T
E
L
K
O
M
NIKA
T
elec
o
mm
un
ica
t
io
n Co
m
pu
t
i
ng
E
lect
ro
nics
a
nd
Co
ntr
o
l
Vo
l.
23
,
No
.
5
,
Octo
b
er
20
25
,
p
p
.
1
212
~1
2
2
7
I
SS
N:
1
6
9
3
-
6
9
3
0
,
DOI
: 1
0
.
1
2
9
2
8
/
T
E
L
KOM
NI
K
A
.
v
23
i
5
.
27046
1212
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//jo
u
r
n
a
l.u
a
d
.
a
c.
id
/in
d
ex
.
p
h
p
/TELK
OM
N
I
K
A
DDo
S
a
tt
a
c
k det
e
ction using
opti
ma
l scrutiny bo
o
ste
d gra
ph
co
nv
o
lutiona
l and bidirec
tiona
l lon
g
sho
rt
-
ter
m
m
e
m
o
ry
H
ud
a
M
o
ha
m
m
e
d Ib
a
di,
A
s
g
ha
r
Asg
ha
ria
n Sa
rdro
ud
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
En
g
i
n
e
e
r
i
n
g
,
U
r
mi
a
U
n
i
v
e
r
si
t
y
,
U
r
mi
a
,
I
r
a
n
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Mar
3
,
2025
R
ev
i
s
ed
J
u
n
18
,
2025
A
cc
ep
ted
A
u
g
1
,
2025
T
h
e
d
istri
b
u
ted
d
e
n
ial
o
f
se
rv
ice
(DD
o
S
)
a
tt
a
c
k
o
c
c
u
rs
w
h
e
n
m
a
ss
iv
e
tra
ff
i
c
f
ro
m
n
u
m
e
ro
u
s
c
o
m
p
u
ters
is
d
irec
ted
to
a
se
rv
e
r
o
r
n
e
tw
o
rk
,
c
a
u
si
n
g
c
ra
sh
e
s
a
n
d
d
isr
u
p
t
in
g
f
u
n
c
ti
o
n
a
li
ty
.
S
u
c
h
a
tt
a
c
k
s
o
f
t
e
n
sh
u
t
d
o
w
n
w
e
b
sites
o
r
a
p
p
li
c
a
ti
o
n
s
tem
p
o
ra
ril
y
a
n
d
re
m
a
in
a
m
o
n
g
th
e
m
o
st
c
rit
ica
l
c
y
b
e
rse
c
u
rit
y
c
h
a
ll
e
n
g
e
s.
De
tec
ti
n
g
D
Do
S
is
d
iff
icu
lt
a
n
d
m
u
st
o
c
c
u
r
b
e
f
o
re
m
it
ig
a
ti
o
n
.
Re
c
e
n
tl
y
,
m
a
c
h
in
e
lea
rn
i
n
g
a
n
d
d
e
e
p
lea
rn
in
g
(M
L
/DL
)
h
a
v
e
b
e
e
n
e
m
p
lo
y
e
d
f
o
r
d
e
tec
ti
o
n
;
h
o
w
e
v
e
r,
a
r
c
h
it
e
c
tu
ra
l
li
m
it
a
ti
o
n
s
re
strict
th
e
ir
e
f
fe
c
ti
v
e
n
e
s
s
a
g
a
in
st
e
v
o
lv
in
g
a
tt
a
c
k
m
e
th
o
d
s.
T
h
is
p
a
p
e
r
p
re
se
n
ts
a
n
o
v
e
l
f
r
a
m
e
w
o
rk
,
sc
ru
ti
n
y
b
o
o
ste
d
g
ra
p
h
c
o
n
v
o
lu
ti
o
n
a
l
–
b
i
d
irec
ti
o
n
a
l
l
o
n
g
sh
o
rt
-
ter
m
m
e
m
o
r
y
a
n
d
v
isio
n
tran
sf
o
rm
e
r
(S
B
GC
-
BiL
S
T
M
-
V
iT
),
w
h
ich
in
teg
ra
tes
g
ra
p
h
c
o
n
v
o
lu
ti
o
n
a
l,
BiL
S
T
M
,
a
n
d
V
i
T
m
o
d
e
ls
w
it
h
m
a
c
h
in
e
lea
rn
in
g
c
las
si
f
iers
su
c
h
a
s
s
u
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
(S
V
M
)
,
Na
ïv
e
Ba
y
e
s
(NB),
ra
n
d
o
m
f
o
re
s
t
(RF
),
a
n
d
K
-
n
e
a
re
st
n
e
ig
h
b
o
rs
(
KN
N).
T
h
e
in
teg
ra
ti
o
n
e
n
a
b
les
a
u
to
n
o
m
o
u
s
e
x
trac
ti
o
n
o
f
c
rit
ica
l
f
e
a
tu
re
s,
e
n
h
a
n
c
in
g
p
re
c
isio
n
i
n
d
e
tec
ti
n
g
a
n
d
c
las
si
fy
in
g
DD
o
S
a
tt
a
c
k
s.
T
o
f
u
rth
e
r
b
o
o
st
p
e
rf
o
r
m
a
n
c
e
,
a
Ba
y
e
sia
n
o
p
ti
m
iza
ti
o
n
a
lg
o
rit
h
m
(BO
A
)
is
a
p
p
li
e
d
f
o
r
h
y
p
e
rp
a
ra
m
e
t
e
r
tu
n
i
n
g
o
f
S
BG
C
a
n
d
M
L
m
e
th
o
d
s.
Ev
a
lu
a
ti
o
n
o
n
b
e
n
c
h
m
a
rk
d
a
ta
se
ts
UN
S
W
-
N
B1
5
a
n
d
CICDD
o
S
2
0
1
9
d
e
m
o
n
stra
tes
th
a
t
th
e
p
r
o
p
o
se
d
a
p
p
r
o
a
c
h
a
c
h
iev
e
s
h
ig
h
e
r
a
c
c
u
ra
c
y
a
n
d
e
ffe
c
ti
v
e
l
y
id
e
n
ti
f
ies
n
e
w
DD
o
S
v
a
rian
ts,
o
u
tp
e
rf
o
r
m
in
g
c
o
n
v
e
n
ti
o
n
a
l
m
e
th
o
d
s
.
K
ey
w
o
r
d
s
:
A
r
ti
f
icial
i
n
tel
lig
e
n
ce
Dee
p
lear
n
in
g
Dis
tr
ib
u
ted
d
en
ial
o
f
s
er
v
ice
Ma
ch
i
n
e
lear
n
i
n
g
Un
k
n
o
w
n
attac
k
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
:
Asg
h
ar
A
s
g
h
ar
ia
n
Sar
d
r
o
u
d
Dep
ar
t
m
en
t o
f
C
o
m
p
u
ter
E
n
g
i
n
ee
r
in
g
,
Ur
m
ia
U
n
i
v
er
s
it
y
Ur
m
ia
5
7
5
6
1
5
1
8
1
8
,
I
r
an
E
m
ail:
a.
as
g
h
ar
ian
@
u
r
m
ia.
ac
.
ir
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
h
ar
m
o
n
io
u
s
n
atu
r
e
o
f
f
u
tu
r
e
b
ased
o
n
tech
n
o
lo
g
y
w
ith
th
e
o
n
-
g
o
in
g
tech
n
o
lo
g
ies
lik
e
ar
tif
icia
l
in
tellig
en
ce
,
j
o
in
ed
d
ev
ices
an
d
w
id
e
-
o
p
en
in
f
o
r
m
atio
n
d
ev
elo
p
m
en
t
u
n
leash
es
a
n
ew
im
ag
er
y
o
f
p
o
s
s
ib
ilit
y
.
T
h
e
s
er
v
ices
an
d
o
p
er
atio
n
s
p
r
o
v
id
ed
b
y
th
ese
co
n
n
ec
ted
d
ev
ices
co
n
tin
u
e
to
r
esh
ap
e
all
f
ac
ets
o
f
s
o
ciety
,
in
clu
d
in
g
p
eo
p
le
’
s
d
aily
liv
es,
h
ea
lth
ca
r
e,
tr
an
s
p
o
r
tatio
n
,
an
d
h
o
m
es.
I
n
ter
n
et
o
f
t
h
in
g
s
(
I
o
T
)
d
ev
ices
ar
e
p
r
o
j
ec
ted
to
attain
3
8
.
6
b
illi
o
n
b
y
2
0
2
5
an
d
ex
p
an
d
to
5
0
b
illi
o
n
b
y
2
0
3
0
.
Qu
ick
j
o
in
in
g
d
ev
ices
s
h
o
o
t
u
p
an
d
n
ew
m
o
d
els
ar
e
s
o
m
etim
es
lef
t
w
ith
o
u
t
u
p
d
ates,
s
ec
u
r
ity
an
d
p
atch
es.
T
h
e
w
id
esp
r
ea
d
ad
o
p
tio
n
o
f
s
u
c
h
tech
n
o
lo
g
ies
h
as
cr
ea
ted
o
p
p
o
r
tu
n
ities
f
o
r
m
alicio
u
s
ac
to
r
s
to
ex
p
lo
it
v
u
ln
er
ab
ilit
ies,
p
o
ten
tially
g
ain
in
g
co
n
tr
o
l
o
v
er
th
ese
d
ev
ices
to
lau
n
ch
attac
k
s
o
n
cr
itical
in
f
r
astru
ctu
r
e
an
d
w
eb
s
ites
[
1
]
.
I
n
r
ec
en
t
y
ea
r
s
,
cy
b
er
-
atta
c
k
s
h
av
e
em
er
g
ed
as
a
m
aj
o
r
is
s
u
e
d
u
e
to
th
eir
h
eig
h
ten
ed
s
o
p
h
is
ticatio
n
,
in
clu
d
in
g
d
en
ial
-
of
-
s
er
v
ice
(
Do
S)
m
eth
o
d
s
an
d
th
e
escalatin
g
th
r
ea
t
o
f
ze
r
o
-
d
ay
attac
k
s
tar
g
etin
g
in
d
u
s
tr
ial
n
etw
o
r
k
s
,
g
o
v
er
n
m
en
t
en
titi
es,
an
d
m
ilit
ar
y
o
p
er
atio
n
s
.
C
o
n
v
en
tio
n
al
s
ig
n
atu
r
e
-
b
ased
Diag
n
o
s
is
s
tr
ateg
ies
m
ay
n
o
t
b
e
s
u
f
f
icien
t
to
tack
le
th
e
v
ar
iab
ilit
y
in
cy
b
er
-
attac
k
s
s
ee
n
n
o
w
ad
ay
s
.
Z
h
ao
et
a
l.
[
2
]
n
ew
an
o
m
aly
d
iag
n
o
s
is
m
eth
o
d
s
ca
n
b
e
im
p
r
o
v
e
d
f
o
r
b
etter
lear
n
in
g
,
ad
ap
tin
g
an
d
also
d
iag
n
o
s
in
g
th
r
ea
ts
o
n
d
if
f
er
en
t
n
etw
o
r
k
ar
ea
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
DDo
S
a
tta
ck
d
etec
tio
n
u
s
in
g
o
p
tima
l scr
u
tin
y
b
o
o
s
ted
g
r
a
p
h
co
n
v
o
lu
tio
n
a
l …
(
Hu
d
a
Mo
h
a
mme
d
I
b
a
d
i
)
1213
I
n
a
n
e
w
s
i
m
p
li
f
ied
lear
n
in
g
-
k
n
o
w
led
g
e
d
is
co
v
er
y
a
n
d
d
ata
m
i
n
in
g
(
NS
L
-
KDD
)
v
er
s
io
n
,
ex
a
m
p
le
s
o
f
n
et
w
o
r
k
la
y
er
at
tack
s
ar
e
d
iv
id
ed
in
to
4
b
asic
c
lass
e
s
i
n
p
r
o
v
i
d
e
d
ata,
w
h
ic
h
is
r
ef
er
r
ed
to
as:
i
)
d
en
ial
o
f
s
er
v
ic
e
(
Do
S
)
,
w
h
e
n
th
e
attac
k
er
tr
ies
to
g
ain
ac
ce
s
s
to
th
e
h
o
s
ts
/
s
e
r
v
ice
lo
ck
ed
b
y
th
e
le
g
iti
m
ate
u
s
er
;
ii
)
pr
o
b
e,
th
e
attac
k
er
atte
m
p
t
s
to
o
b
tain
in
f
o
r
m
at
io
n
ab
o
u
t
th
e
tar
g
et
n
e
t
w
o
r
k
b
y
p
er
f
o
r
m
i
n
g
s
ca
n
n
i
n
g
p
r
o
g
r
am
s
th
at
i
n
s
tal
l
o
r
n
et
w
o
r
k
s
ca
n
s
;
iii
)
u
s
er
to
r
o
o
t
(
U2
R
)
,
th
e
attac
k
er
h
as
lev
er
ag
e
s
o
m
e
o
f
th
e
m
o
s
t
w
id
el
y
u
s
ed
attac
k
s
tr
ateg
ie
s
s
u
c
h
as
m
al
w
ar
e
in
f
ec
tio
n
an
d
s
to
len
cr
ed
en
tials
t
o
s
h
if
t
h
er
ac
ce
s
s
f
r
o
m
lo
w
p
r
iv
ile
g
e
to
s
u
p
er
/r
o
o
t
u
s
er
le
v
el
ac
ce
s
s
;
an
d
iv
)
r
e
m
o
te
to
lo
ca
l
(
R
2
L
)
,
th
e
at
tack
er
c
an
p
er
f
o
r
m
a
s
i
m
u
lat
io
n
o
f
lo
ca
l
u
s
er
s
a
n
d
ac
q
u
ir
e
ac
ce
s
s
to
th
e
tar
g
et
s
y
s
te
m
[
3
]
.
