o
r
m
atio
n
tr
an
s
f
er
s
b
etwe
en
d
ev
ice
an
d
ce
n
tr
al
s
er
v
er
[
3
8
]
.
Fig
u
r
e
1
.
Gen
e
r
al
f
lo
w
o
f
atta
ck
d
etec
tio
n
a
p
p
r
o
ac
h
es.
B
ased
o
n
d
iag
r
am
s
f
r
o
m
Fu
et
a
l.
[
2
8
]
,
Sh
ah
in
et
a
l.
[
3
0
]
,
T
r
an
et
a
l.
[
3
1
]
,
a
n
d
C
h
ak
r
a
b
o
r
ty
et
a
l.
[
3
2
]
Fig
u
r
e
2
.
Gen
e
r
al
f
lo
w
o
f
f
ed
e
r
ated
lear
n
in
g
ap
p
r
o
ac
h
es.
B
ased
o
n
d
iag
r
am
s
f
r
o
m
Ao
u
ed
i
e
t a
l.
[
2
]
an
d
L
i
et
a
l.
[
3
3
]
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
.
4
,
Au
g
u
s
t
20
25
:
3
8
5
1
-
3866
3854
T
ab
le
1
s
u
m
m
ar
izes
th
e
m
o
s
t
co
m
m
o
n
ly
p
r
o
p
o
s
ed
tec
h
n
iq
u
es
an
d
th
eir
ca
teg
o
r
ies.
T
ab
le
2
s
u
m
m
ar
izes
th
e
co
m
p
o
n
en
ts
p
r
esen
t
in
t
h
e
r
e
v
iewe
d
liter
a
tu
r
e.
C
o
m
p
o
n
en
ts
ar
e
co
m
p
r
i
s
ed
o
f
t
h
e
s
o
f
twar
e
r
elate
d
to
o
ls
u
s
ed
b
y
r
esear
ch
er
s
,
th
e
d
atasets
u
s
ed
to
ev
alu
ate
th
e
ML
m
o
d
els,
th
e
attac
k
ty
p
es
in
clu
d
ed
in
th
o
s
e
d
ata
s
ets,
th
e
tech
n
iq
u
es
u
s
ed
in
th
e
p
r
o
p
o
s
ed
m
o
d
els,
th
e
m
etr
ics
u
s
ed
to
m
ea
s
u
r
e
m
o
d
el
p
er
f
o
r
m
a
n
ce
,
an
d
th
e
v
ar
iab
le
s
th
at
wer
e
ad
ju
s
ted
to
ex
am
in
e
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
els.
T
h
ese
co
m
p
o
n
e
n
ts
ar
e
o
r
g
an
ized
in
to
f
o
u
r
ca
te
g
o
r
ies
o
f
to
o
ls
,
in
p
u
t,
tech
n
iq
u
es
an
d
o
u
tp
u
t.
T
ab
le
3
b
r
ea
k
s
d
o
wn
th
ese
co
m
p
o
n
en
ts
b
y
p
a
p
er
.
As
ca
n
b
e
s
ee
n
in
T
ab
le
2
tech
n
iq
u
e
s
s
ec
tio
n
as
well
as
T
ab
le
3
,
t
h
e
cu
r
r
e
n
t
liter
atu
r
e
p
r
o
p
o
s
es
m
an
y
d
if
f
er
en
t
ML
tech
n
iq
u
es
u
s
ed
in
v
ar
io
u
s
c
o
m
b
in
atio
n
s
f
o
r
cy
b
er
s
ec
u
r
ity
in
I
I
o
T
.
B
r
o
ad
l
y
,
th
ese
tech
n
iq
u
es
in
clu
d
e
ca
teg
o
r
ies
o
f
ar
tific
ial
n
eu
r
al
n
etwo
r
k
s
,
s
u
p
er
v
is
ed
,
u
n
s
u
p
er
v
is
ed
an
d
s
em
i
-
s
u
p
er
v
is
e
d
lear
n
in
g
,
d
ee
p
lear
n
in
g
,
en
s
em
b
le
lear
n
in
g
,
an
d
en
s
em
b
le
m
eth
o
d
s
.
Du
e
to
t
h
e
lim
ited
s
co
p
e
o
f
th
is
r
ev
iew,
th
e
f
o
cu
s
will
b
e
o
n
th
e
co
m
m
o
n
tec
h
n
iq
u
es
as
p
r
esen
ted
i
n
T
ab
le
1
.
On
e
p
a
p
er
d
id
n
o
t
s
p
ec
if
y
th
eir
tech
n
iq
u
e,
m
er
ely
s
tatin
g
m
ac
h
i
n
e
lear
n
in
g
(
ML
)
an
d
d
ee
p
lear
n
in
g
(
DL
)
alg
o
r
ith
m
s
[
3
3
]
,
m
a
k
in
g
its
co
m
p
ar
is
o
n
i
n
co
m
p
atib
le
with
o
th
er
s
p
r
esen
ted
h
er
e.
T
ab
le
1
.
Mo
s
t c
o
m
m
o
n
ML
tech
n
iq
u
es
C
a
t
e
g
o
r
y
A
b
b
r
.
Te
c
h
n
i
q
u
e
A
r
t
i
f
i
c
i
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
(
A
N
N
)
C
N
N
C
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
F
C
N
N
F
u
l
l
y
c
o
n
v
o
l
u
t
i
o
n
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
s
M
LP
M
u
l
t
i
l
a
y
e
r
p
e
r
c
e
p
t
r
o
n
LSTM
Lo
n
g
sh
o
r
t
-
t
e
r
m m
e
m
o
r
y
S
u
p
e
r
v
i
se
d
l
e
a
r
n
i
n
g
DT
D
e
c
i
s
i
o
n
t
r
e
e
RF
R
a
n
d
o
m
f
o
r
e
s
t
En
se
mb
l
e
m
e
t
h
o
d
s
X
G
B
o
o
st
Ex
t
r
e
me
g
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
T
ab
le
2
.
C
o
m
p
o
n
en
ts
F
a
c
t
o
r
s
A
t
t
r
i
b
u
t
e
s
I
n
st
a
n
c
e
s
To
o
l
s
S
o
f
t
w
a
r
e
O
P
N
e
t
N
e
t
w
o
r
k
si
mu
l
a
t
i
o
n
,
N
e
t
f
l
o
w
,
R
e
d
i
s,
A
n
a
c
o
n
d
a
N
a
v
i
g
a
t
o
r
,
Te
n
s
o
r
f
l
o
w
,
G
o
o
g
l
e
C
o
l
a
b
,
La
b
V
I
EW,
C
O
N
TA
C
T
El
e
me
n
t
P
l
a
t
f
o
r
m
F
r
a
mew
o
r
k
s,
l
i
b
r
a
r
i
e
s,
l
a
n
g
u
a
g
e
s
P
y
t
o
r
c
h
,
F
l
a
s
k
,
K
e
r
a
s,
S
c
i
k
i
t
-
L
e
a
r
n
,
P
y
t
h
o
n
I
n
p
u
t
D
a
t
a
s
e
t
G
a
s
p
i
p
e
l
i
n
e
S
C
A
D
A
sy
s
t
e
m
,
w
a
t
e
r
s
t
o
r
a
g
e
t
a
n
k
c
o
n
t
r
o
l
s
y
s
t
e
m,
S
e
c
u
r
e
W
a
t
e
r
Tr
e
a
t
me
n
t
,
C
I
C
-
I
D
S
-
2
0
1
8
,
D
S
2
O
S
,
U
N
S
W
-
N
B
1
5
,
S
C
A
D
A
p
o
w
e
r
sy
st
e
m,
X
I
I
o
TI
D
,
B
o
T
-
I
o
T,
T
o
N
-
I
o
T,
G
l
i
t
c
h
e
s,
B
o
t
a
t
t
a
c
k
sa
mp
l
e
s
,
i
n
d
u
c
t
i
o
n
mo
t
o
r
b
e
a
r
i
n
g
c
o
n
d
i
t
i
o
n
s
A
t
t
a
c
k
t
y
p
e
N
M
R
I
,
C
M
R
I
,
M
S
C
I
,
M
P
C
I
,
M
F
C
I
,
D
o
S
,
D
D
o
S
,
R
e
c
o
n
.
,
H
e
a
r
t
b
l
e
e
d
,
w
e
b
a
t
t
a
c
k
s,
b
o
t
n
e
t
,
I
N
F
I
,
U
t
R
,
M
C
,
M
O
,
W
S
,
sp
y
i
n
g
,
sca
n
,
D
TP,
f
u
z
z
e
r
s,
b
a
c
k
d
o
o
r
,
a
n
a
l
y
s
i
s,
e
x
p
l
o
i
t
,
g
e
n
e
r
i
c
,
s
h
e
l
l
c
o
d
e
,
w
o
r
m
,
w
e
a
p
o
n
i
z
a
t
i
o
n
,
L
M
,
C
&C
,
r
a
n
so
m D
o
S
,
e
x
f
i
l
t
r
a
t
i
o
n
,
c
r
y
p
t
o
-
r
a
n
s
o
mw
a
r
e
/
r
a
n
s
o
mw
a
r
e
,
k
e
y
l
o
g
g
i
n
g
,
i
n
j
e
c
t
i
o
n
,
M
I
TM
,
p
a
ssw
o
r
d
,
X
S
S
,
S
S
-
S
P
A
S
S
M
P
A
,
M
S
-
S
P
A
,
M
S
-
M
P
A
,
B
o
t
n
e
t
,
I
R
F
,
O
R
F
,
c
y
b
e
r
-
a
t
t
a
c
k
D
a
t
a
s
e
t
t
y
p
e
S
o
u
r
c
e
d
,
s
e
l
f
-
c
r
e
a
t
e
d
Te
c
h
n
i
q
u
e
s
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
a
p
p
r
o
a
c
h
e
s
F
L,
S
S
L,
EL
,
D
L,
M
L,
A
E
,
F
C
N
,
M
L
P
,
EL
M
,
C
N
N
,
G
R
U
,
LSTM
,
F
C
N
N
,
A
LSTM
,
R
F
,
X
G
B
o
o
s
t
,
Li
g
h
t
G
B
M
,
A
d
a
B
o
o
st
,
L
R
,
S
V
M
,
k
-
N
N
,
D
T,
C
A
,
H
C
A
,
P
R
U
,
R
a
N
N
O
t
h
e
r
t
e
c
h
n
i
q
u
e
s
F
S
A
,
F
P
C
A
,
P
a
i
l
l
i
e
r
,
A
ES
O
u
t
p
u
t
Ev
a
l
u
a
t
i
o
n
m
e
t
r
i
c
s
A
c
c
u
r
a
c
y
,
p
r
e
c
i
s
i
o
n
,
r
e
c
a
l
l
,
F
1
s
c
o
r
e
,
l
o
g
l
o
ss
,
c
o
mm
u
n
i
c
a
t
i
o
n
o
v
e
r
h
e
a
d
,
A
U
C
/
R
O
C
,
s
a
f
e
t
y
f
a
c
t
o
r
,
M
C
C
,
TP
R
,
TN
R
,
F
P
R
,
F
N
R
,
TP,
F
P
,
TN
,
F
N
,
d
e
t
e
c
t
i
o
n
t
i
me
V
a
r
i
a
b
l
e
s
N
u
mb
e
r
o
f
c
l
i
e
n
t
s,
l
o
c
a
l
c
l
i
e
n
t
e
p
o
c
h
s
,
c
o
mm
u
n
i
c
a
t
i
o
n
r
o
u
n
d
s,
a
m
o
u
n
t
o
f
u
n
l
a
b
e
l
e
d
d
a
t
a
,
S
e
g
me
n
t
si
z
e
,
T
i
me
a
l
l
o
c
a
t
e
d
f
o
r
d
e
c
i
s
i
o
n
ma
k
i
n
g
,
l
i
n
e
a
r
/
n
o
n
-
l
i
n
e
a
r
se
n
so
r
s
,
N
u
mb
e
r
o
f
f
e
a
t
u
r
e
s,
Le
a
r
n
i
n
g
r
a
t
e
,
T
i
me
s
l
o
t
s,
B
i
n
a
r
y
c
l
a
ss
i
f
i
c
a
t
i
o
n
/
m
u
l
t
i
c
l
a
ssi
f
i
c
a
t
i
o
n
,
d
a
t
a
set
,
l
e
a
r
n
e
r
s,
mo
d
e
l
,
t
y
p
e
o
f
a
t
t
a
c
k
,
t
r
a
i
n
i
n
g
/
t
e
s
t
i
n
g
,
d
e
v
i
c
e
S
u
m
m
a
r
y
o
f
i
n
s
t
a
n
c
e
s
o
f
a
t
t
r
i
b
u
t
e
s
f
r
o
m
p
a
p
e
r
s
i
n
c
l
u
d
e
d
i
n
t
h
e
l
i
t
e
r
a
t
u
r
e
r
e
v
i
e
w
.