T
h
e
Do
S
attac
k
s
ar
e
t
h
e
s
p
ec
if
ic
attac
k
le
v
el
th
a
t
is
co
m
m
o
n
i
n
o
v
er
-
th
e
-
n
et
co
m
m
u
n
icati
n
g
b
et
w
ee
n
d
if
f
e
r
en
t
n
et
s
th
at
e
m
p
lo
y
in
ter
n
et
s
er
v
ices
f
o
r
u
tili
t
y
,
s
to
r
ag
e
a
n
d
p
r
o
ce
s
s
.
T
h
ese
attac
k
s
ar
e
ch
ar
ac
ter
ized
b
y
n
et
w
o
r
k
co
n
g
est
io
n
ca
u
s
ed
b
y
“
zo
m
b
i
e
p
ac
k
et
s
”
,
T
h
ese
ar
e
m
alicio
u
s
d
ata
u
n
it
s
th
at
tr
av
er
s
e
th
e
co
m
m
u
n
icati
o
n
la
y
er
s
,
g
en
er
ati
n
g
s
i
g
n
i
f
ica
n
t
d
is
r
u
p
tio
n
b
y
o
v
er
w
h
el
m
in
g
n
et
w
o
r
k
ch
a
n
n
e
l
s
an
d
co
n
ta
m
in
at
in
g
s
u
b
s
ta
n
tial
v
o
lu
m
es o
f
leg
i
ti
m
ate
tr
af
f
ic
d
u
r
in
g
tr
an
s
m
is
s
io
n
[
4
]
.
T
h
ese
attac
k
s
s
h
ar
e
s
im
ilar
ities
in
th
eir
ex
ec
u
tio
n
m
ec
h
an
is
m
s
,
w
ith
v
ar
iatio
n
s
p
r
im
ar
ily
o
b
s
er
v
ed
in
th
e
s
ca
le
an
d
in
ten
s
ity
o
f
th
e
ass
au
lt
.
I
n
a
Do
S a
ttack
a
s
in
g
le
s
y
s
tem
an
d
I
n
ter
n
et
co
n
n
ec
tio
n
attac
k
s
a
v
icti
m
.
I
n
co
n
tr
ast,
d
is
tr
ib
u
ted
d
en
ial
o
f
s
er
v
ice
(
DDo
S
)
attac
k
lev
er
ag
es
m
o
r
e
co
m
p
u
ter
s
as
w
ell
as
i
n
ter
n
e
t
co
n
n
ec
tio
n
s
to
s
atu
r
ate
th
e
v
ictim
;
it
is
o
f
ten
d
o
n
e
b
y
u
s
in
g
b
o
tn
ets,
w
h
ich
ar
e
n
etw
o
r
k
s
o
f
co
m
p
r
o
m
is
e
d
m
ac
h
in
es
[
5
]
.
Dep
en
d
in
g
o
n
w
h
ich
p
r
o
to
co
l
is
in
v
o
lv
ed
in
an
attac
k
th
er
e
ar
e
a
r
an
g
e
o
f
d
if
f
er
en
t
w
ay
s
to
ex
ec
u
te
an
y
o
f
s
u
ch
attac
k
s
.
A
s
id
en
tif
ied
b
y
Ma
h
j
ab
in
et
a
l.
[
6
]
s
u
ch
w
ay
s
as
u
s
er
d
atag
r
am
p
r
o
to
co
l
(
UDP
)
f
lo
o
d
ass
au
lt,
tr
an
s
m
is
s
io
n
co
n
tr
o
l
p
r
o
to
co
l
s
y
n
c
h
r
o
n
ize
s
eq
u
en
ce
n
u
m
b
er
s
(
T
C
P
SYN
)
,
an
d
h
y
p
er
tex
t tr
an
s
f
e
r
p
r
o
to
co
l
(
HT
T
P
)
f
lo
o
d
.
HT
T
P
f
lo
o
d
attac
k
s
f
ea
tu
r
e
s
tim
u
latin
g
r
eq
u
ir
em
en
ts
o
f
GE
T
/P
OST
s
u
p
p
lied
ac
r
o
s
s
HT
T
P
.
W
h
en
th
e
u
s
er
w
an
ts
to
r
etr
iev
e
d
ata
f
r
o
m
a
s
er
v
er
,
GE
T
is
u
s
ed
,
an
d
w
h
en
th
e
u
s
er
w
an
ts
to
s
en
d
d
ata
to
a
s
er
v
er
,
s
u
ch
as
u
p
lo
ad
in
g
a
f
ile,
P
OST
is
u
s
ed
.
T
r
an
s
f
er
r
in
g
th
o
u
s
an
d
s
o
f
r
eq
u
ests
to
a
s
er
v
er
o
r
clu
s
ter
w
ill
g
r
ea
tly
in
cr
ea
s
e
its
w
o
r
k
lo
ad
,
p
o
ten
tial
l
y
o
v
er
w
h
elm
i
n
g
th
e
s
er
v
er
an
d
ca
u
s
in
g
d
is
r
u
p
tio
n
s
.
T
h
is
r
en
d
er
s
th
e
s
er
v
er
s
u
n
r
ea
ch
ab
l
e
to
au
th
o
r
ized
u
s
er
s
.
B
y
s
en
d
in
g
a
SYN
p
ac
k
et,
g
ettin
g
a
s
y
n
ch
r
o
n
ize
-
ac
k
n
o
w
led
g
m
en
t
(
SYN
-
A
C
K
)
p
ac
k
et
f
r
o
m
th
e
s
er
v
er
,
an
d
co
m
p
letin
g
th
e
h
an
d
s
h
ak
e
w
ith
an
A
C
K
p
ac
k
et,
T
C
P
SYN
attac
k
ex
p
lo
its
th
e
th
r
ee
-
w
ay
h
an
d
s
h
ak
e
p
r
o
ce
d
u
r
e
in
a
T
C
P
co
n
n
ec
tio
n
.
T
h
is
k
in
d
o
f
attac
k
f
ak
es
SYN
p
ac
k
et
d
esti
n
atio
n
ad
d
r
ess
.
T
h
is
r
esu
lts
to
en
tr
ies
p
er
s
is
tin
g
in
th
e
co
n
n
ec
tio
n
d
atab
ase
o
f
th
e
s
er
v
er
an
d
SYN
-
A
C
K
p
ac
k
ets
b
ein
g
d
eliv
er
ed
co
n
tin
u
ally
at
th
e
s
p
o
o
f
ad
d
r
ess
.
A
s
th
e
item
s
p
ile
u
p
,
th
e
s
er
v
er
is
u
n
ab
le
to
p
r
o
ce
s
s
leg
itim
ate
r
eq
u
ests
an
y
m
o
r
e.
UDP
f
lo
o
d
attac
k
:
A
UDP
f
lo
o
d
attac
k
in
clu
d
es
s
en
d
in
g
an
o
v
er
w
h
elm
i
n
g
am
o
u
n
t
o
f
UDP
p
ac
k
ets
w
ith
p
h
o
n
y
in
ter
n
et
p
r
o
to
co
l
(
IP
)
ad
d
r
ess
es
an
d
ar
b
itra
r
y
p
o
r
t
n
u
m
b
er
s
.
W
h
en
th
ese
p
ac
k
ets
r
ea
ch
th
e
s
er
v
er
,
it
ch
ec
k
s
f
o
r
p
r
o
g
r
am
s
ass
o
ciate
d
w
ith
th
e
s
p
ec
if
ied
p
o
r
ts
.
I
f
n
o
th
in
g
m
atch
es,
a
“
d
esti
n
atio
n
u
n
r
ea
ch
ab
le
”
p
ac
k
et
is
r
etu
r
n
ed
b
y
th
e
s
er
v
er
as
a
r
esp
o
n
s
e.
Sin
ce
m
o
r
e
p
ac
k
ets
ar
r
iv
ed
th
an
th
e
s
er
v
er
co
u
ld
h
an
d
le,
th
e
s
er
v
er
co
u
ld
n
o
t
ass
is
t
au
th
o
r
ized
u
s
er
s
b
ec
au
s
e
o
f
to
o
m
u
ch
tr
af
f
ic.
Du
e
to
th
e
w
ea
k
s
ec
u
r
ity
ca
p
ab
ilit
ies
o
f
I
o
T
d
ev
ices,
s
ig
n
if
ic
a
n
t
r
esear
ch
h
as
em
er
g
ed
in
id
en
tif
y
in
g
an
d
m
itig
atin
g
Do
S
an
d
DDo
S
tr
af
f
ic.
B
u
t
m
ac
h
in
e
lear
n
in
g
(
ML
)
m
eth
o
d
s
ar
e
n
o
t u
s
ed
w
h
o
le
alg
o
r
ith
m
s
.
A
m
o
n
g
t
h
e
M
L
-
b
a
s
e
d
i
n
t
r
u
s
i
o
n
d
e
t
e
c
t
i
o
n
s
y
s
t
em
s
(
I
D
S
s
)
,
s
u
p
e
r
v
i
s
e
d
/
s
em
i
-
s
u
p
e
r
v
i
s
e
d
l
e
a
r
n
in
g
t
e
c
h
n
i
q
u
e
s
h
a
v
e
b
e
c
o
m
e
t
h
e
m
a
i
n
s
t
r
e
am
o
n
e
s
i
n
r
e
a
l
w
o
r
l
d
.
H
o
w
e
v
e
r
,
t
h
e
c
l
a
s
s
i
c
a
l
M
L
t
e
c
h
n
i
q
u
e
s
f
a
i
l
e
d
t
o
d
o
e
a
r
l
y
d
e
t
e
c
t
i
o
n
o
f
e
m
e
r
g
i
n
g
u
n
k
n
o
w
n
a
t
t
a
c
k
s
,
h
e
n
c
e
u
n
k
n
o
w
n
a
t
t
a
c
k
s
d
i
a
g
n
o
s
i
s
i
s
s
t
i
l
l
a
c
h
a
l
l
e
n
g
i
n
g
p
r
o
b
l
e
m
.
N
o
v
e
l
d
i
s
e
a
s
e
i
d
e
n
t
i
f
i
c
a
t
i
o
n
d
i
f
f
e
r
s
t
o
b
e
a
c
h
a
l
l
e
n
g
i
n
g
s
i
g
n
a
n
d
h
a
s
c
a
u
g
h
t
t
h
e
e
y
e
o
f
r
e
s
e
a
r
c
h
e
r
s
t
o
h
a
n
d
l
e
p
l
en
ty
o
f
c
h
a
l
l
e
n
g
e
t
o
f
o
r
m
u
l
a
t
e
v
a
r
i
o
u
s
a
p
p
r
o
a
c
h
e
s
f
r
o
m
v
a
r
i
o
u
s
s
e
c
t
o
r
s
t
o
g
e
t
h
e
r
w
i
t
h
t
h
e
I
D
S
a
n
d
f
a
c
e
d
e
t
e
c
t
i
o
n
.
I
n
s
u
c
h
f
i
e
l
d
s
,
t
h
e
d
a
t
a
q
u
i
c
k
l
y
t
o
g
g
l
e
s
.
D
a
i
l
y
,
n
e
w
u
n
f
a
m
il
i
a
r
a
t
t
a
c
k
s
a
n
d
i
n
n
o
v
a
t
i
v
e
f
a
m
i
l
i
a
r
a
t
t
a
c
k
v
a
r
i
a
n
t
s
a
r
e
e
m
e
r
g
i
n
g
.
I
n
o
r
d
e
r
t
o
b
e
a
b
l
e
t
o
c
o
n
t
i
n
u
e
t
o
d
i
a
g
n
o
s
e
b
a
d
t
r
a
f
f
i
c
,
I
D
S
m
u
s
t
b
e
a
b
l
e
t
o
a
d
a
p
t
t
o
s
w
i
t
c
h
in
g
a
r
e
a
s
.
T
h
e
s
tr
u
ctu
r
e
o
f
th
e
p
r
esen
t
p
ap
er
is
o
r
g
an
ized
as
f
o
llo
w
s
:
s
ec
tio
n
2
o
u
tlin
es
th
e
ex
p
lan
ato
r
y
co
n
tex
t
an
d
p
r
o
v
id
es
a
s
u
m
m
ar
y
o
f
th
e
r
elate
d
s
tu
d
y
.
Sectio
n
3
d
is
cu
s
s
es
th
e
co
n
ce
p
tu
al
f
o
u
n
d
atio
n
s
o
f
th
e
p
r
o
p
o
s
e
d
s
ch
em
es
an
d
p
r
esen
ts
an
o
v
er
v
iew
o
f
th
e
m
o
d
el
em
p
lo
y
ed
f
o
r
em
p
ir
ical
ev
alu
atio
n
,
in
clu
d
in
g
im
p
lem
en
ta
t
i
o
n
d
etails
.
Sectio
n
4
p
r
esen
ts
th
e
em
p
ir
ical
r
esu
lts
o
b
tain
ed
f
r
o
m
th
e
test
in
g
p
r
o
ce
d
u
r
es
,
an
d
co
m
p
ar
es
th
em
to
p
r
ev
io
u
s
p
ap
er
o
u
tco
m
es.
Fin
ally
,
s
ec
tio
n
5
p
r
o
v
id
es
th
e
co
m
p
ac
t
co
n
clu
s
io
n
o
f
th
e
p
ap
er
,
an
d
m
ap
s
th
e
u
n
i
t
f
o
r
tr
aj
ec
to
r
ies o
f
in
s
ig
h
t
p
ap
er
.
2.
RE
L
AT
E
D
WO
RK
T
h
e
in
cr
ea
s
i
n
g
co
m
p
le
x
it
y
a
n
d
f
r
eq
u
e
n
c
y
o
f
DDo
S
attac
k
s
h
a
v
e
p
r
o
m
p
ted
an
ex
ten
s
i
v
e
b
o
d
y
o
f
r
esear
ch
in
to
d
etec
tio
n
m
ec
h
a
n
is
m
s
,
p
ar
ticu
lar
l
y
w
it
h
i
n
th
e
co
n
tex
t
o
f
s
o
f
t
w
ar
e
d
ef
in
ed
n
et
w
o
r
k
i
n
g
(
SDN)
.
W
h
ile
n
u
m
er
o
u
s
ap
p
r
o
ac
h
es
h
av
e
e
m
er
g
ed
o
v
er
th
e
p
ast
f
i
v
e
y
ea
r
s
,
m
o
s
t
s
t
u
d
ies
f
a
ll
in
to
t
w
o
p
r
i
m
ar
y
ca
teg
o
r
ies:
t
h
o
s
e
e
m
p
lo
y
i
n
g
t
r
ad
itio
n
al
ML
tech
n
iq
u
es
a
n
d
th
o
s
e
le
v
er
ag
i
n
g
ad
v
a
n
ce
d
d
ee
p
lear
n
in
g
(
D
L
)
m
o
d
el
s
.
T
h
e
p
u
r
p
o
s
e
o
f
th
is
s
ec
tio
n
is
n
o
t
j
u
s
t
to
lis
t
p
r
ev
io
u
s
w
o
r
k
s
,
b
u
t
to
p
o
in
t
o
u
t
m
et
h
o
d
o
lo
g
ica
l
d
ef
icien
c
ies an
d
p
er
f
o
r
m
an
ce
li
m
ita
tio
n
s
in
th
e
s
e
ap
p
r
o
ac
h
es,
th
at
m
o
ti
v
ates t
h
e
d
esig
n
o
f
a
h
y
b
r
id
,
o
p
tim
ized
d
etec
tio
n
p
ar
ad
ig
m
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
23
,
No
.
5
,
Octo
b
e
r
20
25
:
1
2
1
2
-
1
227
1214
Sectio
n
2
.
1
p
r
esen
ts
t
h
e
r
elat
ed
w
o
r
k
s
t
h
at
u
s
e
M
L
clas
s
i
f
ier
lik
e
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
i
n
e
(
SVM
)
,
K
-
n
ea
r
est n
eig
h
b
o
r
s
(
KNN
)
a
n
d
d
ec
is
io
n
tr
ee
(
DT
)
,
to
d
etec
t
th
e
a
n
o
m
al
ies
i
n
t
h
e
tr
af
f
ic
o
f
SDN.
Ho
w
ev
er
,
i
n
g
en
er
al,
th
e
s
e
m
et
h
o
d
s
m
a
y
s
u
b
s
tan
tial
l
y
u
n
d
er
-
g
en
er
alize
to
n
e
w
o
r
c
h
an
g
i
n
g
t
h
r
ea
ts
a
s
t
h
e
y
m
a
y
h
a
v
e
(
i
n
t
h
e
w
o
r
s
t
ca
s
e)
n
o
ca
p
ac
it
y
to
ef
f
ec
tiv
el
y
lear
n
f
ea
t
u
r
es.
S
u
b
s
e
ctio
n
2
.
2
f
u
r
t
h
er
ex
ten
d
s
to
m
o
r
e
r
ec
en
t
ad
v
an
ce
s
b
ased
o
n
D
L
m
o
d
els
s
u
ch
as
co
n
v
o
lu
tio
n
al
n
e
u
r
al
n
et
w
o
r
k
s
(
C
NNs
)
,
r
ec
u
r
r
e
n
t
n
eu
r
al
n
et
w
o
r
k
s
(
R
N
Ns
)
,
an
d
lo
n
g
s
h
o
r
t
-
ter
m
m
e
m
o
r
y
s
(
L
S
T
Ms
)
f
o
r
en
h
an
ce
d
ad
ap
tiv
it
y
an
d
f
ea
t
u
r
e
ex
tr
ac
t
io
n
.
Ho
w
ev
er
,
th
e
y
o
f
ten
s
u
f
f
er
f
r
o
m
co
m
p
u
tatio
n
a
l in
e
f
f
icie
n
c
y
an
d
ar
e
n
o
t in
ter
p
r
etab
le
en
o
u
g
h
f
o
r
h
i
g
h
-
s
ta
k
es a
p
p
licati
o
n
s
.
Fro
m
th
i
s
o
v
er
v
ie
w
,
w
e
co
m
e
t
o
t
w
o
f
u
n
d
a
m
e
n
tal
o
b
s
er
v
atio
n
s
:
i
)
tr
ad
itio
n
al
M
L
m
e
th
o
d
s
d
o
n
o
t
h
av
e
d
ee
p
en
o
u
g
h
cu
tt
in
g
f
o
r
te
m
p
o
r
al
o
r
s
em
a
n
tic
u
n
d
er
s
ta
n
d
in
g
an
d
ii
)
ex
i
s
ti
n
g
D
L
m
et
h
o
d
s
en
co
u
n
ter
d
if
f
icu
lt
y
in
ter
m
s
o
f
o
p
ti
m
izatio
n
o
r
i
n
ter
p
r
etab
ilit
y
,
an
d
esp
ec
iall
y
in
o
n
lin
e
ap
p
licatio
n
s
.
T
o
b
r
i
d
g
e
th
is
d
i
v
id
e,
th
e
p
r
esen
t
s
tu
d
y
p
r
o
p
o
s
es
a
h
y
b
r
id
d
etec
tio
n
m
o
d
el
th
at
in
te
g
r
ates
s
cr
u
ti
n
y
b
o
o
s
ted
g
r
ap
h
co
n
v
o
l
u
tio
n
(
SB
GC
)
,
b
id
ir
ec
tio
n
al
L
ST
M
(
B
iL
ST
M)
,
an
d
v
is
io
n
tr
a
n
s
f
o
r
m
er
(
ViT
)
m
o
d
els
—
o
p
ti
m
ized
u
s
i
n
g
th
e
b
a
y
e
s
ian
o
p
tim
izatio
n
al
g
o
r
it
h
m
(
B
O
A
)
.
B
y
co
m
b
in
i
n
g
th
e
r
ep
r
esen
ta
tio
n
al
s
tr
en
g
t
h
o
f
DL
w
it
h
th
e
d
ec
is
io
n
ef
f
icie
n
c
y
o
f
ML
,
t
h
i
s
s
t
u
d
y
ai
m
s
to
d
ev
e
lo
p
a
r
o
b
u
s
t
an
d
ad
ap
ta
b
le
I
D
S
s
u
itab
le
f
o
r
m
o
d
er
n
SDN
-
d
r
iv
en
in
f
r
astr
u
ct
u
r
es.
2
.
1
.
T
ra
ditio
na
l
m
a
ch
ine le
a
rning
a
pp
ro
a
ches
f
o
r
DDo
S
det
ec
t
io
n
in SDN
env
iro
n
m
e
nts
Sek
ar
et
a
l
.
[
7
]
w
h
ich
f
o
cu
s
es
o
n
th
e
DDo
S
d
iag
n
o
s
is
an
d
ass
a
u
lt
p
r
ev
en
tio
n
d
o
m
ain
in
SDN
r
ea
lity
.
T
h
e
s
p
ec
ialized
to
p
ic
o
f
tech
n
o
lo
g
y
in
th
e
r
esear
ch
co
m
p
r
is
es
m
ac
h
in
e
lear
n
in
g
ap
p
r
o
ac
h
es,
s
p
ec
if
ical
l
y
d
ec
is
io
n
tr
ee
an
d
tech
n
iq
u
es
lik
e
ex
tr
em
ely
r
an
d
o
m
ized
tr
ee
s
(E
x
tr
a
T
r
ee
)
,
an
d
ca
teg
o
r
ical
b
o
o
s
tin
g
(C
at
B
o
o
s
t
)
in
d
o
m
ain
s
o
f
SDN
to
en
h
an
ce
n
etw
o
r
k
s
ec
u
r
ity
.
T
h
e
p
r
im
ar
y
ch
allen
g
e
u
n
d
er
r
ev
iew
in
th
is
w
o
r
k
,
en
tr
en
ch
e
d
in
SDN
to
DDo
S
attac
k
v
u
ln
er
ab
ilit
y
,
p
r
o
m
o
tes
m
o
r
e
r
o
b
u
s
t
d
etec
tio
n
an
d
m
itig
atio
n
s
o
lu
tio
n
s
cr
ea
tio
n
.
Desp
ite
th
e
im
p
lem
en
tatio
n
o
f
co
n
v
en
tio
n
al
s
ec
u
r
ity
m
ea
s
u
r
es,
th
eir
ad
eq
u
ac
y
in
ad
d
r
ess
in
g
th
e
in
cr
ea
s
i
n
g
l
y
co
m
p
lex
an
d
d
y
n
am
ic
n
atu
r
e
o
f
DDo
S a
ttack
s
w
ith
in
SDNs
r
em
ain
s
u
n
ce
r
tain
.
A
p
p
ly
in
g
m
ac
h
in
e
lear
n
in
g
tech
n
iq
u
es,
San
ap
ala
et
a
l.
[
8
]
p
r
o
p
o
s
es
a
u
n
iq
u
e
tech
n
iq
u
e
o
f
DDo
S
attac
k
s
d
etec
tio
n
an
d
m
itig
atio
n
in
th
e
SDN
ar
ch
itectu
r
e.
T
h
e
p
r
o
p
o
s
ed
s
y
s
tem
in
teg
r
ates
d
y
n
am
ic
p
r
o
ce
s
s
in
g
an
d
ev
alu
atio
n
o
f
r
ea
l
-
w
o
r
ld
n
etw
o
r
k
tr
af
f
ic
u
s
in
g
SVM
an
d
DT
class
if
ier
s
.
T
h
is
ap
p
r
o
ac
h
ac
h
iev
es
a
h
ig
h
ac
cu
r
ac
y
r
ate
in
id
en
tif
y
in
g
an
d
m
itig
atin
g
p
o
ten
tial
DDo
S
attac
k
s
.
I
t
u
s
es
th
e
m
ac
h
in
e
lear
n
in
g
ap
p
r
o
ac
h
to
d
etec
t a
n
d
m
itig
ate
DDo
S a
s
s
au
lts
in
SDNs
in
ter
m
s
o
f
n
etw
o
r
k
s
ec
u
r
ity
.
T
ah
ir
o
u
et
a
l.
[
9
]
,
w
h
o
s
tu
d
y
th
e
u
s
e
o
f
m
ac
h
in
e
lear
n
in
g
tech
n
iq
u
es.
T
h
is
r
esear
ch
em
p
h
asizes
th
e
s
ec
u
r
ity
ch
allen
g
es
o
f
SDN
f
r
am
ew
o
r
k
cr
ea
ted
b
y
th
e
s
ep
ar
atio
n
o
f
co
n
tr
o
l
an
d
d
ata
p
lan
e.
T
h
is
p
ap
er
’
s
ch
allen
g
e
in
d
icate
s
th
at
w
ell
-
tar
g
eted
,
ac
cu
r
ate
d
iag
n
o
s
is
an
d
im
p
r
o
v
ed
DDo
S
attac
k
m
itig
atio
n
o
f
SDN
ca
n
b
e
to
u
g
h
to
ad
d
r
ess
w
ith
ess
en
tiality
b
ein
g
s
tr
es
s
f
o
r
h
ig
h
ac
cu
r
ac
y
an
d
lo
w
f
alse
p
o
s
itiv
e
r
ates.
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
n
o
w
in
clu
d
es
an
aly
s
is
b
ased
o
n
tr
af
f
i
c
p
atter
n
,
en
tr
o
p
y
,
an
d
ML
b
ased
ap
p
r
o
ac
h
es.
T
h
e
class
if
i
er
s
u
s
ed
f
o
r
co
n
s
tr
u
ctin
g
a
class
if
icatio
n
m
o
d
el
f
o
r
th
e
p
r
o
p
o
s
ed
m
eth
o
d
co
m
p
r
is
es
KNN,
SVM,
an
d
Naiv
e
B
ay
es.
A
ls
u
d
an
i
an
d
Saee
a
[
1
0
]
b
eliev
e
th
at
ML
-
b
ase
d
s
o
lu
tio
n
s
m
ig
h
t
b
e
em
p
lo
y
ed
to
r
esear
ch
th
e
to
p
ic
o
f
DDo
S
attac
k
s
d
etec
tio
n
in
SDN
ar
ch
itectu
r
e.
T
h
e
ex
is
ti
n
g
liter
atu
r
e
in
th
is
p
ap
er
o
n
e
o
f
th
e
s
ec
u
r
ity
ch
allen
g
es o
f
in
ter
n
et
tech
n
o
lo
g
y
is
d
is
cu
s
s
ed
as
DDo
S
ass
au
lt
w
h
i
c
h
ad
d
r
ess
es
th
e
DDo
S
attac
k
o
n
SDN
ar
ch
itectu
r
e
o
r
r
eso
u
r
ce
allo
ca
tio
n
is
co
m
p
ar
ab
ly
less
an
d
co
v
er
s
k
e
y
im
p
licatio
n
s
f
o
r
f
u
tu
r
e
SDN
tech
n
o
lo
g
y
.