A
b
b
r
e
v
i
a
t
i
o
n
s
u
s
e
d
i
n
t
a
b
l
e
a
r
e
l
i
s
t
e
d
i
n
t
h
e
A
p
p
e
n
d
i
x
.
Ar
tific
ial
n
eu
r
al
n
etwo
r
k
s
(
ANN)
ar
e
p
ar
t
o
f
DL
,
a
s
u
b
s
ec
tio
n
o
f
ML
.
T
h
e
y
ca
n
b
e
u
tili
ze
d
in
m
o
d
els
f
o
r
c
y
b
er
s
ec
u
r
ity
to
d
etec
t
m
alwa
r
e
o
r
a
n
aly
ze
n
et
wo
r
k
b
eh
av
i
o
r
[
1
6
]
.
ANN
tec
h
n
iq
u
es
ca
n
also
b
e
u
tili
ze
d
f
o
r
tim
e
s
er
ies
p
r
ed
ictio
n
o
r
s
p
ee
ch
r
ec
o
g
n
itio
n
[
4
1
]
.
I
n
th
e
r
e
v
iewe
d
p
ap
er
s
,
ANNs
wer
e
lar
g
ely
u
s
ed
f
o
r
d
ata
class
if
icatio
n
[
4
2
]
an
d
to
ex
tr
ac
t
f
ea
tu
r
es
[
3
0
]
.
T
h
is
p
r
o
ce
s
s
o
f
class
if
i
ca
tio
n
an
d
f
ea
tu
r
e
ex
tr
ac
tio
n
allo
ws
n
ew
d
ata
to
b
e
ea
s
ily
clas
s
if
ied
o
r
f
ilter
ed
b
ased
o
n
p
r
ev
io
u
s
ly
p
r
o
ce
s
s
ed
in
f
o
r
m
atio
n
[
4
2
]
.
ANNs
g
en
er
ally
co
n
s
is
t
o
f
a
n
u
m
b
er
o
f
co
n
n
ec
ted
n
o
d
es
th
at
ea
ch
p
er
f
o
r
m
d
ata
p
r
o
ce
s
s
in
g
[
4
3
]
.
ANN
n
o
d
es
co
n
s
is
t o
f
th
r
ee
lay
er
s
: o
n
e
f
o
r
in
p
u
t,
o
n
e
f
o
r
o
u
tp
u
t a
n
d
o
n
e
h
id
d
en
la
y
er
f
o
r
p
r
o
ce
s
s
in
g
[
3
0
]
,
[
4
3
]
,
[
4
4
]
.
T
h
e
m
o
s
t
co
m
m
o
n
ly
p
r
o
p
o
s
ed
ANNs
in
th
e
liter
atu
r
e
in
clu
d
e
co
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
s
(
C
NN)
[
2
9
]
,
[
3
3
]
,
[
3
4
]
,
f
u
lly
c
o
n
v
o
lu
tio
n
al
n
eu
r
al
n
etwo
r
k
s
(
FC
NN)
[
2
9
]
,
[
3
0
]
,
[
3
4
]
,
m
u
ltil
ay
e
r
p
er
ce
p
tr
o
n
(
ML
P)
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
Ma
ch
in
e
lea
r
n
in
g
a
p
p
r
o
a
c
h
e
s
to
cy
b
ers
ec
u
r
ity
in
th
e
in
d
u
s
tr
ia
l
…
(
Mela
n
ie
Heie
r
)
3855
[
5
]
,
[
3
3
]
an
d
l
o
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
(
L
STM
)
[
3
]
,
[
2
9
]
.
As
s
ee
n
in
T
ab
le
3
,
t
h
e
d
atasets
u
s
ed
to
test
th
ese
m
o
d
els
in
clu
d
ed
:
UNSW
-
NB
1
5
,
B
o
T
-
I
o
T
,
T
o
N
-
I
o
T
,
a
g
as
p
ip
elin
e
s
y
s
tem
,
a
s
u
p
er
v
is
o
r
y
co
n
tr
o
l
an
d
d
ata
ac
q
u
is
itio
n
(
SC
ADA)
s
y
s
tem
,
an
d
XI
I
o
T
I
D.
Als
o
s
h
o
wn
in
T
ab
le
3
,
th
ese
d
atasets
ad
d
r
ess
ed
a
r
an
g
e
o
f
attac
k
ty
p
es,
in
clu
d
in
g
b
u
t n
o
t
lim
ited
to
: D
o
S,
b
ac
k
d
o
o
r
,
r
a
n
s
o
m
war
e,
m
an
-
in
-
th
e
-
m
id
d
le
(
MI
T
M)
,
cr
o
s
s
s
i
te
s
cr
ip
tin
g
(
XSS),
r
ec
o
n
n
aiss
an
ce
(
r
ec
o
n
.
)
an
d
wo
r
m
s
.
T
ab
le
3
.
C
lass
if
icatio
n
R
e
f
[
#
]
To
o
l
s
I
n
p
u
t
Te
c
h
n
i
q
u
e
s
O
u
t
p
u
t
S
W
,
F
W
,
Li
b
s
,
L
a
n
g
s
.
D
a
t
a
s
e
t
A
t
t
a
c
k
t
y
p
e
D
a
t
a
s
e
t
t
y
p
e
M
L
a
n
d
o
t
h
e
r
a
p
p
r
o
a
c
h
e
s
Ev
a
l
u
a
t
i
o
n
m
e
t
r
i
c
s
V
a
r
i
a
b
l
e
s
[
2
]
P
y
t
o
r
c
h
,
S
c
i
k
i
t
-
L
e
a
r
n
,
P
y
t
h
o
n
G
a
s
p
i
p
e
l
i
n
e
S
C
A
D
A
sy
s
t
e
m
d
a
t
a
se
t
,
w
a
t
e
r
s
t
o
r
a
g
e
t
a
n
k
c
o
n
t
r
o
l
sy
s
t
e
m
N
M
R
I
,
C
M
R
I
,
M
S
C
I
,
M
P
C
I
,
M
F
C
I
,
D
o
S
,
R
e
c
o
n
.
S
o
u
r
c
e
d
A
E,
F
C
N
,
F
L,
SSL
A
c
c
u
r
a
c
y
,
p
r
e
c
i
si
o
n
,
r
e
c
a
l
l
,
F
1
sc
o
r
e
,
c
o
mm
u
n
i
c
a
t
i
o
n
o
v
e
r
h
e
a
d
N
u
m.
c
l
i
e
n
t
s
,
l
o
c
a
l
c
l
i
e
n
t
e
p
o
c
h
s,
C
R
,
a
mo
u
n
t
o
f
u
n
l
a
b
e
l
e
d
d
a
t
a
[
3
]
P
y
t
o
r
c
h
,
P
y
t
h
o
n
S
C
A
D
A
p
o
w
e
r
sy
st
e
m
d
a
t
a
s
e
t
s
(
1
5
d
a
t
a
s
e
t
s)
U
n
sp
e
c
i
f
i
e
d
(
t
h
o
u
s
a
n
d
s
o
f
d
i
s
t
i
n
c
t
a
t
t
a
c
k
s)
S
o
u
r
c
e
d
P
R
U
,
D
T,
LSTM
,
EL
A
c
c
u
r
a
c
y
,
F
P
R
,
TP
,
F
P
,
TN
,
F
N
B
i
n
a
r
y
/
mu
l
t
i
c
l
a
ssi
f
i
c
a
t
i
o
n
,
d
a
t
a
se
t
,
l
e
a
r
n
e
r
s
[
5
]
A
n
a
c
o
n
d
a
N
a
v
i
g
a
t
o
r
,
Te
n
s
o
r
f
l
o
w
,
K
e
r
a
s
D
S
2
O
S
,
U
N
S
W
-
N
B
1
5
D
o
S
,
M
C
,
M
O
,
W
S
,
sp
y
i
n
g
,
sca
n
,
D
TP,
f
u
z
z
e
r
s
,
b
a
c
k
d
o
o
r
,
a
n
a
l
y
s
i
s,
e
x
p
l
o
i
t
,
g
e
n
e
r
i
c
,
sh
e
l
l
c
o
d
e
,
w
o
r
m
S
o
u
r
c
e
d
M
LP,
R
a
N
N
A
c
c
u
r
a
c
y
,
p
r
e
c
i
si
o
n
,
r
e
c
a
l
l
,
F
1
sc
o
r
e
,
l
o
g
l
o
ss
,
A
U
C
-
R
O
C
Le
a
r
n
i
n
g
r
a
t
e
[
4
5
]
G
o
o
g
l
e
C
o
l
a
b
G
l
i
t
c
h
e
s
G
l
i
t
c
h
e
s (
8
8
9
0
o
v
e
r
1
2
h
o
u
r
s)
S
e
l
f
-
C
r
e
a
t
e
d
H
C
A
,
EL
M
,
SSL
A
c
c
u
r
a
c
y
,
p
r
e
c
i
si
o
n
,
r
e
c
a
l
l
,
F
1
sc
o
r
e
Ti
me
sl
o
t
s
[
4
6
]
U
n
sp
e
c
i
f
i
e
d
B
o
t
a
t
t
a
c
k
samp
l
e
s
B
o
t
n
e
t
S
e
l
f
-
C
r
e
a
t
e
d
D
L,
M
L
A
c
c
u
r
a
c
y
,
p
r
e
c
i
si
o
n
,
r
e
c
a
l
l
,
F
1
sc
o
r
e
,
M
C
C
,
FPR
[
2
8
]
O
p
n
e
t
,
N
e
t
f
l
o
w
,
R
e
d
i
s
C
I
C
-
I
D
S
-
2
0
1
8
D
o
S
,
R
e
c
o
n
.
,
H
e
a
r
t
b
l
e
e
d
,
w
e
b
a
t
t
a
c
k
s,
b
o
t
n
e
t
,
i
n
s
i
d
e
,
U
t
R
S
o
u
r
c
e
d
CA
S
a
f
e
t
y
f
a
c
t
o
r
,
TP,
F
P
,
d
e
t
e
c
t
i
o
n
t
i
me
N
u
mb
e
r
o
f
f
e
a
t
u
r
e
s
[
2
9
]
S
c
i
k
i
t
-
L
e
a
r
n
U
N
S
W
-
N
B
1
5
,
B
o
T
-
I
o
T,
T
o
N
-
I
o
T
F
u
z
z
e
r
s,
b
a
c
k
d
o
o
r
,
a
n
a
l
y
si
s
,
e
x
p
l
o
i
t
,
g
e
n
e
r
i
c
,
sh
e
l
l
c
o
d
e
,
w
o
r
m,
D
o
S
,
D
D
o
S
,
R
e
c
o
n
.
,
sca
n
,
e
x
f
i
l
t
r
a
t
i
o
n
,
r
a
n
s
o
mw
a
r
e
,
k
e
y
l
o
g
g
i
n
g
,
i
n
j
e
c
t
i
o
n
,
M
I
TM
,
p
a
ssw
o
r
d
,
X
S
S
S
o
u
r
c
e
d
C
N
N
,
LST
M
,
F
C
N
N
A
c
c
u
r
a
c
y
,
p
r
e
c
i
si
o
n
,
r
e
c
a
l
l
,
l
o
g
l
o
ss
D
a
t
a
s
e
t
,
m
o
d
e
l
[
3
0
]
S
c
i
k
i
t
-
L
e
a
r
n
To
N
-
I
o
T
D
o
S
,
D
D
o
S
,
R
e
c
o
n
.