T
h
e
p
u
r
p
o
s
e
o
f
th
e
cu
r
r
en
t
r
esear
ch
is
to
u
n
d
er
tak
e
a
co
m
p
ar
ativ
e
p
er
f
o
r
m
an
ce
ev
alu
atio
n
to
d
is
co
v
er
th
e
b
est
-
p
er
f
o
r
m
in
g
m
ac
h
in
e
-
lear
n
i
n
g
ap
p
r
o
ac
h
es
f
o
r
DDo
S
attac
k
d
etec
tio
n
.
Ma
ch
in
e
lear
n
in
g
m
eth
o
d
s
s
u
ch
as
R
F,
lo
g
is
tic
r
eg
r
ess
io
n
(
LR
)
,
DT
,
KNN,
an
d
SVM
ar
e
u
s
ed
to
id
en
tif
y
an
d
d
etec
t
DDo
S
attac
k
s
.
T
h
e
p
u
r
p
o
s
e
is
to
d
eter
m
in
e
th
e
b
est
tech
n
iq
u
e
to
d
etec
t
DDo
S
ass
au
lts
in
SD
N
s
y
s
tem
s
.
Fen
g
et
a
l.
[
1
1
]
f
o
cu
s
o
n
in
co
r
p
o
r
atin
g
ML
p
ar
ticu
lar
ly
f
o
r
DDo
S a
ttack
d
etec
tio
n
.
T
h
e
c
e
n
t
r
a
l
i
z
e
d
c
o
n
t
r
o
l
p
l
a
n
e
o
f
S
D
N
m
a
k
e
s
i
t
i
n
h
e
r
e
n
t
ly
v
u
l
n
e
r
a
b
l
e
t
o
D
D
o
S
a
t
t
a
c
k
s
,
w
h
i
c
h
c
a
n
r
e
n
d
e
r
t
h
e
e
n
t
i
r
e
n
e
t
w
o
r
k
i
n
f
r
a
s
t
r
u
c
t
u
r
e
i
n
o
p
e
r
a
t
i
v
e
.
T
o
e
n
h
a
n
c
e
D
D
o
S
d
e
t
e
c
t
i
o
n
a
c
c
u
r
a
c
y
a
n
d
r
e
d
u
c
e
f
a
l
s
e
p
o
s
i
t
i
v
e
s
a
n
d
d
e
t
e
c
t
i
o
n
t
i
m
e
,
s
e
v
e
r
a
l
M
L
-
b
a
s
e
d
s
o
l
u
t
i
o
n
s
h
a
v
e
b
e
e
n
p
r
o
p
o
s
e
d
.
O
n
e
s
t
u
d
y
i
n
t
r
o
d
u
c
e
d
a
h
y
b
r
i
d
S
V
M
-
K
N
N
m
o
d
el
t
h
a
t
p
r
o
c
e
s
s
e
s
p
a
c
k
e
t
h
e
a
d
e
r
d
a
t
a
f
o
r
t
r
a
i
n
i
n
g
a
n
d
c
l
a
s
s
i
f
i
c
a
t
i
o
n
.
A
l
a
s
h
h
a
b
e
t
a
l
.
[
1
2
]
p
r
o
p
o
s
e
d
a
n
o
n
l
i
n
e
ML
-
b
a
s
e
d
L
D
D
o
S
d
e
t
e
c
t
i
o
n
a
p
p
r
o
a
c
h
u
s
i
n
g
e
x
p
l
a
i
n
a
b
l
e
b
o
o
s
t
i
n
g
m
a
c
h
i
n
e
(
E
B
M
)
w
i
t
h
s
t
o
c
h
a
s
t
i
c
g
r
a
d
i
e
n
t
d
e
s
c
e
n
t
(
S
G
D
)
,
o
u
t
p
e
r
f
o
r
m
i
n
g
t
r
a
d
i
t
i
o
n
a
l
m
e
t
h
o
d
s
l
i
k
e
S
V
M
,
NB
,
a
n
d
DT
.
A
l
m
o
h
a
g
r
i
e
t
a
l
.
[
1
3
]
e
x
p
l
o
r
e
d
S
DN
v
u
l
n
e
r
a
b
i
l
i
t
i
e
s
u
n
d
e
r
D
D
o
S
t
h
r
e
a
t
s
a
n
d
e
m
p
h
a
s
i
z
e
d
t
h
e
p
o
t
e
n
t
i
a
l
o
f
M
L
t
e
c
h
n
i
q
u
e
s
i
n
a
t
t
a
c
k
d
i
a
g
n
o
s
i
s
.
A
n
o
th
er
s
t
u
d
y
a
s
s
e
s
s
e
d
t
h
e
e
f
f
i
c
i
e
n
cy
o
f
D
T
,
e
x
t
r
e
m
e
g
r
a
d
i
e
n
t
b
o
o
s
t
i
n
g
(
X
G
B
o
o
s
t
)
,
a
n
d
r
a
n
d
o
m
f
o
r
e
s
t
(
R
F
)
i
n
S
D
N
-
b
a
s
ed
D
D
o
S
d
e
t
e
c
t
i
o
n
.
B
e
r
e
i
e
t
a
l
.
[
1
4
]
a
p
p
l
i
e
d
a
d
u
a
l
-
f
e
a
t
u
r
e
-
r
e
d
u
c
t
i
o
n
m
e
t
h
o
d
t
o
t
r
a
i
n
n
i
n
e
m
o
d
e
l
s
o
n
S
D
N
t
r
a
f
f
ic
a
n
d
v
a
l
i
d
a
t
e
d
t
h
e
i
r
p
e
r
f
o
r
m
a
n
c
e
u
s
i
n
g
Py
S
h
a
r
k
a
n
d
C
I
C
F
l
o
w
M
e
t
e
r
f
o
r
r
e
a
l
-
t
i
m
e
a
n
a
l
y
s
i
s
.
D
a
s
e
t
a
l
.
[
1
5
]
i
n
t
r
o
d
u
c
e
d
a
h
y
b
r
i
d
e
n
s
e
m
b
l
e
M
L
f
r
a
m
ew
o
r
k
c
o
m
b
i
n
i
n
g
s
u
p
e
r
v
i
s
e
d
(
f
o
r
k
n
o
w
n
a
t
t
a
c
k
s
)
a
n
d
u
n
s
u
p
e
r
v
i
s
e
d
(
f
o
r
z
e
r
o
-
d
a
y
a
t
t
a
c
k
s
)
c
l
a
s
s
i
f
i
e
r
s
u
s
i
n
g
o
u
t
l
i
e
r
d
e
t
e
c
t
i
o
n
.
A
l
m
o
m
an
i
e
t
a
l
.
[
1
6
]
b
e
n
c
h
m
a
r
k
e
d
s
e
v
e
r
a
l
M
L
c
l
a
s
s
i
f
i
er
s
,
i
n
c
l
u
d
i
n
g
X
G
B
o
o
s
t
,
G
a
u
s
s
i
a
n
N
B
,
D
T
,
R
F
,
S
V
M
,
a
n
d
L
R
,
f
o
r
D
o
S
d
e
t
e
c
t
i
o
n
i
n
I
o
T
e
n
v
i
r
o
n
m
e
n
t
s
.
A
i
r
l
a
n
g
g
a
[
1
7
]
c
o
n
d
u
c
t
e
d
a
c
o
m
p
a
r
a
t
i
v
e
a
n
a
l
y
s
i
s
o
f
R
F
,
D
T
,
a
n
d
L
R
m
o
d
e
l
s
w
i
t
h
s
y
n
t
h
e
t
i
c
m
i
n
o
r
i
ty
o
v
e
r
-
s
a
m
p
l
i
n
g
t
e
c
h
n
i
q
u
e
(
S
M
O
T
E
)
-
e
n
h
a
n
c
e
d
d
a
t
a
t
o
c
o
u
n
t
e
r
c
l
a
s
s
i
m
b
a
l
a
n
c
e
.
N
i
s
a
e
t
a
l
.
[
1
8
]
p
r
o
p
o
s
e
d
a
t
w
o
-
p
h
ase
a
u
t
h
e
n
t
i
c
a
t
i
o
n
m
e
t
h
o
d
b
a
s
e
d
o
n
M
L
-
d
r
i
v
e
n
p
a
c
k
e
t
f
i
l
t
e
r
i
n
g
t
o
i
m
p
r
o
v
e
S
D
N
s
e
c
u
r
i
t
y
w
i
t
h
o
u
t
d
i
s
r
u
p
t
i
n
g
h
o
s
t
o
p
e
r
a
t
i
o
n
s
.
T
a
b
l
e
1
s
u
m
m
a
r
iz
e
s
t
h
e
e
v
a
l
u
a
t
e
d
m
o
d
e
l
s
,
d
a
t
a
s
e
t
s
,
a
n
d
t
h
e
i
r
c
o
m
p
a
r
a
t
i
v
e
a
d
v
a
n
t
a
g
e
s
a
n
d
l
i
m
i
t
a
t
io
n
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
DDo
S
a
tta
ck
d
etec
tio
n
u
s
in
g
o
p
tima
l scr
u
tin
y
b
o
o
s
ted
g
r
a
p
h
co
n
v
o
lu
tio
n
a
l …
(
Hu
d
a
Mo
h
a
mme
d
I
b
a
d
i
)
1215
T
ab
le
1
.
Su
m
m
ar
y
o
f
M
L
ap
p
r
o
ac
h
es f
o
r
DDo
S d
etec
tio
n
R
e
f
Y
e
a
r
C
l
a
ssi
f
i
e
r
D
a
t
a
se
t
P
r
o
s
C
o
n
s
[
7
]
2
0
2
3
D
T
,
C
a
t
B
o
o
st
,
Ex
t
r
a
T
r
e
e
K
a
g
g
l
e
En
h
a
n
c
e
d
se
c
u
r
i
t
y
D
a
t
a
se
t
l
i
mi
t
a
t
i
o
n
s
[
8
]
2
0
2
3
S
V
M
,
D
T
N
e
t
w
o
r
k
t
r
a
f
f
i
c
d
a
t
a
S
c
a
l
a
b
l
e
,
e
f
f
e
c
t
i
v
e
,
r
e
a
l
-
t
i
me
d
e
t
e
c
t
i
o
n
F
e
a
t
u
r
e
se
l
e
c
t
i
o
n
,
p
o
t
e
n
t
i
a
l
d
a
t
a
o
v
e
r
f
i
t
t
i
n
g
[
9
]
2
0
2
3
K
N
N
,
S
V
M
,
G
a
u
ssi
a
n
N
a
ï
v
e
B
a
y
e
s
(
G
N
B
)
C
I
C
-
2
0
1
9
,
M
i
n
i
n
e
t
e
mu
l
a
t
o
r
Ef
f
e
c
t
i
v
e
D
D
o
S
d
e
t
e
c
t
i
o
n
a
n
d
mi
t
i
g
a
t
i
o
n
N
e
e
d
f
o
r
l
a
r
g
e
t
r
a
i
n
i
n
g
d
a
t
a
[
1
0
]
2
0
2
3
S
V
M
,
R
F
,
D
T
,
L
R
,
K
N
N
M
e
n
d
e
l
e
y
d
a
t
a
C
e
n
t
r
a
l
i
z
e
d
c
o
n
t
r
o
l
,
e
f
f
i
c
i
e
n
t
d
e
t
e
c
t
i
o
n
S
p
e
c
i
f
i
c
d
o
mai
n
a
p
p
l
i
c
a
b
i
l
i
t
y
[
1
1
]
2
0
2
3
S
V
M
,
K
N
N
M
i
n
i
n
e
t
e
mu
l
a
t
o
r
H
i
g
h
e
r
d
e
t
e
c
t
i
o
n
p
e
r
f
o
r
man
c
e
P
o
t
e
n
t
i
a
l
s
c
a
l
a
b
i
l
i
t
y
c
h
a
l
l
e
n
g
e
s
[
1
2
]
2
0
2
3
O
n
l
i
n
e
M
L
w
i
t
h
S
G
D
a
n
d
E
B
M
M
i
n
i
n
e
t
e
mu
l
a
t
o
r
R
e
a
l
t
i
me
t
i
me
p
r
o
c
e
ssi
n
g
,
i
n
t
e
r
p
r
e
t
a
b
i
l
i
t
y
S
i
mu
l
a
t
e
d
e
n
v
i
r
o
n
me
n
t
,
l
i
mi
t
e
d
e
v
a
l
u
a
t
i
o
n
[
1
3
]
2
0
2
3
X
G
B
o
o
st
,
R
F
,
D
T
C
I
C
D
D
o
S
2
0
1
9
En
h
a
n
c
e
d
n
e
t
w
o
r
k
se
c
u
r
i
t
y
L
i
mi
t
e
d
r
e
a
l
w
o
r
l
d
[
1
4
]
2
0
2
4
R
F
,
K
N
N
,
a
n
d
S
V
M
C
I
C
F
l
o
w
M
e
t
e
r
a
n
d
P
y
sh
a
r
k
I
mp
r
o
v
i
n
g
o
v
e
r
f
i
t
t
i
n
g
,
r
e
d
u
c
i
n
g
t
h
e
u
n
n
e
c
e
ssary
c
o
mp
l
e
x
i
t
y
o
f
t
h
e
i
n
i
t
i
a
l
mo
d
e
l
s,
a
n
d
d
e
c
r
e
a
si
n
g
b
o
t
h
t
r
a
i
n
i
n
g
a
n
d
p
r
e
d
i
c
t
i
o
n
t
i
me
s
D
a
t
a
se
t
l
i
mi
t
a
t
i
o
n
s
[
1
5
]
2
0
2
4
L
R
,
D
T
,
N
B
,
N
N
,
S
V
M
,
l
o
c
a
l
o
u
t
l
i
e
r
f
a
c
t
o
r
(
L
O
F
)
,
i
so
l
a
t
i
o
n
f
o
r
e
st
(
I
S
O
F
)
,
e
n
se
mb
l
e
l
e
a
r
n
i
n
g
(
EL
E
)
,
O
C
S
P
N
S
L
-
K
D
D
,
U
N
S
W
-
N
B
1
5
,
a
n
d
C
I
C
I
D
S
2
0
1
7
C
o
r
r
e
c
t
l
y
d
e
t
e
c
t
i
n
g
D
D
o
S
a
t
t
a
c
k
s
w
i
t
h
o
u
t
ma
n
y
f
a
l
se
a
l
a
r
ms
S
e
n
si
t
i
v
e
t
o
t
h
e
c
h
o
i
c
e
o
f
k
e
r
n
e
l
a
n
d
i
t
s
p
a
r
a
me
t
e
r
s
[
1
6
]
2
0
2
4
Ex
t
r
e
me
g
r
a
d
i
e
n
t
b
o
o
st
i
n
g
(
X
G
b
o
o
st
)
,
G
a
u
ssi
a
n
N
B
,
L
R
,
R
F
,
S
V
M
,
a
n
d
D
T
U
N
S
W
-
N
B
1
5
A
mo
n
g
t
h
e
c
l
a
ssi
f
i
e
r
s t
h
a
t
h
a
v
e
b
e
e
n
st
u
d
i
e
d
,
t
h
e
R
F
i
s
t
h
e
mo
st
e
f
f
e
c
t
i
v
e
i
n
d
e
t
e
c
t
i
n
g
D
o
S
a
t
t
a
c
k
s
A
ss
e
ssi
n
g
a
f
e
w
M
L
c
l
a
ssi
f
i
e
r
s u
s
i
n
g
p
a
r
t
i
c
u
l
a
r
a
t
t
a
c
k
s
[
1
7
]
2
0
2
4
L
R
,
R
F
,
a
n
d
D
T
K
a
g
g
l
e
I
mp
r
o
v
e
me
n
t
t
h
r
o
u
g
h
s
c
a
l
i
n
g
a
n
d
d
a
t
a
b
a
l
a
n
c
i
n
g
S
p
e
c
i
f
i
c
d
o
mai
n
a
p
p
l
i
c
a
b
i
l
i
t
y
[
1
8
]
2
0
2
4
S
V
M
,
K
N
N
C
I
C
D
o
S
2
0
1
7
R
e
d
u
c
t
i
o
n
o
f
f
a
l
se
p
o
si
t
i
v
e
s,
t
h
e
mi
n
i
mi
z
a
t
i
o
n
o
f
c
e
n
t
r
a
l
p
r
o
c
e
ssi
n
g
u
n
i
t
u
t
i
l
i
z
a
t
i
o
n
a
n
d
c
o
n
t
r
o
l
c
h
a
n
n
e
l
b
a
n
d
w
i
d
t
h
c
o
n
su
m
p
t
i
o
n
,
t
h
e
i
mp
r
o
v
e
me
n
t
o
f
p
a
c
k
e
t
d
e
l
i
v
e
r
y
r
a
t
i
o
,
a
n
d
t
h
e
d
e
c
r
e
a
se
i
n
t
h
e
n
u
mb
e
r
o
f
f
l
o
w
r
e
q
u
e
st
s
s
u
b
mi
t
t
e
d
t
o
t
h
e
c
o
n
t
r
o
l
l
e
r
S
e
n
si
t
i
v
i
t
y
t
o
t
h
e
c
h
o
i
c
e
o
f
k
i
n
K
N
N
,
S
V
M
d
o
e
s
n
o
t
p
e
r
f
o
r
m v
e
r
y
w
e
l
l
w
h
e
n
t
h
e
d
a
t
a
se
t
h
a
s
mo
r
e
n
o
i
se
2
.
2
.
Dee
p
lea
rning
-
ba
s
ed
t
ec
hn
iq
ues
f
o
r
DDo
S
det
ec
t
io
n
in SDN
co
nte
x
t
s
T
o
h
elp
in
id
en
tif
y
in
g
an
d
co
m
b
atin
g
DDo
S
attac
k
s
in
SDN
d
o
m
ain
s
,
A
ls
u
d
an
i
et
a
l.
[
1
9
]
p
r
o
v
id
e
th
e
ad
v
er
s
ar
ial
DB
N
-
L
ST
M
m
eth
o
d
em
p
lo
y
in
g
g
en
er
ativ
e
ad
v
er
s
ar
ial
n
etw
o
r
k
(
GA
N
)
,
d
ee
p
b
elief
n
etw
o
r
k
s
an
d
lo
n
g
s
h
o
r
t
-
ter
m
m
em
o
r
y
(
DB
N
-
L
ST
M)
to
m
in
im
ize
th
e
s
en
s
itiv
ity
o
f
th
e
s
y
s
tem
to
ad
v
er
s
ar
ial
attac
k
s
an
d
ex
p
ed
ite
f
ea
tu
r
e
ex
tr
ac
tio
n
.
Yu
n
g
aice
la
-
Nau
la
et
a
l.
[
2
0
]
p
r
esen
t
an
SDN
-
b
ased
m
eth
o
d
to
au
to
m
ate
d
etec
tio
n
an
d
m
itig
atio
n
o
f
s
lo
w
-
r
ate
DDo
S
attac
k
s
.
R
ein
f
o
r
ce
m
en
t
lear
n
in
g
(
R
L
)
is
ap
p
lie
d
in
th
e
f
r
am
ew
o
r
k
to
r
ed
u
ce
attac
k
s
,
w
h
i
l
e
d
ee
p
lear
n
in
g
(
DL
)
ass
is
ts
to
id
en
tif
y
th
em
.
Fu
r
th
er
m
o
r
e,
a
m
o
v
in
g
tar
g
et
d
ef
en
s
e
(
MT
D)
m
ec
h
an
is
m
b
ased
o
n
n
etw
o
r
k
f
u
n
ctio
n
v
ir
tu
aliza
tio
n
(
NFV)
is
p
r
esen
ted
to
im
p
r
o
v
e
th
e
ef
f
ec
tiv
en
ess
an
d
ad
ap
tab
ilit
y
o
f
th
e
s
y
s
tem
.
DL
tech
n
iq
u
es
th
at
Mo
u
s
a
an
d
A
b
d
u
llah
[
2
1
]
ar
e
u
tili
ze
d
f
o
r
d
iag
n
o
s
tics
o
f
DDo
S
attac
k
s
as
w
ell
as
ch
ec
k
p
o
in
t n
etw
o
r
k
s
,
a
f
au
lt
-
to
ler
an
t
m
eth
o
d
o
lo
g
y
f
o
r
ex
ten
d
ed
o
p
er
atio
n
s
.
A
li
et
a
l.
[
2
2
]
ex
p
lo
r
e
a
n
u
m
b
er
o
f
class
if
icatio
n
alg
o
r
ith
m
s
(
e.
g
.
,
SVMs,
KNNs,
DT
s
,
m
u
ltil
ay
er
p
er
ce
p
tr
o
n
s
(
ML
P
s
)
an
d
C
NNs).
T
h
e
ap
p
r
o
ac
h
p
r
esen
ted
b
y
Ma
n
s
o
o
r
et
a
l.
[
2
3
]
co
n
s
is
ts
o
f
th
r
ee
s
tep
s
:
d
ata
p
r
ep
r
o
ce
s
s
in
g
,
cr
o
s
s
-
f
ea
tu
r
e
s
elec
tio
n
(
w
h
ich
ai
m
s
to
f
in
d
ess
en
tial
f
ea
tu
r
es
f
o
r
DDo
S d
etec
tio
n
)
an
d
d
etec
tio
n
w
ith
th
e
u
s
e
o
f
th
e
R
NNs
m
o
d
el.
W
an
g
et
a
l.
[
2
4
]
p
r
esen
ted
s
ix
m
o
d
els:
DNN,
C
NN,
R
NN,
L
ST
M,
C
NN
+
R
NN,
C
NN
+
L
ST
M
to
id
en
tif
y
w
h
eth
er
n
etw
o
r
k
t
r
af
f
ic
w
as
a
m
alicio
u
s
attac
k
.
C
o
n
s
id
er
th
e
C
NN
an
d
B
iL
ST
M
u
tili
zin
g
an
atten
tio
n
tech
n
iq
u
e,
Said
an
d
A
s
lk
er
za
d
e
[
2
5
]
in
tr
u
s
io
n
d
iag
n
o
s
tic
m
u
ltip
le
m
o
d
els.
T
h
e
m
o
d
el
co
m
p
r
is
es
o
f
th
r
ee
p
r
im
ar
y
co
m
p
o
n
en
t
s
:
C
NN,
B
iL
ST
M,
an
d
atte
n
tio
n
lay
er
.
T
h
e
C
NN
lay
er
f
o
cu
s
es
o
n
v
ar
io
u
s
n
etw
o
r
k
tr
af
f
ic
d
ata
to
d
er
iv
e
lo
ca
l
f
ea
tu
r
es.
T
h
e
B
iL
ST
M
lay
er
lear
n
s
th
e
tem
p
o
r
al
r
elatio
n
s
h
ip
s
o
f
lo
ca
l
f
ea
tu
r
es.
T
h
e
atten
tio
n
lay
er
p
ick
s
u
p
th
e
m
o
s
t
r
elev
an
t
p
r
o
p
er
ties
f
r
o
m
th
e
r
esp
ec
tiv
e
B
iL
ST
M
r
esu
lt
f
o
r
ea
ch
ty
p
e
o
f
in
tr
u
s
io
n
f
o
r
DL
,
R
ao
an
d
Su
b
b
ar
ao
[
2
6
]
em
p
lo
y
th
e
SVM,
ML
P
,
an
d
L
ST
M
alg
o
r
ith
m
s
.
T
h
e
DL
m
o
d
el
th
at
is
p
r
o
p
o
s
ed
lear
n
s
an
d
g
en
er
ates
b
in
ar
y
an
d
m
u
lticlas
s
class
if
icatio
n
m
o
d
els
to
d
if
f
er
en
tiate
b
etw
ee
n
r
o
u
tin
e
tr
af
f
ic
an
d
n
etw
o
r
k
attac
k
f
u
n
ctio
n
s
m
o
d
els
an
d
d
atasets
ar
e
an
aly
ze
d
to
d
etec
t
an
o
m
alies
an
d
in
d
icato
r
s
o
f
p
o
ten
tial
attac
k
s
.
R
ig
o
r
o
u
s
an
d
p
r
ec
is
e
an
aly
s
is
is
co
n
d
u
cted
w
ith
in
th
e
d
ee
p
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
23
,
No
.
5
,
Octo
b
e
r
20
25
:
1
2
1
2
-
1
227
1216
lear
n
in
g
m
o
d
el.
Du
r
in
g
d
iag
n
o
s
tics
,
th
e
s
y
s
tem
d
if
f
er
en
tiates
b
etw
ee
n
m
alicio
u
s
ac
tiv
ity
an
d
n
o
r
m
al
n
etw
o
r
k
tr
af
f
ic.
A
d
d
itio
n
ally
,
tr
af
f
ic
p
atter
n
s
an
d
p
r
o
to
co
l
b
eh
av
io
r
ar
e
ex
am
in
ed
to
en
h
an
ce
d
etec
tio
n
ac
cu
r
ac
y
.
T
r
af
f
ic
f
ilter
in
g
w
as
th
en
d
o
n
e
to
r
em
o
v
e
d
u
b
io
u
s
tr
af
f
ic
o
r
attac
k
s
ig
n
atu
r
es.
Salih
an
d
A
b
d
u
lr
az
aq
im
p
r
o
v
ed
a
in
n
o
v
ativ
e
DL
f
r
am
ew
o
r
k
ca
lled
C
y
b
er
n
et
[
2
7
]
.
R
ec
o
g
n
izin
g
an
d
d
iag
n
o
s
in
g
th
e
s
o
r
ts
o
f
DDo
S
attac
k
s
ac
cu
r
ately
,
an
d
u
n
d
er
s
tan
d
in
g
s
o
m
e
o
f
th
e
m
o
s
t
f
u
n
d
am
en
tal
tactics
in
th
e
f
ield
o
f
cy
b
er
p
r
o
tectio
n
w
er
e
th
e
o
b
j
ec
tiv
es.
T
h
e
k
ey
ar
ch
itectu
r
al
co
m
p
o
n
en
t
im
p
lem
en
ted
is
th
at
th
e
o
u
tp
u
t
f
r
o
m
th
e
L
ST
M
an
d
th
e
1
D
-
C
N
N
ar
e
ad
d
ed
p
r
io
r
to
f
ee
d
in
g
th
e
h
id
d
en
o
u
tp
u
t
to
th
e
n
ex
t
lay
er
.