,
sc
a
n
,
b
a
c
k
d
o
o
r
,
r
a
n
s
o
mw
a
r
e
,
i
n
j
e
c
t
i
o
n
,
M
I
TM
,
p
a
ssw
o
r
d
,
X
S
S
S
o
u
r
c
e
d
X
G
B
o
o
st
,
A
d
a
B
o
o
st
,
F
C
N
N
,
A
LSTM
A
c
c
u
r
a
c
y
,
p
r
e
c
i
si
o
n
,
r
e
c
a
l
l
,
F
1
sc
o
r
e
M
o
d
e
l
,
d
e
v
i
c
e
,
a
t
t
a
c
k
t
y
p
e
[
3
1
]
C
O
N
TA
C
T
El
e
m
e
n
t
,
La
b
V
I
EW
i
n
d
u
c
t
i
o
n
mo
t
o
r
b
e
a
r
i
n
g
c
o
n
d
i
t
i
o
n
s
I
R
F
,
O
R
F
,
c
y
b
e
r
a
t
t
a
c
k
S
e
l
f
-
c
r
e
a
t
e
d
D
T,
R
F
,
X
G
B
o
o
st
,
A
c
c
u
r
a
c
y
,
A
U
R
O
C
,
TPR,
F
P
R
,
TP
,
F
P
,
F
N
M
o
d
e
l
,
mo
t
o
r
st
a
t
u
s
(
a
t
t
a
c
k
t
y
p
e
)
[
3
2
]
U
n
sp
e
c
i
f
i
e
d
S
W
a
T
SS
-
S
P
A
,
S
S
-
M
P
A
,
MS
-
S
P
A
,
M
S
-
M
P
A
S
o
u
r
c
e
d
LR
,
S
V
M
,
k
-
N
N
,
R
F
,
F
S
A
,
F
P
C
A
A
c
c
u
r
a
c
y
,
p
r
e
c
i
si
o
n
,
r
e
c
a
l
l
,
F
1
sc
o
r
e
,
TP,
F
P
,
TN
,
F
N
S
e
g
m
e
n
t
s
i
z
e
,
Ti
me
f
o
r
d
e
c
i
s
i
o
n
mak
i
n
g
,
l
i
n
e
a
r
/
n
o
n
-
l
i
n
e
a
r
s
e
n
so
r
s,
M
o
d
e
l
[
3
3
]
F
l
a
s
k
,
K
e
r
a
s,
P
y
t
h
o
n
G
a
s
p
i
p
e
l
i
n
e
S
C
A
D
A
sy
s
t
e
m
N
M
R
I
,
C
M
R
I
,
M
S
C
I
,
M
P
C
I
,
M
F
C
I
,
D
o
S
,
R
e
c
o
n
.
S
o
u
r
c
e
d
M
LP,
C
N
N
,
G
R
U
,
F
L,
P
a
i
l
l
i
e
r
,
A
ES
A
c
c
u
r
a
c
y
,
p
r
e
c
i
si
o
n
,
r
e
c
a
l
l
,
F
1
sc
o
r
e
N
u
m
.
c
l
i
e
n
t
s
,
C
R
,
l
o
c
a
l
/
i
d
e
a
l
/
p
r
o
p
o
s
e
d
m
o
d
e
l
,
t
y
p
e
o
f
a
t
t
a
c
k
[
3
4
]
G
o
o
g
l
e
C
o
l
a
b
,
P
y
t
o
r
c
h
X
I
I
o
T
I
D
R
e
c
o
n
.
,
e
x
p
l
o
i
t
,
w
e
a
p
o
n
i
z
a
t
i
o
n
,
LM
,
C
&
C
,
r
a
n
s
o
m
D
o
S
,
e
x
f
i
l
t
r
a
t
i
o
n
,
r
a
n
s
o
mw
a
r
e
S
o
u
r
c
e
d
R
F
,
X
G
B
o
o
st
,
Li
g
h
t
G
B
M
,
C
N
N
,
LST
M
,
F
L,
P
a
i
l
l
i
e
r
P
r
e
c
i
s
i
o
n
,
r
e
c
a
l
l
,
F
1
sc
o
r
e
,
TPR,
TN
R
,
F
P
R
,
F
N
R
N
u
m.
c
l
i
e
n
t
s,
mo
d
e
l
,
t
r
a
i
n
i
n
g
/
t
e
s
t
i
n
g
T
h
e
c
o
n
t
e
n
t
s
o
f
t
h
e
c
o
m
p
o
n
e
n
t
t
a
b
l
e
b
ro
k
e
d
o
w
n
b
y
p
a
p
e
r.
Ab
b
re
v
i
a
t
i
o
n
s
u
s
e
d
i
n
t
h
e
t
a
b
l
e
a
r
e
l
i
s
t
e
d
i
n
t
h
e
Ap
p
e
n
d
i
x
.
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
.
4
,
Au
g
u
s
t
20
25
:
3
8
5
1
-
3866
3856
C
NNs
ar
e
o
f
ten
u
s
ed
f
o
r
v
is
u
al
r
ec
o
g
n
itio
n
ac
tiv
ities
[
4
1
]
,
[
4
7
]
,
an
d
a
r
e
ab
le
to
ex
tr
a
ct
f
ea
tu
r
es
au
to
m
atica
lly
[
2
9
]
.
C
NNs
co
n
s
is
t
o
f
a
n
u
m
b
er
o
f
co
n
v
o
l
u
tio
n
al
lay
er
s
u
s
ed
to
ex
tr
ac
t
f
ea
t
u
r
es
an
d
a
n
u
m
b
er
o
f
f
u
lly
c
o
n
n
ec
te
d
lay
er
s
u
s
e
d
to
class
if
y
th
ese
f
ea
tu
r
es,
th
er
eb
y
p
r
o
v
id
in
g
t
h
e
co
m
b
i
n
ed
o
u
tp
u
t
[
2
8
]
,
[
4
1
]
,
[
4
4
]
,
[
4
7
]
.
As with
o
th
er
ANN
s
,
C
NN
alg
o
r
ith
m
s
ca
n
b
e
u
s
e
d
in
co
m
b
in
atio
n
with
o
th
er
M
L
tech
n
iq
u
es.
An
FC
NN
i
s
an
C
NN
co
m
p
r
is
ed
o
n
ly
o
f
co
n
v
o
lu
tio
n
al
lay
e
r
s
[
3
0
]
,
[
4
8
]
.
FC
NN
in
p
ar
ticu
lar
is
ab
le
to
p
er
f
o
r
m
well
in
ter
m
s
o
f
ti
m
e
an
d
r
eso
u
r
ce
s
wh
en
th
er
e
ar
e
m
an
y
v
ar
iatio
n
s
in
th
e
d
ata
[
2
9
]
,
[
3
0
]
,
[
4
9
]
,
[
5
0
]
.
T
h
is
is
b
ec
au
s
e
th
e
n
e
u
r
o
n
s
in
ea
ch
lay
er
ar
e
n
o
t f
u
lly
c
o
n
n
ec
ted
[
2
9
]
.
T
h
e
ML
P is
also
k
n
o
wn
as
a
f
ee
d
-
f
o
r
war
d
f
u
lly
-
co
n
n
ec
ted
m
u
lti
-
lay
er
n
eu
r
al
n
etwo
r
k
[
5
1
]
.
An
ML
P
cr
ea
tes
co
r
r
elatio
n
s
b
etwe
en
t
h
e
in
p
u
t
a
n
d
o
u
tp
u
t
d
ata
b
y
a
d
ju
s
tin
g
th
e
n
eu
r
o
n
s
in
its
lay
er
s
[
3
0
]
.
C
o
n
tin
u
in
g
in
v
esti
g
atio
n
in
to
th
e
u
s
e
o
f
ML
P
d
is
co
v
er
ed
th
at
p
er
f
o
r
m
an
ce
was
ab
le
to
b
e
im
p
r
o
v
ed
b
y
s
eq
u
e
n
tially
p
r
e
-
tr
ain
in
g
la
y
er
s
[
5
1
]
.
L
STM
is
a
ty
p
e
o
f
r
ec
u
r
r
e
n
t
n
eu
r
al
n
etwo
r
k
th
at
is
ab
le
t
o
r
ec
all
p
r
io
r
in
f
o
r
m
atio
n
,
le
ar
n
f
ea
tu
r
e
d
ep
en
d
e
n
cies
[
5
2
]
a
n
d
lear
n
o
r
d
er
d
ep
e
n
d
en
c
y
in
s
eq
u
e
n
c
e
p
r
ed
ictio
n
[
2
9
]
,
[
3
0
]
.
L
ST
M
u
tili
ze
s
g
ates
f
o
r
in
p
u
t,
o
u
tp
u
t
an
d
f
o
r
g
ettin
g
to
p
r
o
ce
s
s
m
em
o
r
y
d
ata
[
2
9
]
,
[
4
1
]
,
[
5
2
]
.
L
STM
is
ab
le
to
b
e
u
tili
ze
d
with
o
th
er
ML
tech
n
iq
u
es
to
ass
is
t
with
ac
cu
r
ate
attac
k
p
r
e
d
ictio
n
.
L
STM
tech
n
iq
u
es
h
av
e
b
ee
n
a
p
p
lied
in
I
I
o
T
s
y
s
tem
s
in
in
d
u
s
tr
ies s
u
ch
as f
in
an
ce
,
h
ea
lth
ca
r
e,
an
d
t
r
an
s
p
o
r
tatio
n
[
4
1
]
.
T
h
ese
ANN
tech
n
iq
u
es
h
av
e
b
ee
n
co
m
b
in
e
d
with
v
ar
io
u
s
o
th
er
ML
ap
p
r
o
ac
h
es
to
f
o
r
m
u
late
m
o
d
els.
T
h
ese
m
o
d
els
h
av
e
b
ee
n
co
m
p
ar
e
d
in
d
if
f
er
en
t
wa
y
s
in
th
e
r
ev
iewe
d
liter
atu
r
e.
Sh
ah
in
et
a
l.
[
2
9
]
co
m
p
ar
ed
two
m
o
d
els:
o
n
e
co
m
b
in
in
g
L
STM
with
C
NN
an
d
th
e
o
th
er
c
o
m
b
in
in
g
L
STM
with
FC
N
N.
Ma
k
k
ar
et
a
l.
[
3
4
]
co
m
p
ar
ed
f
o
u
r
m
o
d
els
with
in
a
f
ed
er
a
ted
lear
n
in
g
ar
c
h
itectu
r
e:
C
NN,
L
STM
an
d
two
en
s
em
b
le
m
eth
o
d
m
o
d
els.
E
ac
h
o
f
th
ese
m
o
d
els
u
s
ed
r
a
n
d
o
m
f
o
r
est
(
R
F)
f
o
r
f
ea
tu
r
e
o
r
g
an
izatio
n
,
an
d
en
s
em
b
le
m
eth
o
d
s
f
o
r
tr
ain
in
g
.
T
h
ese
k
in
d
s
o
f
m
o
d
el
c
o
m
p
ar
is
o
n
s
ar
e
u
s
ef
u
l
in
th
e
an
aly
s
is
o
f
s
p
ec
if
ic
tech
n
iq
u
e
p
er
f
o
r
m
an
ce
.
Oth
er
s
in
th
e
r
ev
iewe
d
liter
atu
r
e
f
o
r
m
ed
th
eir
m
o
d
els
with
a
co
m
b
in
atio
n
o
f
ANN
an
d
n
o
n
-
ANN
ML
ap
p
r
o
ac
h
es.
Hu
m
a
et
a
l.
[
5
]
an
d
Kh
an
et
a
l.
[
3
]
b
o
th
co
m
b
in
e
d
ANN
tech
n
i
q
u
es
with
d
ee
p
lear
n
in
g
,
th
o
u
g
h
in
d
if
f
er
e
n
t
way
s
.
H
u
m
a
et
a
l.
[
5
]
u
tili
ze
d
a
d
e
ep
r
an
d
o
m
n
eu
r
al
n
etwo
r
k
(
R
aNN
)
with
ML
P.
K
h
an
et
a
l.
[
3
]
p
r
o
p
o
s
ed
a
p
y
r
am
id
al
r
ec
u
r
r
en
t
u
n
it
(
PR
U)
m
o
d
el
th
at
in
co
r
p
o
r
ate
d
L
ST
M.
I
n
th
is
way
,
ANN
tech
n
iq
u
es h
av
e
th
e
f
lex
i
b
ilit
y
to
b
e
ap
p
lied
in
m
a
n
y
d
if
f
er
e
n
t m
o
d
el
ty
p
es.
L
i
et
a
l.
[
3
3
]
u
tili
ze
d
ANN
ap
p
r
o
ac
h
es
with
in
a
f
ed
er
ate
d
lear
n
in
g
a
r
ch
itectu
r
e
alo
n
g
s
id
e
o
th
er
tech
n
iq
u
es.
T
h
ey
p
r
o
p
o
s
ed
a
m
o
d
el
u
tili
zin
g
b
o
th
ML
P
an
d
C
NN
alo
n
g
with
a
n
o
th
e
r
A
NN
tech
n
iq
u
e
-
g
ated
r
ec
u
r
r
en
t
u
n
it
(
GR
U)
.
GR
U
m
eth
o
d
s
o
f
f
e
r
an
ef
f
icien
t
o
p
tio
n
th
at
r
eq
u
ir
es
less
co
m
p
u
tatio
n
al
r
eso
u
r
ce
s
[
5
3
]
.
ANN
ap
p
ea
r
t
o
b
e
a
p
o
p
u
lar
ML
m
eth
o
d
f
o
r
I
I
o
T
cy
b
e
r
s
ec
u
r
ity
,
a
n
d
t
h
ese
tech
n
iq
u
es
h
a
v
e
th
e
f
lex
ib
ilit
y
to
b
e
ap
p
lied
in
d
if
f
er
e
n
t
way
s
an
d
with
d
if
f
er
e
n
t
ML
an
d
n
o
n
-
ML
tech
n
i
q
u
es.