No
n
-
tim
e
s
er
ies
d
ata
is
also
ap
p
lied
in
p
r
o
d
u
cin
g
th
e
f
ea
tu
r
es
in
th
e
L
ST
M
lay
er
s
,
an
d
it
h
elp
s
th
e
m
o
d
el
to
lear
n
r
elev
an
t
attr
ib
u
tes
au
to
m
atica
lly
w
ith
th
e
in
clu
s
io
n
o
f
lay
er
s
s
u
ch
as
co
n
v
o
lu
tio
n
al
lay
er
s
.
B
en
m
o
h
am
ed
e
t
a
l.
[
2
8
]
p
r
o
p
o
s
ed
a
n
ew
tech
n
iq
u
e
ca
lled
E
n
co
d
er
-
Stack
ed
d
ee
p
n
eu
r
al
n
etw
o
r
k
s
th
at
lev
er
ag
e
Stack
ed
/b
ag
g
ed
ML
P
.
T
h
e
s
u
g
g
ested
s
o
lu
tio
n
em
p
l
o
y
s
an
en
co
d
er
to
id
en
tif
y
r
elev
an
t
elem
en
ts
f
r
o
m
a
p
r
o
ce
s
s
ed
d
ata
s
et
f
o
r
p
r
ec
is
e
attac
k
id
en
tif
icatio
n
,
th
e
co
m
p
ar
is
o
n
o
f
v
ar
io
u
s
d
ee
p
lear
n
in
g
-
b
ased
s
tr
ateg
ies
f
o
r
DDo
S
d
etec
tio
n
,
alo
n
g
w
ith
th
eir
r
esp
ec
tiv
e
class
if
ier
s
,
d
atasets
,
an
d
ev
alu
atio
n
asp
ec
ts
,
is
s
u
m
m
ar
ized
in
T
ab
le
2
.
T
ab
le
2
.
Su
m
m
ar
y
o
f
D
L
ap
p
r
o
ac
h
es f
o
r
DDo
S d
etec
tio
n
R
e
f
Y
e
a
r
C
l
a
ssi
f
i
e
r
D
a
t
a
se
t
P
r
o
s
C
o
n
s
[
1
9
]
2
0
2
3
G
A
N
a
n
d
D
B
N
-
L
S
T
M
C
I
C
D
D
o
S
2
0
1
9
Ef
f
e
c
t
i
v
e
D
D
o
S
d
e
t
e
c
t
i
o
n
R
e
a
l
-
t
i
me
p
e
r
f
o
r
man
c
e
,
r
e
p
r
o
d
u
c
i
b
i
l
i
t
y
[
2
0
]
2
0
2
3
L
S
T
M
a
n
d
R
L
C
I
C
D
o
S
2
0
1
7
Ef
f
i
c
i
e
n
t
d
e
t
e
c
t
i
o
n
,
mi
n
i
m
a
l
c
o
mp
l
e
x
i
t
y
V
a
l
i
d
a
t
i
t
h
e
o
n
e
o
f
t
h
e
p
r
o
p
o
se
d
f
r
a
me
w
o
r
k
[
2
1
]
2
0
2
3
L
S
T
M
a
n
d
C
N
N
M
e
n
d
e
l
e
y
D
a
t
a
U
N
S
W
_
2
0
1
8
_
I
o
T
_
B
o
t
n
e
t
S
i
n
g
l
e
d
a
t
a
se
t
u
se
d
R
e
a
l
-
t
i
me
t
e
st
i
n
g
,
d
i
v
e
r
se
d
a
t
a
se
t
s
[
2
2
]
2
0
2
3
S
V
M
,
K
N
N
,
D
T
,
M
L
P
,
C
N
C
I
D
S
2
0
1
7
a
n
d
C
I
C
D
D
o
S
2
0
1
9
En
h
a
n
c
e
d
d
e
t
e
c
t
i
o
n
,
sc
a
l
a
b
i
l
i
t
y
L
i
mi
t
e
d
r
e
a
l
-
t
i
me
a
sse
ssm
e
n
t
[
2
3
]
2
0
2
3
R
N
N
M
e
n
d
e
l
e
y
D
a
t
a
L
o
w
F
P
R
,
i
n
f
o
r
ma
t
i
v
e
f
e
a
t
u
r
e
s
L
i
mi
t
e
d
t
o
c
o
mm
u
n
i
c
a
t
i
o
n
c
h
a
n
n
e
l
s
[
2
4
]
2
0
2
3
C
N
N
+
R
N
N
,
D
N
N
,
C
N
N
,
R
N
N
,
L
S
T
M
,
a
n
d
C
N
N
+
L
S
T
M
C
S
E
-
C
I
C
-
I
D
S
2
0
1
8
En
h
a
n
c
e
d
d
e
t
e
c
t
i
n
g
c
a
p
a
b
i
l
i
t
i
e
s
o
v
e
r
c
u
r
r
e
n
t
sy
st
e
ms
A
d
d
i
t
i
o
n
a
l
v
a
l
i
d
a
t
i
o
n
i
s
r
e
q
u
i
r
e
d
i
n
v
a
r
i
o
u
s
n
e
t
w
o
r
k
c
o
n
t
e
x
t
s
[
2
5
]
2
0
2
3
C
N
N
-
B
i
L
S
T
M
I
n
S
D
N
En
h
a
n
c
e
d
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
R
e
l
i
a
n
c
e
o
n
sp
e
c
i
f
i
c
d
a
t
a
se
t
s
[
2
6
]
2
0
2
4
S
V
M
,
M
L
P
,
a
n
d
L
S
T
M
N
S
L
-
K
D
D
D
I
t
r
e
d
u
c
e
s t
h
e
i
m
p
a
c
t
o
f
D
o
S
/
D
D
o
S
a
t
t
a
c
k
s
o
n
n
e
t
w
o
r
k
s
[
2
7
]
2
0
2
4
L
S
T
M
,
C
N
N
C
I
C
-
D
D
o
S
2
0
1
9
su
p
e
r
i
o
r
d
e
t
e
c
t
i
o
n
a
n
d
r
e
c
o
g
n
i
t
i
o
n
c
a
p
a
b
i
l
i
t
i
e
s,
p
a
r
t
i
c
u
l
a
r
l
y
w
h
e
n
a
p
p
l
i
e
d
t
o
u
n
se
e
n
d
a
t
a
T
h
e
l
a
c
k
o
f
i
n
-
d
e
p
t
h
e
x
p
l
o
r
a
t
i
o
n
i
n
t
o
t
h
e
i
n
t
e
r
p
r
e
t
a
b
i
l
i
t
y
o
f
t
h
e
p
r
o
p
o
se
d
mo
d
e
l
s
[
2
8
]
2
0
2
4
S
t
a
c
k
e
d
d
e
e
p
n
e
u
r
a
l
n
e
t
w
o
r
k
s
(E
-
S
D
N
N
)
C
I
C
D
S
2
0
1
7
a
n
d
C
I
C
D
D
o
S
2
0
1
9
t
h
e
e
f
f
e
c
t
i
v
e
n
e
ss o
f
t
h
e
E
-
S
D
N
N
mo
d
e
l
t
h
e
mo
d
e
l
’
s c
o
mp
l
e
x
i
t
y
a
n
d
r
u
n
t
i
me
a
n
a
l
y
si
s n
e
c
e
ssi
t
a
t
e
a
d
d
i
t
i
o
n
a
l
t
r
a
i
n
i
n
g
t
i
me
3.
M
E
T
H
O
D
T
h
e
p
r
esen
t
p
ap
er
p
r
o
v
id
ed
th
e
o
p
tim
ized
s
im
ilar
ity
-
b
ased
g
r
ap
h
clu
s
ter
in
g
(
O
-
SB
GC
)
,
s
im
ilar
i
t
y
-
b
ased
g
r
ap
h
clu
s
ter
in
g
(
SB
GC
)
d
ev
elo
p
ed
b
y
th
e
b
u
tter
f
ly
o
p
tim
izatio
n
alg
o
r
ith
m
(
B
OA
)
f
o
r
tu
n
i
n
g
h
y
p
er
p
ar
am
eter
s
f
o
r
DDo
S a
ttack
d
iag
n
o
s
is
an
d
ea
r
ly
p
r
ev
en
tio
n
.
T
h
is
s
ec
tio
n
is
p
r
im
ar
ily
co
m
p
r
is
ed
o
f
eig
h
t
s
u
b
s
ec
tio
n
s
:
d
ataset
p
r
ep
ar
atio
n
,
p
r
ep
r
o
ce
s
s
in
g
,
s
p
litt
in
g
d
at
a,
b
alan
cin
g
d
ata,
B
OA
f
o
r
o
p
tim
izi
n
g
h
y
p
er
p
ar
am
eter
s
,
Featu
r
e
ex
tr
ac
tio
n
w
ith
O
-
SB
GC
-
B
iL
ST
M
an
d
ViT
,
ML
o
p
tim
izatio
n
w
ith
B
ay
esi
a
n
o
p
tim
izatio
n
alg
o
r
ith
m
(
SVM,
KNN,
NB
,
R
F),
an
d
E
v
alu
atio
n
o
f
m
o
d
els
b
ased
o
n
ac
cu
r
ac
y
,
r
ec
all
,
f
als
e
p
o
s
itiv
e
r
ate
(
FP
R
)
,
fa
ls
e
n
eg
ativ
e
r
ate
(
FNR
)
,
tr
u
e
n
eg
ativ
e
r
ate
(
T
NR
)
,
p
r
ec
is
io
n
,
F1
-
s
co
r
e,
r
ec
eiv
er
o
p
er
atin
g
ch
ar
ac
ter
is
tic
(
R
OC
)
,
ar
ea
u
n
d
er
th
e
cu
r
v
e
(
A
UC
)
.
T
h
e
o
v
er
all
w
o
r
k
f
lo
w
o
f
th
e
p
r
o
p
o
s
ed
DDo
S
d
etec
tio
n
s
y
s
tem
,
in
co
r
p
o
r
atin
g
O
-
SB
GC
-
B
iL
ST
M
an
d
ViT
m
o
d
els,
is
illu
s
tr
a
ted
in
Fig
u
r
e
1
.
3
.
1
.
B
a
y
esia
n o
ptim
iza
t
io
n a
lg
o
rit
h
m
B
a
y
e
s
i
a
n
o
p
t
i
m
i
z
a
t
i
o
n
r
e
f
e
r
s
t
o
t
h
e
m
e
t
h
o
d
o
f
o
p
t
i
m
i
z
a
t
i
o
n
g
i
v
e
n
t
h
e
M
L
t
h
a
t
i
s
n
o
r
m
a
l
l
y
b
e
i
n
g
a
p
p
l
i
e
d
f
o
r
o
p
t
i
m
i
z
i
n
g
o
b
je
c
t
i
v
e
b
l
a
c
k
-
b
o
x
t
a
s
k
s
.
T
h
e
m
e
c
h
a
n
i
s
m
o
f
B
a
y
e
s
i
a
n
o
p
t
i
m
i
z
a
t
i
o
n
a
p
p
l
i
e
s
p
r
o
b
a
b
i
l
i
s
t
i
c
m
o
d
el
i
n
t
e
g
r
a
t
i
o
n
t
o
h
e
l
p
o
p
t
i
m
i
z
a
t
i
o
n
p
r
o
c
e
s
s
e
s
a
n
d
t
e
c
h
n
i
q
u
e
s
t
o
m
o
d
e
l
a
n
d
s
a
m
p
l
e
B
a
y
e
s
i
a
n
n
e
t
w
o
r
k
s
.
T
h
e
m
e
c
h
a
n
i
s
m
p
r
e
s
e
n
t
s
i
n
s
t
a
n
c
e
s
o
l
u
t
i
o
n
s
p
o
p
u
l
a
t
i
o
n
a
p
p
l
y
i
n
g
B
a
y
e
s
i
a
n
n
e
t
w
o
r
k
s
.
B
a
y
e
s
i
a
n
o
p
t
i
m
i
z
a
t
i
o
n
t
e
c
h
n
iq
u
e
i
s
b
e
i
n
g
a
p
p
l
i
e
d
f
o
r
s
o
l
v
i
n
g
b
r
o
a
d
i
s
s
u
e
s
’
r
a
n
g
e
.
O
n
e
o
f
t
h
e
m
o
s
t
c
o
m
m
o
n
i
s
s
u
e
s
t
h
a
t
s
u
c
h
a
m
e
c
h
a
n
i
s
m
i
s
a
p
p
l
i
e
d
f
o
r
o
p
t
i
m
i
z
a
t
i
o
n
r
e
f
e
r
s
t
o
h
y
p
e
r
-
p
a
r
am
e
te
r
M
L
m
o
d
e
l
s
t
u
n
i
n
g
f
o
r
d
e
v
e
l
o
p
i
n
g
p
r
e
d
i
c
t
i
o
n
p
e
r
f
o
r
m
a
n
c
e
.
T
h
i
s
a
c
h
i
e
v
e
s
i
n
f
o
f
r
o
m
t
h
e
l
a
s
t
i
n
s
t
a
n
c
e
s
a
n
d
u
p
d
a
t
e
s
o
p
t
i
m
i
z
a
t
i
o
n
f
u
n
c
t
i
o
n
s
a
m
p
l
e
p
o
i
n
t
s
b
e
l
o
w
.