Neu
r
al
n
e
two
r
k
tech
n
iq
u
es
d
o
h
av
e
d
r
awb
ac
k
s
h
o
we
v
er
wh
en
it
co
m
es
to
ap
p
lic
atio
n
in
I
I
o
T
.
T
h
ese
tech
n
iq
u
es
ca
n
h
av
e
a
h
ig
h
co
m
p
u
tatio
n
al
c
o
s
t
an
d
b
e
s
u
s
ce
p
tib
le
to
o
v
e
r
f
itti
n
g
[
1
4
]
.
N
eu
r
al
n
etwo
r
k
m
o
d
els
also
tak
e
tim
e
to
co
m
p
lete
th
eir
tr
ain
in
g
p
h
ase,
an
d
ca
n
r
eq
u
ir
e
lar
g
e
am
o
u
n
ts
o
f
d
ata
[
1
4
]
.
T
h
e
two
m
o
s
t
co
m
m
o
n
s
u
p
er
v
is
ed
lear
n
in
g
tech
n
iq
u
es
u
tili
ze
d
in
th
e
r
ev
iewe
d
p
ap
er
s
wer
e
d
ec
is
io
n
tr
ee
(
DT
)
[
3
]
,
[
3
1
]
an
d
R
F
[
3
2
]
,
[
3
4
]
.
As
ca
n
b
e
s
ee
n
i
n
T
a
b
le
3
,
d
atasets
u
s
ed
to
test
th
ese
m
o
d
els
in
clu
d
ed
m
u
lt
ip
le
SC
ADA
p
o
wer
s
y
s
tem
d
atasets
,
a
s
elf
-
cr
ea
ted
eq
u
ip
m
en
t
-
b
ased
d
ataset,
SW
aT
an
d
XI
I
o
T
I
D
d
atasets
.
Als
o
s
h
o
wn
in
T
a
b
le
3
,
th
ese
d
atasets
co
v
er
ed
a
r
an
g
e
o
f
attac
k
s
in
clu
d
i
n
g
b
u
t
n
o
t
lim
ited
to
cy
b
er
attac
k
,
Do
S,
r
an
s
o
m
war
e
,
r
ec
o
n
n
aiss
an
ce
,
an
d
s
in
g
le
a
n
d
m
u
lti
-
p
o
in
t a
ttack
s
.
T
h
e
DT
tech
n
iq
u
e
b
u
ild
s
its
tr
ain
in
g
m
o
d
els
b
y
lear
n
in
g
r
u
les
f
o
r
d
ec
is
io
n
m
ak
in
g
[
5
4
]
,
[
5
5
]
.
T
h
i
s
tech
n
iq
u
e
b
eg
in
s
with
a
s
in
g
l
e
n
o
d
e
an
d
th
en
b
r
an
c
h
es
o
u
t
to
cr
ea
te
m
o
r
e
n
o
d
es
f
o
r
ea
c
h
p
o
s
s
ib
ilit
y
[
5
6
]
.
E
ac
h
n
ew
n
o
d
e
h
as
th
e
p
o
ten
t
ial
to
b
r
an
ch
o
u
t
f
u
r
th
er
[
5
6
]
.
DT
is
ab
le
to
tr
ai
n
m
o
d
els
q
u
i
ck
ly
an
d
with
less
r
eq
u
ir
ed
m
em
o
r
y
[
3
1
]
.
DT
ca
n
b
e
u
tili
ze
d
in
a
n
u
m
b
er
o
f
way
s
,
in
clu
d
in
g
im
ag
e
p
r
o
ce
s
s
in
g
,
class
if
y
in
g
d
ata
an
d
p
atter
n
r
ec
o
g
n
itio
n
[
1
6
]
.
R
F
is
a
class
if
ier
co
n
s
is
tin
g
o
f
a
n
u
m
b
er
o
f
d
ec
is
io
n
tr
ee
s
[
5
7
]
.
T
h
e
u
s
e
o
f
R
F
o
f
f
er
s
ac
cu
r
ac
y
to
a
m
o
d
el
[
5
7
]
as we
ll a
s
s
p
ee
d
o
f
lear
n
in
g
[
5
8
]
.
Kh
an
et
a
l.
[
3
]
u
tili
ze
d
DT
al
o
n
g
with
e
n
s
em
b
le
-
lear
n
i
n
g
t
o
p
r
o
ce
s
s
th
e
o
u
tp
u
t
o
f
p
r
e
v
io
u
s
lay
er
s
o
f
th
e
m
o
d
el
b
e
f
o
r
e
m
ak
in
g
t
h
e
f
in
al
d
ec
is
io
n
o
n
wh
eth
er
th
e
d
ata
s
ig
n
if
ied
a
n
attac
k
.
T
r
a
n
et
a
l.
[
3
1
]
c
o
m
p
ar
e
d
a
s
tan
d
alo
n
e
DT
m
o
d
el
with
an
R
F
m
o
d
el
an
d
a
n
o
th
er
m
o
d
el
u
s
in
g
ex
tr
em
e
g
r
ad
ie
n
t
b
o
o
s
tin
g
(
XGBo
o
s
t)
.
C
o
n
v
er
s
ely
to
o
th
er
m
o
d
els
p
r
esen
ted
h
er
e,
C
h
ak
r
ab
o
r
ty
et
a
l
.
[
3
2
]
p
r
im
ar
ily
u
tili
ze
d
n
o
n
-
ML
tech
n
iq
u
es
f
o
r
th
eir
m
o
d
el,
b
u
t
u
tili
ze
d
d
if
f
er
en
t
s
u
p
er
v
is
ed
lear
n
in
g
tec
h
n
iq
u
es
f
o
r
attac
k
class
if
icatio
n
.
T
h
e
y
co
m
p
ar
ed
lo
g
is
tic
r
eg
r
ess
io
n
(
L
R
)
,
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e
(
SVM)
,
k
-
n
ea
r
est
n
eig
h
b
o
r
(
k
-
NN)
an
d
R
F.
L
ik
e
ANN
tech
n
iq
u
es,
s
u
p
er
v
is
ed
lear
n
in
g
ap
p
r
o
ac
h
es
p
r
o
v
id
e
s
o
m
e
f
lex
ib
ilit
y
to
b
e
u
tili
ze
d
in
d
if
f
er
en
t
way
s
with
d
if
f
er
en
t M
L
tec
h
n
iq
u
es.
T
h
er
e
wer
e
th
r
ee
en
s
em
b
le
m
eth
o
d
s
u
tili
ze
d
in
th
e
r
ev
iewe
d
liter
atu
r
e,
all
o
f
wh
ich
wer
e
g
r
ad
ien
t
b
o
o
s
tin
g
alg
o
r
ith
m
s
.
Gr
ad
ien
t
b
o
o
s
tin
g
alg
o
r
ith
m
s
ca
lcu
late
th
e
m
is
tak
es
o
f
ea
r
lier
m
o
d
el
s
b
y
cr
ea
tin
g
a
n
ew
m
o
d
el
[
5
9
]
.
T
h
ey
th
e
n
m
ak
e
a
ch
o
ice
b
ased
o
n
th
e
am
al
g
a
m
atio
n
o
f
t
h
e
n
ew
an
d
o
ld
m
o
d
els.
Gen
er
ally
,
th
e
in
clu
s
io
n
o
f
a
b
o
o
s
tin
g
alg
o
r
ith
m
ca
n
im
p
r
o
v
e
p
er
f
o
r
m
a
n
ce
[
5
9
]
.
T
h
e
m
o
s
t
co
m
m
o
n
en
s
em
b
le
m
eth
o
d
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
Ma
ch
in
e
lea
r
n
in
g
a
p
p
r
o
a
c
h
e
s
to
cy
b
ers
ec
u
r
ity
in
th
e
in
d
u
s
tr
ia
l
…
(
Mela
n
ie
Heie
r
)
3857
p
r
o
p
o
s
ed
was
XGBo
o
s
t
[
3
0
]
,
[
3
1
]
,
[
3
4
]
.
T
h
is
alg
o
r
ith
m
is
c
o
n
s
id
er
ed
to
h
av
e
l
o
w
r
eso
u
r
c
e
d
ep
e
n
d
en
c
y
an
d
a
f
ast s
p
ee
d
[
6
0
]
–
[
6
2
]
.
Sh
ah
in
e
t a
l.
[
3
0
]
c
o
n
f
ir
m
e
d
th
ese
o
b
s
e
r
v
atio
n
s
,
wh
ile
also
s
tatin
g
th
at
it p
er
f
o
r
m
ed
well
in
ter
m
s
o
f
n
etwo
r
k
i
n
tr
u
s
io
n
d
etec
tio
n
.
T
r
an
et
a
l.
[
3
1
]
s
im
ilar
ly
s
tated
th
at
th
is
m
et
h
o
d
h
as
p
er
f
o
r
m
ed
well
in
ter
m
s
o
f
f
au
lt
d
etec
tio
n
,
th
o
u
g
h
ca
n
,
if
n
o
t
u
s
ed
with
o
th
er
p
r
o
ce
s
s
e
s
,
in
cr
ea
s
e
th
e
r
eso
u
r
ce
s
r
eq
u
ir
e
d
.
I
t
is
ap
p
ar
en
t
f
r
o
m
th
e
liter
atu
r
e
t
h
at
in
ter
m
s
o
f
en
s
em
b
le
m
eth
o
d
s
,
g
r
ad
ien
t
b
o
o
s
tin
g
alg
o
r
ith
m
s
in
p
ar
ticu
lar
ar
e
p
o
p
u
lar
m
eth
o
d
s
to
u
s
e
in
ML
ap
p
r
o
ac
h
es to
I
I
o
T
cy
b
er
s
ec
u
r
ity
.
Sem
i
-
s
u
p
er
v
is
ed
lear
n
in
g
ca
n
p
r
o
ce
s
s
b
o
th
u
n
lab
eled
an
d
lab
elled
d
ata
[
2
]
,
[
6
3
]
.
T
h
is
tech
n
iq
u
e
u
tili
ze
s
u
n
s
u
p
er
v
is
ed
lear
n
in
g
w
ith
u
n
lab
eled
d
ata
to
ex
tr
ac
t
f
ea
tu
r
es
f
r
o
m
it.
I
t
th
e
n
u
s
es
s
u
p
er
v
is
ed
lear
n
i
n
g
to
in
co
r
p
o
r
ate
a
s
m
all
am
o
u
n
t
o
f
lab
elled
d
ata
to
ca
lib
r
ate
th
e
f
ea
tu
r
es
an
d
c
o
n
s
tr
u
ct
t
h
e
m
o
d
el
f
o
r
u
s
e
in
attac
k
d
etec
tio
n
[
2
]
,
[
6
3
]
.
T
h
is
tech
n
iq
u
e
is
u
s
ed
b
y
Ao
u
ed
i
et
a
l.
[
2
]
with
in
th
eir
f
ed
er
ate
d
lear
n
in
g
m
o
d
el
t
o
s
o
lv
e
th
e
is
s
u
e
in
I
I
o
T
cy
b
e
r
s
ec
u
r
ity
o
f
n
ee
d
i
n
g
to
ex
a
m
in
e
lar
g
e
am
o
u
n
ts
o
f
u
n
lab
ele
d
d
ata
to
d
eter
m
in
e
wh
eth
er
an
attac
k
wo
u
ld
h
av
e
o
cc
u
r
r
ed
.
C
o
n
v
er
s
ely
,
J
ian
g
[
4
5
]
u
tili
ze
s
s
em
i
-
s
u
p
er
v
is
ed
lear
n
in
g
in
th
eir
m
o
d
el
to
d
etec
t
v
o
ltag
e
g
litch
attac
k
s
(
VGA)
f
r
o
m
g
litch
es
in
p
o
wer
s
ig
n
als
f
r
o
m
a
n
I
I
o
T
m
ac
h
in
e.
T
a
b
le
3
s
h
o
ws
th
e
d
atasets
u
ti
lized
b
y
th
ese
ap
p
r
o
ac
h
es
to
ex
a
m
in
e
p
er
f
o
r
m
an
ce
.
T
h
ese
d
atasets
in
clu
d
ed
b
o
th
n
etwo
r
k
o
r
d
ev
ice
d
ata
a
n
d
e
q
u
ip
m
en
t
g
litch
es,
co
v
er
in
g
a
r
an
g
e
o
f
p
o
ten
tial
attac
k
s
u
r
f
ac
es
in
I
I
o
T
.
T
h
ese
d
if
f
er
in
g
ap
p
licatio
n
s
o
f
s
em
i
-
s
u
p
er
v
is
ed
lear
n
i
n
g
d
em
o
n
s
tr
ate
th
at
it
is
a
v
e
r
s
atile
an
d
f
lex
ib
le
a
p
p
r
o
ac
h
s
u
it
ab
le
f
o
r
cy
b
er
s
ec
u
r
ity
a
p
p
l
icatio
n
s
in
I
I
o
T
.
As
ca
n
b
e
s
ee
n
i
n
T
ab
les 2
an
d
3
,
p
r
o
p
o
s
ed
s
o
lu
tio
n
s
in
th
e
cu
r
r
en
t
liter
atu
r
e
u
s
e
n
o
t
ju
s
t
a
m
u
ltit
u
d
e
o
f
ML
tech
n
i
q
u
es,
b
u
t
test
ag
ain
s
t
m
an
y
d
if
f
e
r
en
t
d
atasets
,
co
v
er
in
g
a
wid
e
r
a
n
g
e
o
f
atta
ck
s
ce
n
ar
io
s
.
T
h
e
y
also
u
tili
ze
m
an
y
d
if
f
er
en
t
ev
a
lu
atio
n
m
etr
ics
an
d
v
ar
ia
b
les,
m
ak
in
g
it
d
if
f
icu
lt
to
d
r
aw
c
o
m
p
ar
is
o
n
s
b
etwe
en
th
e
p
er
f
o
r
m
a
n
ce
o
f
d
if
f
er
en
t
m
o
d
els
as
th
is
p
a
p
er
attem
p
ts
.
T
h
er
e
wer
e
s
o
m
e
c
o
m
m
o
n
alities
am
o
n
g
th
e
p
r
o
g
r
a
m
m
in
g
lan
g
u
a
g
es,
to
o
ls
an
d
lib
r
a
r
ies w
h
er
e
th
ese
wer
e
m
en
tio
n
ed
in
th
e
s
tu
d
ies.
I
n
ter
m
s
o
f
d
atasets
,
th
er
e
wer
e
th
r
ee
m
o
s
t
co
m
m
o
n
in
u
s
e
in
th
e
liter
atu
r
e:
a
g
as
p
ip
elin
e
SC
AD
A
s
y
s
tem
d
ataset
[
2
]
,
[
3
3
]
UNS
W
-
NB
1
5
[
5
]
,
[
2
9
]
a
n
d
T
o
N
-
I
o
T
[
2
9
]
,
[
3
0
]
.
T
h
ese
d
atasets
ar
e
all
b
ased
o
n
I
I
o
T
an
d
c
o
v
er
a
wid
e
r
an
g
e
o
f
att
ac
k
s
ce
n
ar
io
s
,
as
ca
n
b
e
s
ee
n
in
T
a
b
les
2
a
n
d
3
.
Usi
n
g
th
e
s
am
e
d
atasets
ca
n
m
ak
e
it
ea
s
ier
to
m
ak
e
co
m
p
ar
is
o
n
s
b
etwe
en
d
if
f
er
en
t
ap
p
r
o
ac
h
es
[
6
4
]
.
Fo
r
e
x
am
p
le,
Hu
m
a
et
a
l.
[
5
]
a
n
d
Sh
ah
in
et
a
l.
[
2
9
]
b
o
t
h
u
s
e
t
h
e
UNSW
-
NB
1
5
d
ataset,
m
ak
in
g
it
ea
s
ier
t
o
co
m
p
ar
e
th
e
r
esu
lts
ac
h
iev
ed
b
y
th
eir
r
esp
ec
tiv
e
m
o
d
els.
Similar
ly
,
d
if
f
er
in
g
ev
alu
atio
n
m
etr
ics m
ak
e
m
o
d
els d
if
f
icu
lt to
co
m
p
ar
e.
Fo
r
ex
am
p
le,
Fu
et
a
l.
[
2
8
]
p
r
o
p
o
s
ed
a
h
ier
a
r
ch
ical
ab
n
o
r
m
al
tr
af
f
ic
d
etec
tio
n
m
eth
o
d
u
tili
zin
g
an
u
n
s
u
p
er
v
is
ed
clu
s
ter
in
g
alg
o
r
ith
m
.
T
h
is
m
o
d
el
was
ab
le
to
d
ete
ct
an
o
m
alies
in
th
e
s
h
o
r
test
am
o
u
n
t
o
f
tim
e
i
n
c
o
m
p
ar
is
o
n
to
o
th
e
r
s
elec
ted
m
o
d
els.
Ho
wev
er
,
as
th
is
m
o
d
el
d
id
n
o
t
u
s
e
an
y
o
f
th
e
m
etr
ics
co
m
m
o
n
to
o
t
h
er
r
e
v
iewe
d
m
o
d
els,
its
p
er
f
o
r
m
an
ce
is
n
o
t
ea
s
ily
co
m
p
ar
ab
le
in
th
is
r
ev
iew.
T
h
eir
d
ata
was
also
m
ain
ly
p
r
esen
ted
in
th
e
f
o
r
m
o
f
b
a
r
g
r
ap
h
s
,
r
at
h
er
th
a
n
n
u
m
er
ically
,
m
ak
in
g
s
co
r
e
in
ter
p
r
etatio
n
p
o
ten
tially
in
ac
cu
r
ate.
T
h
e
r
an
g
e
o
f
v
ar
iab
les
s
h
o
wn
in
T
ab
les
2
an
d
3
,
wh
ile
p
r
o
v
i
d
in
g
ex
ce
llen
t
d
ata
with
in
s
in
g
le
p
ap
er
s
,
ca
n
ag
ain
m
ak
e
c
o
m
p
ar
is
o
n
t
r
o
u
b
leso
m
e
b
etwe
en
s
ep
ar
ate
ex
p
er
im
e
n
ts
.
As
ca
n
b
e
s
ee
n
in
T
ab
le
3
,
m
o
d
el
co
m
p
ar
is
o
n
s
wer
e
th
e
m
o
s
t p
r
o
m
in
en
t v
ar
iab
l
e
[
2
9
]
–
[
3
4
]
.
C
o
m
p
ar
in
g
m
o
d
els
u
s
in
g
th
e
s
a
m
e
v
ar
ia
b
les
ca
n
b
e
v
er
y
u
s
ef
u
l
to
d
eter
m
in
e
th
e
p
er
f
o
r
m
an
c
e
o
f
d
if
f
er
en
t
ML
tech
n
iq
u
es.
C
o
m
p
ar
i
n
g
p
er
f
o
r
m
an
ce
b
ased
o
n
d
ataset
o
r
attac
k
ty
p
e
was
als
o
co
m
m
o
n
[
3
]
,
[
2
9
]
–
[
3
1
]
.
W
h
en
it
ca
m
e
to
ac
tu
al
p
ar
am
eter
s
o
f
th
e
m
o
d
e
ls
,
th
e
n
u
m
b
er
o
f
clien
ts
was
th
e
m
o
s
t
co
m
m
o
n
v
ar
iab
le
f
o
r
c
o
m
p
ar
in
g
p
er
f
o
r
m
an
ce
[
2
]
,
[
3
]
,
[
3
4
]
.
T
h
is
is
an
im
p
o
r
tan
t p
o
in
t o
f
co
m
p
ar
is
o
n
,
as th
e
n
u
m
b
er
o
f
d
ev
ices w
it
h
in
an
I
I
o
T
e
n
v
ir
o
n
m
en
t c
o
u
ld
v
ar
y
.
W
h
ile
in
p
u
t
a
n
d
o
u
tp
u
t
c
o
m
p
o
n
en
ts
wer
e
m
an
y
a
n
d
v
ar
ied
,
to
o
ls
u
s
ed
b
y
t
h
e
d
if
f
er
e
n
t
a
p
p
r
o
ac
h
es
wer
e
f
ewe
r
.
Of
th
o
s
e
th
at
m
e
n
tio
n
ed
th
e
p
r
o
g
r
am
m
i
n
g
lan
g
u
ag
e
u
s
ed
,
all
u
tili
ze
d
p
y
th
o
n
[
2
]
,
[
3
]
,
[
4
6
]
.
T
h
e
m
o
s
t
co
m
m
o
n
ly
u
s
ed
to
o
l
wa
s
Go
o
g
le
C
o
llab
o
r
ato
r
y
[
3
4
]
,
[
4
5
]
.
Of
t
h
e
f
r
am
ewo
r
k
s
an
d
l
ib
r
ar
ies
m
en
tio
n
ed
,
Py
to
r
ch
[
2
]
,
[
3
]
,
[
3
4
]
a
n
d
Scik
it
-
L
ea
r
n
[
2
]
,
[
2
9
]
,
[
3
0
]
wer
e
m
o
s
t
co
m
m
o
n
.
So
m
e
to
o
ls
an
d
lib
r
ar
ies
wer
e
u
n
s
p
ec
if
ied
in
t
h
e
r
ev
iewe
d
lit
er
atu
r
e
[
3
2
]
,
[
4
6
]
.
T
h
e
m
o
s
t
r
ec
en
tly
p
r
o
p
o
s
ed
s
o
lu
tio
n
s
to
I
I
o
T
cy
b
e
r
s
ec
u
r
ity
th
at
u
tili
ze
ML
ap
p
r
o
ac
h
es
h
av
e
n
o
t
y
et
b
ee
n
co
n
s
o
lid
ated
a
n
d
e
v
alu
at
ed
.
T
h
is
p
a
p
er
will
r
e
v
iew
th
e
s
e
s
o
lu
tio
n
s
to
p
r
o
v
id
e
an
o
v
e
r
v
iew
o
f
th
e
c
u
r
r
en
t
s
tate
o
f
ML
ap
p
r
o
ac
h
es to
cy
b
er
s
ec
u
r
ity
in
I
I
o
T
.
T
h
e
k
ey
r
esear
ch
q
u
esti
o
n
s
in
clu
d
e:
a.
W
h
at
ar
e
th
e
cy
b
er
s
ec
u
r
ity
co
n
ce
r
n
s
with
in
th
e
I
I
o
T
?
b.
W
h
at
ar
e
th
e
m
o
s
t
r
ec
en
t
m
ac
h
in
e
lear
n
in
g
ap
p
r
o
ac
h
es
b
ein
g
p
r
o
p
o
s
ed
to
s
o
lv
e
th
es
e
cy
b
e
r
s
ec
u
r
ity
co
n
ce
r
n
s
?
c.
W
h
at
ar
e
th
e
ad
v
an
ta
g
es a
n
d
d
is
ad
v
an
tag
es o
f
th
ese
ap
p
r
o
ac
h
es?
d.
W
h
at
s
o
f
twar
e
an
d
p
r
o
g
r
am
m
in
g
la
n
g
u
a
g
es
ar
e
b
ein
g
u
s
ed
t
o
im
p
lem
en
t
m
ac
h
in
e
lear
n
in
g
ap
p
r
o
ac
h
es
to
cy
b
er
s
ec
u
r
ity
f
o
r
I
I
o
T
?
2.
M
E
T
H
O
D
T
h
is
r
ev
iew
u
tili
ze
d
C
h
ar
les
S
tu
r
t
Un
iv
er
s
ity
lib
r
ar
y
r
eso
u
r
c
es,
s
p
ec
if
ically
h
ttp
s
:
//p
r
i
mo
.
csu
.
ed
u
.
a
u
to
lo
ca
te
ap
p
r
o
p
r
iate
ar
ticles
f
o
r
th
e
to
p
ic.
I
n
clu
s
io
n
cr
iter
ia
:
i)
Pu
b
lis
h
ed
in
2
0
2
1
o
r
later
,
ii)
Peer
r
ev
iewe
d
,
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
.
4
,
Au
g
u
s
t
20
25
:
3
8
5
1
-
3866
3858
iii)
Hav
e
a
jo
u
r
n
al
r
atin
g
o
f
Q1
o
r
Q
2
ac
c
o
r
d
in
g
to
Scim
ag
o
J
o
u
r
n
al
an
d
C
o
u
n
tr
y
R
an
k
(
SJ
R
)
jo
u
r
n
a
l
r
an
k
in
g
s
,
a
n
d
iv
)
Pro
p
o
s
e
a
ML
ap
p
r
o
ac
h
to
c
y
b
er
s
ec
u
r
ity
in
I
I
o
T
.
T
h
e
r
esear
ch
m
et
h
o
d
o
lo
g
y
f
o
r
th
is
p
r
o
ject
is
o
u
tlin
ed
i
n
Fig
u
r
e
3
an
d
is
as
f
o
llo
ws.
First,
s
ea
r
ch
es
wer
e
p
er
f
o
r
m
ed
in
o
r
d
er
to
f
o
r
m
u
late
th
e
to
p
ic.
I
n
itial
k
ey
wo
r
d
s
u
s
ed
wer
e
“so
f
twar
e
d
esig
n
OR
s
o
f
twar
e
d
ev
elo
p
m
e
n
t”
a
n
d
“c
y
b
er
s
ec
u
r
ity
”.