T
h
e
t
a
s
k
o
f
u
t
ilit
y
i
s
a
p
p
l
i
e
d
t
o
c
h
o
o
s
i
n
g
t
h
e
n
e
x
t
i
n
s
t
a
n
c
e
p
o
i
n
t
s
f
o
r
i
n
c
r
e
a
s
i
n
g
t
h
e
f
u
n
c
t
i
o
n
o
u
t
p
u
t
a
b
o
v
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
DDo
S
a
tta
ck
d
etec
tio
n
u
s
in
g
o
p
tima
l scr
u
tin
y
b
o
o
s
ted
g
r
a
p
h
co
n
v
o
lu
tio
n
a
l …
(
Hu
d
a
Mo
h
a
mme
d
I
b
a
d
i
)
1217
B
OA
as
th
e
m
eth
o
d
o
f
o
p
tim
izatio
n
s
tar
ts
w
ith
cr
ea
tin
g
a
r
an
d
o
m
b
asic
p
o
p
u
latio
n
o
f
p
r
o
m
is
i
n
g
s
o
lu
tio
n
s
k
n
o
w
n
as
s
tr
in
g
s
,
w
ith
a
u
n
iq
u
e
s
h
ar
e
o
v
er
w
h
o
le
f
ea
s
ib
le
s
tr
in
g
s
.
A
f
ter
th
at,
s
u
ch
p
o
p
u
latio
n
w
o
u
l
d
b
e
u
p
d
ated
o
v
er
s
ev
er
al
iter
atio
n
s
in
clu
d
in
g
f
o
u
r
s
tag
es.
Su
ch
f
o
u
r
b
asic
s
tag
es
ar
e
r
ep
ea
ted
till
p
r
o
m
is
i
n
g
ac
tu
al
v
ar
iab
les.
B
asically
,
b
y
p
er
f
o
r
m
in
g
th
e
m
eth
o
d
o
f
ch
o
ice
f
o
r
th
e
p
r
es
en
t
p
o
p
u
latio
n
,
s
atis
f
ac
to
r
y
r
esp
o
n
s
es
w
o
u
ld
b
e
s
elec
ted
.
Nex
t,
B
ay
esian
n
etw
o
r
k
is
m
ad
e
f
o
r
f
itti
n
g
s
atis
f
ac
to
r
y
p
r
o
m
is
in
g
s
o
lu
tio
n
p
o
p
u
latio
n
.
Fig
u
r
e
1
.
Flo
w
c
h
ar
t p
r
o
p
o
s
ed
m
et
h
o
d
3
.
2
.
F
e
a
t
ure
ex
t
ra
ct
io
n
T
h
e
p
r
o
p
o
s
ed
m
o
d
el
in
co
r
p
o
r
a
tes
a
d
u
al
-
s
tr
ea
m
f
ea
t
u
r
e
ex
tr
a
ctio
n
p
r
o
ce
s
s
,
co
m
b
in
i
n
g
th
e
c
ap
ab
ilit
ies
o
f
SB
GC
an
d
B
i
L
ST
M
to
ca
p
tu
r
e
s
p
atial
an
d
te
m
p
o
r
al
p
atter
n
s
f
r
o
m
n
et
w
o
r
k
tr
af
f
ic
d
ata.
T
h
is
p
r
o
ce
s
s
is
f
u
r
t
h
er
en
h
a
n
ce
d
th
r
o
u
g
h
in
te
g
r
atio
n
w
i
th
t
h
e
ViT
,
w
h
ic
h
c
o
n
tr
ib
u
tes
d
ee
p
atte
n
tio
n
-
b
ase
d
em
b
ed
d
in
g
s
to
t
h
e
o
v
er
all
r
ep
r
esen
tatio
n
.
T
h
e
in
itial
s
ta
g
e
b
eg
i
n
s
w
it
h
t
h
e
SB
GC
m
o
d
u
le,
w
h
ich
tr
an
s
f
o
r
m
s
th
e
r
a
w
i
n
p
u
t
f
ea
t
u
r
es
in
to
g
r
ap
h
-
s
tr
u
ct
u
r
ed
d
ata.
I
n
t
h
is
co
n
te
x
t
,
ea
ch
i
n
p
u
t
f
ea
t
u
r
e
v
ec
to
r
co
r
r
esp
o
n
d
s
to
a
n
o
d
e,
an
d
th
e
ed
g
es
b
et
w
ee
n
n
o
d
es
ar
e
d
y
n
a
m
icall
y
d
eter
m
i
n
ed
b
ased
o
n
s
i
m
ilar
it
y
m
etr
ics
o
r
co
r
r
elatio
n
co
ef
f
icie
n
ts
.
T
h
is
s
tr
u
ctu
r
e
allo
w
s
SB
G
C
to
ac
cu
m
u
late
an
d
tr
an
s
f
er
co
n
tex
t
b
et
w
ee
n
f
ea
t
u
r
es,
ca
p
tu
r
in
g
n
o
n
-
l
in
ea
r
s
p
atial
in
f
o
r
m
atio
n
t
h
at
is
n
o
t
p
r
eser
v
ed
in
f
lat
r
ep
r
ese
n
tatio
n
s
.
T
h
e
SB
GC
b
lo
ck
is
co
m
p
o
s
ed
o
f
an
in
p
u
t
p
r
o
j
ec
tio
n
lay
e
r
,
a
GC
N
b
lo
ck
,
an
d
a
d
r
o
p
o
u
t
m
ec
h
an
is
m
.
T
h
is
s
ch
e
m
e
e
n
ab
les
t
h
e
n
et
w
o
r
k
t
o
ca
p
tu
r
e
co
m
p
le
x
in
ter
-
f
ea
t
u
r
e
r
elatio
n
s
w
h
i
le
k
ee
p
in
g
th
e
g
e
n
er
alit
y
a
n
d
w
it
h
o
u
t o
v
er
f
itti
n
g
.
T
h
e
n
,
th
e
en
h
a
n
ce
d
f
ea
t
u
r
es
g
en
er
ated
b
y
SB
G
C
ar
e
f
u
s
e
d
w
it
h
f
r
a
m
e
-
d
if
f
er
e
n
ce
v
ec
to
r
s
Vti
th
a
t
r
ef
lect
t
h
e
te
m
p
o
r
al
d
if
f
er
en
ce
b
et
w
ee
n
ad
j
ac
en
t
f
r
a
m
e
s
.
S
u
c
h
d
i
f
f
er
e
n
ce
s
ar
e
cr
u
cial
f
o
r
d
i
s
tin
g
u
is
h
i
n
g
s
u
b
tle
b
eh
av
io
r
al
ch
a
n
g
es
o
f
th
e
tr
af
f
ic
t
h
at
r
ef
lect
DDo
Sa
ttac
k
s
.
T
h
e
s
p
atial
f
ea
t
u
r
e
f
t
ian
d
te
m
p
o
r
al
f
ea
tu
r
e
Vti
ar
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
23
,
No
.
5
,
Octo
b
e
r
20
25
:
1
2
1
2
-
1
227
1218
s
tack
ed
a
n
d
ad
o
p
ted
to
th
e
B
iL
ST
M
n
et
w
o
r
k
f
o
r
ti
m
e
-
s
eq
u
en
ce
m
o
d
eli
n
g
.
B
iL
ST
M
la
y
e
r
(
w
it
h
1
2
8
h
id
d
en
u
n
i
ts
f
o
r
b
o
th
d
ir
ec
tio
n
s
)
p
r
o
ce
s
s
es
th
e
co
n
ca
te
n
ated
i
n
p
u
t
s
i
g
n
al
s
a
n
d
m
o
d
els
t
h
e
b
id
i
r
ec
tio
n
al
te
m
p
o
r
al
d
ep
en
d
en
cies.
T
h
is
ar
ch
itect
u
r
e
is
f
av
o
r
ab
le
f
o
r
d
y
n
a
m
ics
in
tr
u
s
io
n
d
etec
tio
n
,
th
e
m
o
d
el
ca
n
tak
e
th
e
co
n
te
x
t
of
a
p
ar
ticu
lar
tr
af
f
ic
w
i
n
d
o
w
s
,
b
o
th
th
e
p
ast an
d
f
u
t
u
r
e
o
b
s
er
v
atio
n
s
i
n
to
ac
co
u
n
t.
Ma
th
em
atica
lly
,
th
e
B
iL
ST
M
p
r
o
ce
s
s
in
g
ca
n
b
e
ex
p
r
ess
ed
as:
=
(
(
,
)
)
=
(
(
,
(
−
1
)
)
)
(
1
)
T
h
e
ViT
p
r
o
ce
s
s
es
th
e
in
p
u
t
d
ata
in
a
s
i
m
i
lar
w
a
y
b
y
d
iv
id
i
n
g
it
i
n
to
f
i
x
ed
-
s
ize
p
atc
h
to
k
en
s
.
T
h
es
e
to
k
en
s
ar
e
em
b
ed
d
ed
an
d
en
r
ich
ed
w
it
h
p
o
s
itio
n
al
e
n
co
d
in
g
s
,
th
en
f
ed
to
a
s
tack
o
f
T
r
an
s
f
o
r
m
er
en
co
d
er
lay
er
s
.
T
h
e
atten
tio
n
m
ec
h
a
n
i
s
m
i
n
t
h
e
ViT
allo
w
s
f
o
r
th
e
m
o
d
el
to
lear
n
lo
n
g
-
r
a
n
g
e
d
ep
en
d
en
cie
s
an
d
to
ass
ig
n
t
h
e
v
ar
y
i
n
g
w
eig
h
t
s
to
d
if
f
er
en
t
p
a
r
ts
o
f
th
e
i
n
p
u
t.
T
h
e
o
u
tp
u
t
o
f
t
h
e
ViT
w
h
ich
i
s
a
h
i
g
h
-
d
i
m
e
n
s
io
n
al
f
ea
t
u
r
e
v
ec
to
r
ca
p
tu
r
in
g
b
o
th
g
lo
b
al
an
d
lo
ca
l in
f
o
r
m
atio
n
.
Af
ter
t
h
e
p
er
-
f
r
a
m
e
e
x
tr
ac
tio
n
,
th
e
SB
GC
-
B
iLST
M
an
d
ViT
p
ip
elin
es
ar
e
co
n
ca
ten
at
ed
u
s
i
n
g
a
f
ea
t
u
r
es
co
n
ca
te
n
atio
n
la
y
er
,
g
en
er
ati
n
g
a
n
i
n
te
g
r
ated
f
ea
t
u
r
e
v
ec
to
r
t
h
at
p
r
eser
v
es
s
p
a
tial,
te
m
p
o
r
al,
a
n
d
atten
tio
n
-
b
ased
co
n
tex
t.
T
h
is
f
u
s
io
n
g
u
ar
an
tees
t
h
at
t
h
e
m
o
d
el
ca
n
ta
k
e
ad
v
an
ta
g
e
o
f
th
e
co
m
p
le
m
e
n
tar
y
s
tr
en
g
th
s
f
r
o
m
g
r
ap
h
-
b
ased
m
o
d
elin
g
,
s
eq
u
e
n
tial le
ar
n
i
n
g
a
n
d
s
el
f
-
at
ten
t
io
n
m
ec
h
a
n
is
m
s
.
T
h
e
r
esu
lti
n
g
v
ec
to
r
is
t
h
e
n
p
r
o
v
id
ed
to
th
e
cla
s
s
i
f
ic
at
io
n
s
ta
g
e
w
h
er
e
w
ell
-
k
n
o
w
n
m
ac
h
in
e
lear
n
i
n
g
alg
o
r
ith
m
s
a
s
SVM,
KN
N,
NB
(
c
lass
if
icat
io
n
s
)
a
n
d
R
F
ar
e
e
m
p
lo
y
ed
.
T
h
ese
class
i
f
ier
s
o
p
er
ate
o
n
th
e
f
u
s
ed
r
ep
r
esen
tatio
n
to
ac
cu
r
atel
y
d
etec
t a
n
d
class
i
f
y
DDo
S a
ttac
k
in
s
ta
n
ce
s
w
ith
in
t
h
e
n
e
t
w
o
r
k
t
r
af
f
ic
d
ata
[2
9]
.
3
.
3
.
M
a
chine le
a
rning
o
pti
m
iza
t
io
n w
it
h
B
a
y
esia
n o
pti
m
i
za
t
io
n a
lg
o
ri
t
h
m
Me
ch
an
is
m
s
o
f
ML
ap
p
ly
th
e
ex
tr
ac
ted
attr
ib
u
tes
f
r
o
m
ev
er
y
lay
er
o
f
th
e
DL
m
o
d
el
’
s
f
ea
tu
r
e
f
o
r
m
ak
in
g
class
if
icatio
n
s
.
Fo
r
o
b
tain
in
g
th
e
h
ig
h
est
f
ea
s
ib
le
ac
cu
r
ac
y
in
class
if
icatio
n
,
th
is
is
ess
en
tial
f
o
r
r
ec
o
g
n
izin
g
ag
en
ts
th
at
im
p
ac
t
m
ec
h
an
is
m
s
’
ac
cu
r
ac
y
an
d
s
elec
tin
g
th
e
b
est
am
o
u
n
ts
g
iv
en
th
e
ch
o
s
e
n
attr
ib
u
tes.
Her
e,
s
o
m
e
p
ap
er
s
ex
is
t
ab
o
u
t
DL
m
o
d
els
an
d
ML
m
ec
h
an
is
m
s
’
p
ar
am
eter
o
p
tim
izatio
n
.
Her
e,
ev
er
y
ML
alg
o
r
ith
m
h
y
p
er
p
ar
am
eter
s
w
er
e
ch
o
s
en
v
ia
B
ay
esian
o
p
tim
izatio
n
an
d
s
tan
d
ar
d
eig
h
t
-
f
o
ld
cr
o
s
s
-
v
alid
ati
o
n
w
ay
in
th
e
p
r
o
ce
s
s
in
g
o
f
tr
ain
in
g
.