T
h
e
r
esu
lts
o
f
th
is
s
ea
r
ch
wer
e
th
en
g
r
o
u
p
e
d
in
to
c
o
m
m
o
n
to
p
ics,
an
d
ad
d
itio
n
al
k
ey
wo
r
d
s
ad
d
ed
in
clu
d
in
g
“m
ac
h
in
e
lear
n
in
g
O
R
d
ee
p
lear
n
in
g
”
an
d
“I
I
o
T
OR
I
n
d
u
s
tr
ial
I
n
ter
n
et
o
f
T
h
i
n
g
s
OR
in
d
u
s
tr
y
4
.
0
”.
T
h
e
to
p
ic
o
f
ML
ap
p
r
o
ac
h
es
to
cy
b
e
r
s
ec
u
r
ity
was
th
en
s
el
ec
ted
b
ased
o
n
t
h
e
co
m
m
o
n
to
p
ics
o
f
ar
ti
cles
f
o
u
n
d
.
T
h
e
r
esu
ltin
g
co
llectio
n
o
f
ar
ticles
was
th
en
s
cr
ee
n
ed
an
d
s
elec
ted
ac
co
r
d
in
g
to
th
e
in
clu
s
io
n
cr
it
er
ia
o
u
tlin
ed
ab
o
v
e
an
d
th
eir
s
u
itab
ilit
y
f
o
r
th
e
to
p
ic.
T
h
e
s
co
p
e
f
o
r
th
is
r
ev
iew
was lim
ited
to
twelv
e
p
ap
er
s
d
u
e
to
ass
ig
n
m
en
t r
e
q
u
ir
em
e
n
ts
.
Fig
u
r
e
3
.
R
esear
ch
m
eth
o
d
o
l
o
g
y
.
Ad
ap
te
d
f
r
o
m
Dea
k
in
U
n
iv
er
s
ity
[
6
5
]
Data
ex
tr
ac
tio
n
was
p
er
f
o
r
m
ed
with
th
e
u
s
e
o
f
E
x
ce
l
s
p
r
ea
d
s
h
ee
ts
.
A
b
r
o
a
d
f
ea
tu
r
e
a
n
aly
s
is
was
co
m
p
leted
,
wh
ich
i
n
v
o
lv
e
d
s
u
m
m
ar
izin
g
t
h
e
f
o
llo
win
g
f
ea
tu
r
es
o
f
ea
ch
p
ap
e
r
:
i)
p
r
o
b
lem
d
ef
i
n
itio
n
,
ii)
p
r
o
p
o
s
ed
s
o
lu
tio
n
,
iii)
ad
v
an
tag
es
an
d
d
is
ad
v
an
tag
es
,
iv
)
m
eth
o
d
,
s
tep
s
,
an
d
/o
r
s
tag
es
,
v
)
lim
itatio
n
s
an
d
ju
s
tific
atio
n
s
,
v
i)
ch
allen
g
es
,
v
ii)
h
ar
d
wa
r
e,
s
o
f
twar
e
an
d
p
r
o
g
r
am
m
in
g
lan
g
u
a
g
es
,
v
ii
i)
m
o
d
els
u
s
ed
f
o
r
co
m
p
ar
is
o
n
to
th
e
p
r
o
p
o
s
ed
s
o
lu
tio
n
,
an
d
ix
)
f
u
t
u
r
e
wo
r
k
s
u
g
g
ested
.
T
h
e
s
p
ec
if
ic
tech
n
iq
u
es
u
s
ed
in
th
e
r
ev
iewe
d
ar
ticles
wer
e
co
n
s
o
li
d
ated
,
as
wer
e
th
e
d
atasets
,
i
m
p
lem
en
tatio
n
p
r
o
ce
d
u
r
es,
ev
alu
atio
n
cr
iter
ia
an
d
r
esu
lts
.
Fin
ally
,
th
is
ex
t
r
ac
ted
d
ata
was u
tili
ze
d
to
co
m
p
lete
t
h
is
f
in
al
r
ep
o
r
t.
T
h
e
co
n
s
o
lid
ated
tech
n
iq
u
es
wer
e
r
ev
iewe
d
to
d
eter
m
in
e
t
h
o
s
e
th
at
wer
e
m
o
s
t
u
s
ed
b
y
th
e
p
ap
e
r
s
u
n
d
er
r
e
v
iew.
T
h
ese
co
m
m
o
n
tech
n
iq
u
es
wer
e
g
r
o
u
p
ed
in
to
th
e
ca
teg
o
r
ies
o
f
ANN,
s
u
p
er
v
is
ed
lear
n
in
g
an
d
e
n
s
em
b
le
m
eth
o
d
s
as
s
h
o
wn
i
n
T
ab
le
1
.
T
h
ese
co
m
m
o
n
tec
h
n
iq
u
es
wer
e
u
s
ed
f
o
r
d
is
cu
s
s
io
n
an
d
co
m
p
ar
is
o
n
in
o
r
d
er
t
o
m
ain
tain
th
e
f
o
cu
s
an
d
s
co
p
e
o
f
th
e
p
a
p
er
.
Fo
r
e
x
am
p
le,
C
h
ak
r
ab
o
r
ty
et
a
l.
[
3
2
]
u
tili
ze
d
L
R
,
SVM
an
d
k
-
NN,
h
o
wev
er
it
was
th
e
o
n
ly
p
ap
e
r
am
o
n
g
th
e
twel
v
e
r
ev
iewe
d
to
u
s
e
th
ese
tech
n
iq
u
es,
an
d
s
o
t
h
ey
wer
e
ex
clu
d
ed
f
r
o
m
in
-
d
ep
th
d
is
cu
s
s
io
n
.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
3
.
1
.
Resul
t
s
T
ab
le
4
d
is
p
lay
s
th
e
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all
an
d
/
o
r
F1
s
c
o
r
es f
o
r
ea
ch
p
ap
er
’
s
b
est s
co
r
in
g
m
o
d
el.
W
h
er
e
m
o
d
els
wer
e
co
m
p
ar
e
d
to
s
tate
-
of
-
th
e
-
ar
t
tech
n
iq
u
e
s
with
in
th
e
p
ap
er
,
th
ese
co
m
p
ar
is
o
n
s
co
r
es
wer
e
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
Ma
ch
in
e
lea
r
n
in
g
a
p
p
r
o
a
c
h
e
s
to
cy
b
ers
ec
u
r
ity
in
th
e
in
d
u
s
tr
ia
l
…
(
Mela
n
ie
Heie
r
)
3859
tak
en
f
o
r
T
ab
le
4
.
T
h
e
v
ar
iab
les
co
lu
m
n
s
h
o
ws
an
y
v
ar
iab
l
es
ass
o
ciate
d
with
th
e
ac
h
iev
em
en
t
o
f
th
o
s
e
b
est
s
co
r
es.
Fu
et
a
l.
[
2
8
]
was
th
e
o
n
ly
p
a
p
er
th
at
d
id
n
o
t
u
s
e
an
y
o
f
th
ese
co
m
m
o
n
e
v
alu
atio
n
m
etr
ics
as
n
o
ted
in
th
e
tab
le.
Sco
r
es
wer
e
o
n
ly
i
n
clu
d
ed
in
T
a
b
le
4
wh
er
e
th
e
p
r
ec
is
e
s
co
r
e
was
s
tated
b
y
th
e
au
th
o
r
s
;
s
co
r
es
d
is
p
lay
ed
o
n
l
y
in
g
r
ap
h
f
o
r
m
wer
e
ex
clu
d
ed
.
Acc
u
r
ac
y
is
th
e
m
etr
ic
m
o
s
t
u
tili
ze
d
b
y
liter
atu
r
e
an
d
s
o
is
d
is
cu
s
s
ed
h
er
e
in
f
u
r
th
er
d
etail.
Acc
u
r
ac
y
is
d
escr
ib
ed
as
th
e
p
er
ce
n
t
ag
e
o
r
r
atio
o
f
c
o
r
r
ec
t
p
r
e
d
ictio
n
s
[
5
]
,
[
2
9
]
,
[
4
6
]
.
As
ca
n
b
e
s
ee
n
in
T
ab
le
4
,
t
h
e
h
ig
h
est
ac
cu
r
ac
y
o
f
1
0
0
%
was
ac
h
iev
ed
b
y
Sh
a
h
in
et
a
l.
[
2
9
]
o
n
th
e
B
o
T
-
I
o
T
an
d
UNSW
-
NB
15
d
atasets
with
th
eir
m
o
d
el
u
s
i
n
g
L
STM
an
d
FC
NN
tech
n
iq
u
es.
Sh
ah
in
et
a
l.
[
3
0
]
d
id
al
s
o
ac
h
iev
e
a
1
0
0
%
ac
cu
r
ac
y
f
o
r
o
n
e
o
f
th
e
d
ev
ic
es
in
th
eir
s
tu
d
y
,
h
o
wev
er
as
b
o
t
h
o
f
th
eir
m
o
d
els
ac
h
iev
ed
1
0
0
%
in
th
at
ca
s
e,
th
e
n
ex
t
h
ig
h
est
s
co
r
e
wo
u
l
d
h
av
e
b
ee
n
tak
en
th
at
d
if
f
e
r
en
tiated
th
e
m
o
d
els.
Un
f
o
r
t
u
n
ately
,
s
in
ce
th
eir
ac
cu
r
ac
y
s
co
r
es we
r
e
p
r
esen
te
d
o
n
ly
in
b
ar
g
r
a
p
h
f
o
r
m
,
a
s
p
ec
if
ic
ac
cu
r
ac
y
s
co
r
e
was n
o
t
ab
le
to
b
e
d
is
ce
r
n
ed
f
o
r
in
clu
s
io
n
in
t
h
e
tab
le.
T
h
e
lo
west
ac
cu
r
ac
y
s
co
r
e
o
f
.
7
8
was
ac
h
iev
ed
b
y
th
e
m
o
d
el
u
tili
zin
g
p
r
im
ar
ily
non
-
ML
tec
h
n
iq
u
es,
s
u
g
g
esti
n
g
th
at
ML
tech
n
iq
u
es
g
e
n
er
ally
h
av
e
s
u
p
er
io
r
p
er
f
o
r
m
an
ce
in
th
is
ar
ea
.
T
ec
h
n
iq
u
es
u
tili
ze
d
in
m
o
d
el
s
ac
h
iev
in
g
9
9
%
o
r
ab
o
v
e
a
cc
u
r
ac
y
in
clu
d
e
FL
[
3
3
]
,
A
NN
[
5
]
,
[
2
9
]
,
[
3
3
]
,
en
s
em
b
le
m
eth
o
d
s
[
3
1
]
,
s
u
p
e
r
v
is
ed
lear
n
in
g
[
3
1
]
,
an
d
d
ee
p
lear
n
in
g
[
5
]
.
T
h
e
o
n
ly
ML
tech
n
iq
u
e
u
s
ed
b
y
m
o
r
e
th
an
o
n
e
o
f
th
ese
h
ig
h
ac
cu
r
ac
y
ac
h
iev
i
n
g
m
o
d
e
ls
was
ML
P.
T
h
ese
r
esu
lt
s
d
em
o
n
s
tr
ate
th
at
ANN
tech
n
iq
u
es in
p
a
r
ticu
lar
ar
e
s
u
cc
ess
f
u
lly
b
ein
g
u
tili
ze
d
in
ML
s
o
lu
tio
n
s
to
I
I
o
T
cy
b
er
s
ec
u
r
ity
.
T
ab
le
4
.
E
v
alu
atio
n
B
e
st
m
o
d
e
l
R
e
f
[
#
]
Te
c
h
n
i
q
u
e
s
A
c
c
u
r
a
c
y
P
r
e
c
i
s
i
o
n
R
e
c
a
l
l
F1
V
a
r
i
a
b
l
e
s
F
e
d
e
r
a
t
e
d
sem
i
-
su
p
e
r
v
i
s
e
d
l
e
a
r
n
i
n
g
sc
h
e
me
[
2
]
A
E,
F
L,
F
C
N
,
SL
9
5
.
8
4
%
9
7
.
8
9
8
7
.
1
5
O
v
e
r
a
l
l
S
c
o
r
e
s
D
L
b
a
se
d
S
C
A
D
A
n
e
t
w
o
r
k
b
a
s
e
d
c
y
b
e
r
a
t
t
a
c
k
d
e
t
e
c
t
i
o
n
sch
e
me
[
3
]
P
R
U
,
D
T,
LSTM
,
EL
9
8
.
8
9
%
b
i
n
a
r
y
c
l
a
ss
i
f
i
c
a
t
i
o
n
w
i
t
h
d
a
t
a
se
t
1
H
D
R
a
N
N
[
5
]
M
LP,
R
a
N
N
0
.
9
9
1
9
0
.
9
9
0
7
0
.
9
8
9
8
0
.
9
9
0
2
d
a
t
a
se
t
:
U
N
S
W
-
N
B
1
5
H
e
u
r
i
s
t
i
c
s
e
m
i
-
s
u
p
e
r
v
i
s
e
d
l
e
a
r
n
i
n
g
m
e
t
h
o
d
[
4
5
]
C
A
,
E
LM
,
SSL
9
0
.
7
9
0
.
7
9
0
.
7
9
0
.
7
Ti
mes
l
o
t
:
1
0
S
e
c
u
r
e
n
e
t
w
o
r
k
m
o
d
e
l
[
4
6
]
D
L,
M
L
0
.
8
7
0
.
9
0
7
0
.
8
6
4
0
.
8
8
1
O
v
e
r
a
l
l
S
c
o
r
e
s
S
e
c
u
r
e
c
l
u
st
e
r
i
n
g
a
l
g
o
r
i
t
h
m
f
o
r
c
o
mp
l
e
x
a
t
t
r
i
b
u
t
e
f
e
a
t
u
r
e
s
[
2
8
]
CA
u
s
e
d
o
n
l
y
T
P
a
n
d
FP
e
v
a
l
u
a
t
i
o
n
m
e
t
r
i
c
s
a
n
d
o
n
l
y
p
r
e
s
e
n
t
e
d
i
n
b
a
r
g
r
a
p
h
f
o
r
m
L
S
T
M
-
F
C
N
a
n
d
L
S
T
M
-
F
C
N
5
-
f
o
l
d
s
C
V
[
2
9
]
LSTM
,
F
C
N
N
1
0
0
%
d
a
t
a
se
t
s:
B
o
T
-
I
o
T
a
n
d
U
N
S
W
-
N
B
1
5
D
e
e
p
h
y
b
r
i
d
l
e
a
r
n
i
n
g
m
o
d
e
l
[
3
0
]
A
LSTM
,
F
C
N
N
,
A
d
a
B
o
o
st
9
9
.
9
0
%
9
9
.
9
0
%
9
9
.
9
0
%
b
o
o
s
t
e
r
:
A
d
a
B
o
o
s
t
,
d
e
v
i
c
e
:
G
P
S
O
n
l
i
n
e
f
a
u
l
t
d
i
a
g
n
o
s
i
s
w
i
t
h
R
F
[
3
1
]
RF
9
9
.
0
3
%
O
v
e
r
a
l
l
S
c
o
r
e
s
F
M
4
:
f
u
n
c
t
i
o
n
a
l
p
o
si
t
i
o
n
a
n
d
v
e
l
o
c
i
t
y
mo
d
e
l
[
3
2
]
F
S
A
,
F
P
C
A
,
RF
0
.
7
8
1
0
.
7
8
seg
m
e
n
t
s
i
z
e
4
0
o
r
2
0
0
D
e
e
p
F
e
d
[
3
3
]
M
LP,
C
N
N
,
G
R
U
,
F
L,
P
a
i
l
l
i
e
r
,
A
ES
9
9
.
2
0
%
9
8
.
8
5
%
9
7
.
4
7
%
9
8
.
1
4
%
n
u
m
c
l
i
e
n
t
s:
7
,
c
o
mm
.
R
o
u
n
d
s
:
1
0
S
e
c
u
r
e
I
I
o
T
-
C
N
N
m
o
d
e
l
[
3
4
]
F
L,
C
N
N
,
R
F
0
.
5
1
0
.
9
7
0
.
6
7
t
e
st
i
n
g
,
t
r
a
i
n
:
0
S
e
l
e
c
t
e
d
r
e
su
l
t
s
o
f
e
a
c
h
p
a
p
e
r’
s
b
e
s
t
p
e
rf
o
rm
i
n
g
m
o
d
e
l
i
n
c
l
u
d
i
n
g
a
ss
o
c
i
a
t
e
d
v
a
r
i
a
b
l
e
s
i
f
a
l
t
e
rn
a
t
i
v
e
v
a
r
i
a
b
l
e
v
a
l
u
e
s w
e
re
a
ss
o
c
i
a
t
e
d
w
i
t
h
d
i
f
f
e
re
n
t
s
c
o
res.
A
b
b
r
e
v
i
a
t
i
o
n
s
u
se
d
i
n
t
a
b
l
e
a
re
l
i
s
t
e
d
i
n
t
h
e
A
p
p
e
n
d
i
x
.
3
.
2
.
Dis
cu
s
s
io
n
T
h
e
aim
o
f
th
is
r
ev
iew
was
to
p
r
o
v
id
e
a
n
o
v
er
v
iew
o
f
th
e
cu
r
r
en
t
s
tate
o
f
ML
s
o
lu
tio
n
s
to
cy
b
er
s
ec
u
r
ity
in
I
I
o
T
b
y
e
x
am
in
in
g
th
e
p
r
o
p
o
s
ed
s
o
lu
tio
n
s
f
r
o
m
r
ec
en
t
y
ea
r
s
.
B
y
u
n
d
er
ta
k
in
g
th
is
ex
am
in
atio
n
,
th
is
r
ev
iew
p
r
o
v
id
es
in
s
ig
h
ts
in
to
wh
at
wo
r
k
s
in
o
r
d
er
t
o
in
f
o
r
m
f
u
t
u
r
e
r
esear
ch
o
r
th
e
d
ev
elo
p
m
e
n
t
o
f
r
ea
l
-
w
o
r
ld
s
o
lu
tio
n
s
.
T
h
e
m
o
s
t
co
m
m
o
n
l
y
p
r
o
p
o
s
ed
ML
tech
n
iq
u
es
wer
e
d
is
cu
s
s
ed
an
d
co
m
p
ar
ed
,
alo
n
g
with
o
th
e
r
asp
ec
ts
o
f
s
tu
d
ies
in
to
I
I
o
T
cy
b
er
s
ec
u
r
ity
s
o
lu
tio
n
s
s
u
c
h
as
d
ata
s
ets
an
d
ev
alu
atio
n
m
etr
ics.
T
h
is
r
e
v
iew
s
h
o
wed
th
at
s
o
m
e
o
f
t
h
e
m
o
s
t
p
r
o
m
is
in
g
ML
tech
n
iq
u
es
f
o
r
ap
p
licatio
n
i
n
I
I
o
T
cy
b
er
s
ec
u
r
ity
i
n
clu
d
e
FL,
FC
NN,
R
F a
n
d
s
em
i
-
s
u
p
er
v
is
ed
lear
n
in
g
.
I
t
is
clea
r
f
r
o
m
th
e
c
u
r
r
en
t
liter
atu
r
e
th
at
cy
b
er
s
ec
u
r
ity
in
I
I
o
T
is
o
f
g
r
o
win
g
co
n
ce
r
n
d
u
e
to
t
h
e
p
o
ten
tial
co
n
s
eq
u
e
n
ce
s
o
f
an
attac
k
[
2
]
,
[
5
]
,
[
3
4
]
,
an
d
th
e
v
u
ln
er
ab
ilit
y
o
f
I
I
o
T
d
e
v
ices
an
d
n
etwo
r
k
s
[
2
]
,
[
3
]
,
[
5
]
,
[
2
9
]
,
[
3
2
]
–
[
3
4
]
.
T
h
e
liter
atu
r
e
id
e
n
tifi
es
th
at
in
d
u
s
tr
ies
u
tili
zin
g
I
I
o
T
ar
e
in
cr
ea
s
in
g
ly
b
ei
n
g
ta
r
g
eted
b
y
cy
b
er
-
attac
k
s
o
f
v
ar
y
i
n
g
f
o
r
m
s
[
5
]
,
[
2
9
]
,
[
3
0
]
,
[
3
2
]
a
n
d
th
at
th
e
d
ata
b
ein
g
s
to
r
e
d
in
an
d
tr
an
s
f
er
r
ed
b
etwe
en
th
ese
d
ev
ices
r
eq
u
ir
es
p
r
iv
ac
y
p
r
o
tectio
n
[
2
]
,
[
3
]
,
[
5
]
,
[
3
2
]
–
[
3
4
]
.
I
n
te
r
m
s
o
f
in
tr
u
s
io
n
an
d
attac
k
m
ec
h
an
is
m
s
f
o
r
I
I
o
T
,
th
e
liter
atu
r
e
ag
r
ee
s
th
at
ac
cu
r
ac
y
[
3
]
,
[
2
8
]
,
[
3
2
]
a
n
d
ef
f
icien
cy
[
2
]
,
[
3
]
,
[
3
2
]
ar
e
v
er
y
im
p
o
r
tan
t.
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
.
4
,
Au
g
u
s
t
20
25
:
3
8
5
1
-
3866
3860
3
.
2
.
1
.
Arc
hite
ct
ure
I
n
o
r
d
er
to
ac
h
iev
e
ac
c
u
r
ac
y
,
ef
f
icien
c
y
,
an
d
s
ec
u
r
ity
i
n
I
I
o
T
cy
b
e
r
s
ec
u
r
ity
s
o
lu
tio
n
s
,
v
ar
io
u
s
m
ac
h
in
e
lear
n
in
g
tech
n
i
q
u
es
a
r
e
p
r
o
p
o
s
ed
in
th
e
liter
atu
r
e.
T
h
e
s
o
lu
tio
n
s
p
r
o
p
o
s
ed
ar
e
d
iv
i
d
ed
in
t
o
two
ty
p
es
o
f
ar
ch
itectu
r
e:
attac
k
d
etec
ti
o
n
ar
ch
itectu
r
e
a
n
d
FL
ar
ch
it
ec
tu
r
e.
An
attac
k
d
etec
tio
n
a
r
ch
itectu
r
e
p
r
o
v
id
es
f
lex
ib
le
ap
p
licatio
n
,
as
it
ca
n
b
e
ap
p
lied
at
th
e
n
etwo
r
k
lev
e
l
[
3
]
,
[
5
]
,
[
2
8
]
,
[
2
9
]
,
at
th
e
d
ev
ice
lev
el
[
3
0
]
,
[
3
1
]
,
o
r
with
in
m
o
n
ito
r
in
g
s
y
s
tem
s
[
3
2
]
.
T
h
e
d
o
wn
s
id
e
o
f
th
is
ap
p
r
o
ac
h
is
th
at
d
ata
is
o
f
ten
s
e
n
t
f
r
o
m
d
ev
ices
an
d
p
r
o
ce
s
s
ed
elsewh
er
e.
T
h
is
r
a
is
es
co
n
ce
r
n
s
ab
o
u
t
d
ata
p
r
i
v
ac
y
,
as
well
as
d
ata
s
ec
u
r
it
y
in
tr
an
s
m
is
s
io
n
.
Dep
en
d
in
g
o
n
th
e
in
d
u
s
tr
y
u
ti
lizin
g
th
e
s
o
lu
tio
n
,
t
h
er
e
m
ay
also
b
e
leg
alities
to
co
n
s
id
er
i
n
th
is
ar
ea
o
f
d
ata
s
af
ety
.
T
h
e
FL
ar
ch
itectu
r
e
a
p
p
r
o
ac
h
ad
d
r
ess
es
th
is
lim
i
tat
io
n
th
r
o
u
g
h
tr
ai
n
in
g
a
m
o
d
el
o
n
th
e
d
e
v
ice
its
elf
an
d
s
en
d
in
g
tr
ai
n
in
g
m
o
d
el
p
ar
am
eter
s
r
ath
er
th
an
th
e
r
aw
d
ata
to
a
ce
n
tr
al
lo
ca
tio
n
.
So
m
e
m
o
d
els
in
th
e
r
ev
iewe
d
p
ap
er
s
also
in
clu
d
i
n
g
d
ata
en
cr
y
p
tio
n
to
im
p
r
o
v
e
s
ec
u
r
ity
[
2
]
,
[
3
3
]
,
[
3
4
]
.
T
h
e
d
o
wn
s
id
e
o
f
th
is
ap
p
r
o
ac
h
,
h
o
wev
e
r
,
is
th
e
p
r
o
ce
s
s
in
g
p
o
wer
r
eq
u
ir
ed
t
o
co
m
p
lete
tr
ain
in
g
an
d
p
r
o
ce
s
s
in
g
o
n
th
e
d
ev
ice
its
elf
,
wh
ich
I
I
o
T
d
ev
ices m
ay
n
o
t
h
av
e
ca
p
ac
ity
f
o
r
.
3
.
2
.
2
.
T
ec
hn
iqu
e
s
I
t
was
clea
r
f
r
o
m
th
e
liter
atu
r
e
th
at
m
an
y
d
if
f
er
e
n
t
ML
tech
n
iq
u
es
ar
e
b
ein
g
u
tili
ze
d
in
p
r
o
p
o
s
ed
I
I
o
T
s
ec
u
r
ity
s
o
lu
tio
n
s
.
I
t
is
a
ls
o
ap
p
ar
e
n
t
f
r
o
m
t
h
e
liter
atu
r
e
th
at
at
t
h
is
s
tag
e
th
er
e
ar
e
n
o
ag
r
ee
d
u
p
o
n
b
est
m
eth
o
d
s
f
o
r
I
I
o
T
cy
b
er
s
ec
u
r
it
y
s
o
lu
tio
n
s
.
As
d
if
f
er
en
t
alg
o
r
ith
m
s
ca
n
b
e
co
m
b
in
e
d
in
d
i
f
f
er
en
t
way
s
,
th
er
e
ar
e
a
m
u
ltit
u
d
e
o
f
p
o
s
s
ib
ilit
ies
in
th
is
s
p
ac
e.
T
h
e
m
o
s
t
co
m
m
o
n
ly
p
r
o
p
o
s
ed
ML
tech
n
iq
u
es
f
ell
in
to
f
o
u
r
ca
teg
o
r
ies:
ANN,
s
u
p
er
v
is
e
d
lear
n
in
g
,
en
s
em
b
le
m
eth
o
d
s
an
d
s
em
i
-
s
u
p
er
v
is
ed
l
ea
r
n
in
g
.
Of
th
e
ANN
tech
n
iq
u
es,
th
e
m
o
s
t
co
m
m
o
n
ly
u
tili
ze
d
wer
e
C
NN
[
2
9
]
,
[
3
3
]
,
[
3
4
]
,
FC
NN
[
2
9
]
,
[
3
3
]
,
[
3
4
]
,
ML
P
[
5
]
,
[
3
3
]
,
a
n
d
L
STM
[
2
]
,
[
1
2
]
.
FC
NN
in
p
ar
ticu
lar
p
r
o
v
id
e
d
g
o
o
d
p
r
o
ce
s
s
in
g
tim
e
with
a
lo
w
lev
el
o
f
r
eso
u
r
ce
s
r
eq
u
ir
e
d
[
2
9
]
,
[
3
0
]
.
T
h
e
co
m
p
a
r
is
o
n
s
in
clu
d
ed
i
n
s
o
m
e
p
ap
er
s
p
r
o
v
id
e
d
in
s
ig
h
t
in
to
th
e
p
er
f
o
r
m
a
n
ce
o
f
s
p
ec
if
ic
ML
tech
n
iq
u
es,
p
ar
ticu
lar
ly
with
in
th
e
ANN
ca
teg
o
r
y
.
Ma
k
k
ar
et
a
l.
[
3
4
]
c
o
m
p
a
r
ed
m
o
d
els
u
s
in
g
C
NN,
L
STM
an
d
two
en
s
em
b
le
m
eth
o
d
s
,
a
n
d
f
o
u
n
d
th
at
th
e
C
NN
m
o
d
el
h
ad
th
e
b
est
p
er
f
o
r
m
a
n
ce
,
as
ca
p
tu
r
ed
in
T
ab
le
4
.
Ho
wev
er
,
th
ey
f
o
u
n
d
th
at
an
in
cr
ea
s
e
in
th
e
n
u
m
b
e
r
o
f
d
e
v
ices
co
r
r
elate
d
with
an
in
cr
ea
s
e
in
th
e
tim
e
tak
en
to
p
r
o
ce
s
s
d
ata
as
well
as
t
h
e
tim
e
tak
en
t
o
d
etec
t
attac
k
s
[
3
4
]
.
Sh
a
h
in
et
a
l.
[
2
9
]
c
o
m
p
ar
e
d
two
m
o
d
els
,
o
n
e
co
m
b
in
in
g
L
STM
wit
h
C
NN
an
d
th
e
o
th
e
r
co
m
b
in
in
g
L
STM
with
FC
NN.
As
d
is
p
lay
ed
in
T
a
b
le
4
,
th
ey
f
o
u
n
d
th
at
th
e
m
o
d
el
u
tili
zin
g
FC
NN
o
u
tp
er
f
o
r
m
ed
th
e
o
n
e
u
s
in
g
C
NN
ac
r
o
s
s
two
d
if
f
er
en
t
d
atasets
.
Fro
m
th
ese
co
m
p
a
r
is
o
n
s
,
it
ca
n
b
e
s
u
r
m
is
ed
t
h
at
in
ter
m
s
o
f
a
ttack
d
etec
tio
n
f
o
r
I
I
o
T
,
m
o
d
els
u
tili
zin
g
C
NN
tech
n
iq
u
es o
u
t
p
er
f
o
r
m
L
STM
m
o
d
els,
an
d
FC
NN
m
o
d
els o
u
tp
er
f
o
r
m
th
o
s
e
u
s
in
g
C
NN
tech
n
iq
u
es.
T
h
ese
co
m
p
ar
is
o
n
s
,
alo
n
g
wit
h
th
e
b
est
m
o
d
el
r
esu
lts
f
r
o
m
T
ab
le
4
,
ass
is
t
in
d
r
awin
g
c
o
n
clu
s
io
n
s
ab
o
u
t
wh
ich
tech
n
iq
u
es
s
tan
d
o
u
t
in
th
e
cu
r
r
e
n
t
liter
atu
r
e.
T
h
e
r
esu
lts
in
T
ab
le
4
s
h
o
w
th
at
th
e
ANN
tech
n
iq
u
e
m
o
s
t
co
m
m
o
n
l
y
u
s
ed
b
y
th
e
m
o
d
els
th
at
ac
h
ie
v
e
d
an
ac
cu
r
ac
y
a
b
o
v
e
9
9
%
was
ML
P.
T
h
e
r
ef
o
r
e
,
it
is
clea
r
th
at
FC
NN
an
d
ML
P a
r
e
p
ar
ticu
lar
ly
p
r
o
m
is
in
g
ANN
tech
n
iq
u
es f
o
r
I
I
o
T
cy
b
er
s
ec
u
r
ity
s
o
l
u
tio
n
s
.
Su
p
er
v
is
ed
lear
n
in
g
tech
n
iq
u
es
in
th
e
cu
r
r
en
t
liter
atu
r
e
c
an
b
e
ev
alu
ated
i
n
a
s
im
ilar
m
an
n
er
.
C
h
ak
r
ab
o
r
ty
et
a
l.
[
3
2
]
co
m
p
ar
ed
th
eir
m
o
d
el
u
s
in
g
d
if
f
er
en
t
s
u
p
er
v
is
ed
lear
n
i
n
g
tech
n
iq
u
es
f
o
r
class
if
icatio
n
,
f
in
d
in
g
th
at
R
F o
u
tp
er
f
o
r
m
ed
L
R
,
SVN
an
d
k
-
NN
ap
p
r
o
ac
h
es.
Similar
ly
,
T
r
an
et
a
l.
[
3
1
]
f
o
u
n
d
th
at
an
R
F
m
o
d
el
o
u
tp
er
f
o
r
m
e
d
m
o
d
els
u
s
in
g
DT
an
d
XGBo
o
s
t.
I
t
ca
n
b
e
s
ee
n
th
en
,
f
r
o
m
th
ese
co
m
p
ar
is
o
n
s
th
at
R
F is
th
e
p
o
p
u
lar
an
d
b
est
p
er
f
o
r
m
in
g
s
u
p
er
v
is
ed
lear
n
in
g
tech
n
i
q
u
e
with
in
th
e
cu
r
r
e
n
t liter
atu
r
e.
C
o
n
tr
ar
y
to
ML
P,
R
F
was
n
o
t
am
o
n
g
th
e
m
o
s
t
u
tili
ze
d
tech
n
iq
u
e
in
th
e
to
p
-
s
co
r
in
g
s
o
lu
ti
o
n
s
s
h
o
wn
in
T
ab
le
4
.
Ho
wev
er
,
it
d
id
f
e
atu
r
e
in
s
ev
er
al
o
f
th
e
m
o
d
els
th
at
p
er
f
o
r
m
e
d
th
e
b
est
in
th
eir
p
ar
ticu
lar
s
tu
d
y
[
3
1
]
,
[
3
2
]
,
[
3
4
]
.
T
h
is
s
u
p
p
o
r
ts
th
e
r
esu
lts
o
f
th
o
s
e
s
tu
d
ies
th
at
f
o
u
n
d
im
p
r
o
v
e
d
p
er
f
o
r
m
a
n
ce
b
y
in
cl
u
d
in
g
R
F
in
th
eir
m
o
d
els an
d
s
h
o
ws th
at
R
F is
a
p
r
o
m
is
in
g
ML
tech
n
i
q
u
e
f
o
r
I
I
o
T
cy
b
er
s
ec
u
r
ity
s
o
l
u
tio
n
s
.
G
r
ad
ien
t
b
o
o
s
tin
g
alg
o
r
ith
m
s
wer
e
th
e
m
ain
m
eth
o
d
s
u
tili
ze
d
in
th
e
ca
teg
o
r
y
o
f
en
s
em
b
le
m
eth
o
d
s
,
with
XGBo
o
s
t
b
ein
g
th
e
m
o
s
t
co
m
m
o
n
ly
u
s
ed
[
3
0
]
,
[
3
1
]
,
[
3
4
]
a
n
d
s
h
o
wn
t
o
b
e
a
g
o
o
d
tec
h
n
iq
u
e
f
o
r
n
etwo
r
k
in
tr
u
s
io
n
d
etec
tio
n
[
3
0
]
as
well
as
f
au
lt
d
etec
tio
n
[
3
1
]
.
Ho
wev
er
,
wh
en
Sh
ah
in
et
a
l.
[
3
0
]
co
m
p
a
r
ed
th
eir
Atten
tio
n
b
ased
L
STM
(
AL
STM
)
-
FC
NN
m
o
d
el
with
XG
B
o
o
s
t
an
d
Ad
aBo
o
s
t,
th
e
m
o
d
el
u
s
in
g
Ad
aBo
o
s
t
ac
tu
ally
p
r
o
v
id
ed
b
etter
p
er
f
o
r
m
an
ce
in
ter
m
s
o
f
p
r
ec
is
io
n
,
r
ec
all
an
d
F1
s
co
r
e.
Sem
i
-
s
u
p
er
v
is
ed
lea
r
n
in
g
tec
h
n
iq
u
es
wer
e
u
tili
ze
d
b
y
Ao
u
ed
i
et
a
l.
[
2
]
an
d
J
ian
g
[
4
5
]
i
n
d
if
f
er
en
t
way
s
.
T
h
is
ty
p
e
o
f
lear
n
in
g
was
w
ell
s
u
ited
to
d
ea
lin
g
with
lar
g
e
am
o
u
n
ts
o
f
u
n
lab
ele
d
d
a
ta
alo
n
g
with
s
o
m
e
lab
elled
d
ata
as
o
f
ten
e
x
is
ts
in
I
I
o
T
en
v
ir
o
n
m
en
ts
[
2
]
.
Ho
wev
er
,
I
I
o
T
e
n
v
ir
o
n
m
en
ts
m
a
y
also
h
a
v
e
am
o
u
n
ts
o
f
p
u
r
ely
u
n
lab
eled
d
ata,
w
h
ich
wo
u
ld
n
o
t
b
e
a
b
le
to
b
e
p
r
o
ce
s
s
ed
b
y
th
ese
ty
p
es
o
f
m
o
d
els
[
2
]
.
Desp
ite
th
is
lim
itatio
n
,
b
ein
g
a
b
le
to
p
r
o
c
ess
th
e
co
m
b
in
ed
la
b
elled
an
d
u
n
lab
ele
d
d
ata
s
u
g
g
ests
th
at
s
em
i
-
s
u
p
er
v
is
ed
le
ar
n
in
g
is
a
g
o
o
d
o
p
tio
n
f
o
r
I
I
o
T
cy
b
e
r
s
ec
u
r
ity
s
o
lu
tio
n
s
.
Mo
r
e
r
ec
en
t
p
a
p
er
s
h
av
e
lo
o
k
ed
at
r
ein
f
o
r
ce
m
en
t
lear
n
in
g
(
R
L
)
as
an
ad
ap
tiv
e
an
d
f
lex
ib
l
e
ML
to
o
l
in
cy
b
er
s
ec
u
r
ity
,
th
o
u
g
h
th
is
tech
n
iq
u
e
ca
n
also
h
a
v
e
a
h
ig
h
co
m
p
u
tatio
n
al
co
s
t
[
1
8
]
.
R
L
s
ee
k
s
to
tr
ain
an
ag
en
t
in
h
o
w
to
b
eh
av
e
in
i
ts
en
v
ir
o
n
m
e
n
t
in
a
wa
y
th
a
t
will
m
ax
im
ize
r
ewa
r
d
s
[
1
8
]
.
T
h
er
e
ar
e
m
a
n
y
Evaluation Warning : The document was created with Spire.PDF for Python.