Var
io
u
s
ML
m
ec
h
an
is
m
s
h
av
e
u
n
if
o
r
m
h
y
p
er
p
ar
am
eter
s
th
at
co
u
ld
b
e
o
p
tim
ized
.
Fo
r
in
s
tan
ce
,
KNN
h
as
h
y
p
er
p
ar
am
eter
s
s
u
ch
as
d
i
s
tan
ce
w
eig
h
t,
m
etr
ic,
an
d
n
eig
h
b
o
r
s
’
n
u
m
b
e
r
;
SVM
h
as
h
y
p
er
p
ar
am
eter
s
s
u
ch
as
b
o
x
k
er
n
el
k
in
d
,
m
u
lticlas
s
m
eth
o
d
,
an
d
lim
itatio
n
lev
el;
R
F
h
as
h
y
p
er
p
ar
am
eter
s
s
u
ch
as
m
ax
s
p
lits
n
u
m
b
er
an
d
s
p
lit v
ar
iab
le;
NB
h
as h
y
p
er
p
ar
am
eter
s
s
u
ch
as
k
in
d
o
f
k
er
n
e
l
as w
e
ll
as n
am
e
o
f
s
h
ar
e.
NB
alg
o
r
ith
m
an
d
k
er
n
el
k
in
d
ap
p
lied
in
its
im
p
ac
t
h
o
w
ap
p
r
o
p
r
iate
p
r
ed
ictio
n
s
ar
e
f
o
u
r
o
p
tio
n
s
ex
is
t
:
T
r
ian
g
le,
Gau
s
s
ian
,
B
o
x
,
an
d
E
p
an
ec
h
n
ik
o
v
.
E
v
er
y
k
er
n
el
h
as
s
o
m
e
w
ea
k
n
ess
es
an
d
s
tr
en
g
th
s
.
GNB
m
o
d
el
s
ap
p
ly
Gau
s
s
ian
s
h
ar
e
f
o
r
co
m
p
u
tin
g
ev
er
y
lev
el
o
f
f
ea
s
ib
ilit
ies
w
h
en
k
er
n
el
Naiv
e
B
ay
es
m
o
d
els
ap
p
ly
s
elec
ted
k
er
n
el
s
h
ar
e.
T
h
e
SVM
alg
o
r
ith
m
h
as
th
r
ee
h
y
p
er
p
ar
am
eter
s
:
m
u
lticlas
s
tech
n
iq
u
e,
k
er
n
el
k
in
d
,
an
d
b
o
x
lim
itatio
n
class
.
Ker
n
el
k
in
d
d
ec
id
es
h
o
w
d
ata
is
tr
an
s
f
o
r
m
ed
b
ef
o
r
e
clas
s
if
icatio
n
an
d
co
u
ld
b
e
lin
ea
r
,
cu
b
ic,
q
u
ad
r
atic,
Gau
s
s
ian
r
ad
ial
b
asis
f
u
n
ctio
n
(
R
B
F).
KNN
alg
o
r
ith
m
h
as
3
b
asic
ad
j
u
s
tm
en
ts
f
o
r
b
ein
g
o
p
tim
ize
d
to
o
b
tain
th
e
b
est
o
u
tco
m
es:
d
is
tan
ce
w
eig
h
t
an
d
m
etr
ic,
an
d
n
eig
h
b
o
r
s
’
n
u
m
b
er
.
Neig
h
b
o
r
s
’
n
u
m
b
er
s
h
o
w
s
h
o
w
m
a
n
y
n
ea
r
b
y
n
eig
h
b
o
r
s
m
u
s
t
b
e
co
n
s
id
er
ed
to
g
r
o
u
p
ev
er
y
lo
ca
tio
n
.
T
h
e
m
etr
ic
o
f
d
is
tan
ce
co
m
p
u
t
e
s
d
is
tan
ce
am
o
n
g
d
ata
p
o
in
ts
,
v
ar
io
u
s
ac
ce
s
s
ib
le
o
p
tio
n
s
ex
is
t
s
u
ch
as
C
ity
b
lo
ck
,
E
u
clid
ea
n
,
an
d
s
o
o
n
.
R
an
d
o
m
f
o
r
est
r
ef
er
s
to
a
s
et
o
f
b
ig
d
ec
is
io
n
tr
ee
s
’
am
o
u
n
ts
.
E
v
er
y
tr
ee
is
b
u
ilt
b
y
s
elec
tin
g
k
r
an
d
o
m
f
ea
tu
r
es
an
d
ap
p
ly
in
g
a
tr
ain
in
g
s
et
o
f
d
ata.
C
h
o
o
s
e
n
d
ec
is
io
n
tr
ee
s
n
u
m
b
er
(
n
esti
m
ato
r
s
)
to
b
e
cr
ea
ted
in
th
e
f
o
r
est.
A
m
o
r
e
co
n
s
id
er
ab
le
tr
ee
n
u
m
b
er
n
o
r
m
ally
ca
u
s
es
th
e
d
ev
elo
p
ed
ac
cu
r
ac
y
h
o
w
ev
er
at
th
e
d
e
v
elo
p
ed
co
s
t
o
f
co
m
p
u
tatio
n
.
A
lg
o
r
ith
m
1
o
u
tlin
es
a
co
m
p
r
eh
en
s
iv
e
p
ip
elin
e
f
o
r
d
etec
tin
g
DDo
S
attac
k
s
u
s
in
g
th
e
O
-
SB
G
C
-
B
iL
ST
M
-
ML
f
r
am
ew
o
r
k
.
I
n
itially
,
th
e
n
ec
ess
ar
y
co
m
p
o
n
en
ts
,
in
clu
d
in
g
m
o
d
el
h
y
p
er
p
ar
am
eter
s
an
d
tr
ain
i
n
g
p
ar
am
eter
s
,
ar
e
co
n
f
ig
u
r
ed
.
P
r
ep
r
o
ce
s
s
in
g
b
eg
in
s
w
ith
lo
ad
in
g
th
e
DDo
S
d
atasets
,
f
o
llo
w
ed
b
y
s
p
litt
in
g
th
e
m
in
to
tr
ain
in
g
,
v
alid
atio
n
,
an
d
test
s
u
b
s
ets.
T
h
e
B
OA
is
th
en
em
p
lo
y
ed
to
o
p
tim
ize
th
e
h
y
p
er
p
ar
am
eter
s
o
f
th
e
O
-
SB
GC
-
B
iL
ST
M
m
o
d
el,
en
s
u
r
in
g
im
p
r
o
v
ed
ac
cu
r
ac
y
.
T
h
is
o
p
tim
ized
m
o
d
el
ex
t
r
ac
ts
h
ig
h
-
lev
el
tem
p
o
r
a
l
f
ea
tu
r
es
f
r
o
m
th
e
n
etw
o
r
k
tr
af
f
ic
d
ata,
w
h
ich
ar
e
s
u
b
s
eq
u
en
tly
r
ef
in
ed
an
d
en
co
d
ed
b
y
a
ViT
th
r
o
u
g
h
atten
tio
n
-
b
ased
f
ea
tu
r
e
en
h
an
ce
m
en
t.
T
h
e
f
ea
tu
r
es
f
r
o
m
b
o
th
m
o
d
els
ar
e
co
n
ca
ten
ated
to
f
o
r
m
a
h
y
b
r
id
r
ep
r
esen
tatio
n
.
I
n
th
e
class
i
f
icatio
n
p
h
ase,
tr
ad
itio
n
al
ML
class
if
ier
s
—
SVM,
NB
,
KNN,
an
d
RF
—
ar
e
ap
p
lied
to
th
e
h
y
b
r
i
d
f
ea
tu
r
e
s
et.
T
h
e
m
o
d
el
is
tr
ain
ed
,
v
alid
ated
,
an
d
th
en
test
ed
,
w
ith
p
er
f
o
r
m
an
ce
ev
alu
ated
u
s
in
g
m
etr
ics
lik
e
p
r
ec
is
io
n
,
r
ec
all,
F1
-
s
co
r
e,
an
d
ac
cu
r
ac
y
.
T
h
r
esh
o
ld
-
b
ase
d
b
in
ar
y
d
ec
is
io
n
s
ar
e
m
ad
e
d
u
r
in
g
test
in
g
to
class
i
f
y
tr
af
f
ic
as
b
en
ig
n
o
r
DDo
S.
Fin
ally
,
R
OC
an
d
A
UC
m
etr
ics
ar
e
co
m
p
u
ted
to
ass
ess
th
e
m
o
d
el
’
s
o
v
er
all
d
etec
tio
n
ca
p
ab
ilit
y
,
an
d
s
u
s
p
icio
u
s
s
am
p
les
ar
e
lo
g
g
ed
f
o
r
f
u
r
th
er
an
aly
s
is
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
DDo
S
a
tta
ck
d
etec
tio
n
u
s
in
g
o
p
tima
l scr
u
tin
y
b
o
o
s
ted
g
r
a
p
h
co
n
v
o
lu
tio
n
a
l …
(
Hu
d
a
Mo
h
a
mme
d
I
b
a
d
i
)
1219
Algorithm 1.
Algorithm for DDoS detection and classification using O
-
SBGC
-
BiLSTM
-
ML model
Input
:
Dataset of DDoS with corresponding labels
Training parameters
Test dataset for evaluation.
Output
: DD
oS detection and classification Evaluation metrics (accuracy, preci
sion,
and
recall)
1: Load and preprocess the dataset.
2: Split the dataset into train, validation, and test sets.
3:
Bayesian optimization algorithm
i. Implement BOA for optimizing hyperparameters in SBGC
-
BiLSTM.
ii. Implement BOA for optimizing
hyperparameters in ML algorithms.
4: O
-
SBGC
-
BiLSTM Feature Extraction:
i. Implement an O
-
SBGC
-
BiLSTM for feature extraction.
ii. Train the O
-
SBGC
-
BiLSTM using the training dataset.
iii. Extract high
-
level features using the trained O
-
SBGC
-
BiLSTM.
5: Visio
n Transformer (ViT) Encoding:
i. Implement a ViT to extract features.
ii. Fine
-
tune the ViT using the training dataset.
iii. Encode the extracted features with the trained ViT.
6: Concatenate the extracted features of O
-
SBGC
-
BiLSTM and ViT.
7: Impleme
nt ML algorithms as classification heads (SVM, NB, KNN, RF).
8: Train the classification head using the fused features from step 5.
9: Model Evaluation:
i. Evaluate the O
-
SBGC
-
BiLSTM
-
ML model using the validation set.
ii. Fine
-
tune hyperparameters if nec
essary.
10: Model Testing:
i. Test the trained model on the testing dataset.
ii. Compute classification metrics (e.g., Accuracy, Recall, FPR, FNR, TNR,
Precision, F1
-
s
core, ROC, AUC).
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
I
n
th
is
s
ec
tio
n
,
w
e
ev
alu
ate
th
e
p
r
o
p
o
s
ed
I
DS
ex
p
er
im
en
tally
.
T
h
e
p
er
f
o
r
m
an
ce
an
d
ef
f
ec
tiv
en
ess
o
f
th
e
s
y
s
tem
w
er
e
ev
alu
ated
o
n
tw
o
b
en
ch
m
ar
k
d
atasets
:
UNSW
-
NB
1
5
an
d
C
I
C
I
DS2
0
1
9
.
Dete
ctio
n
ac
cu
r
ac
y
an
d
r
o
b
u
s
tn
ess
w
er
e
m
ea
s
u
r
ed
u
s
in
g
s
tan
d
ar
d
ev
alu
atio
n
m
etr
ics.
A
ll
ex
p
er
im
en
ts
w
er
e
im
p
lem
e
n
ted
in
P
y
th
o
n
,
o
n
a
m
ac
h
in
e
w
ith
an
I
n
tel
C
o
r
e
i7
-
7
7
0
0
p
r
o
ce
s
s
o
r
an
d
3
2
GB
o
f
R
A
M.
4
.
1
.
Co
llect
i
o
n a
nd
cha
ra
ct
e
ristics o
f
da
t
a
s
et
I
n
o
r
d
er
to
ca
r
ef
u
lly
r
ev
iew
p
er
f
o
r
m
an
ce
an
d
r
o
b
u
s
tn
ess
o
f
th
e
p
r
o
p
o
s
ed
DDo
S
d
etec
tio
n
m
o
d
el
.
T
w
o
p
u
b
licly
ac
ce
s
s
ib
le
b
en
ch
m
ar
k
d
atasets
w
er
e
u
s
ed
:
UNSW
-
NB
1
5
an
d
C
I
C
DDo
S2
0
1
9
.
T
h
ese
ar
e
r
ea
lis
tic
n
etw
o
r
k
-
w
id
e
DDo
S
attac
k
d
atasets
co
n
tain
in
g
leg
itim
ate
an
d
attac
k
tr
af
f
ic
f
o
r
tr
ain
in
g
an
d
v
alid
atio
n
o
f
in
tr
u
s
io
n
d
etec
tio
n
m
o
d
el,
esp
ec
ially
in
SDN
an
d
an
ad
v
an
ce
d
DDo
S a
ttack
ca
s
e.
4
.
1
.
1
.
UNSW
-
NB
1
5
d
a
t
a
s
et
T
h
e
UNSW
-
NB
1
5
d
ataset
w
as
o
r
ig
i
n
all
y
co
n
s
tr
u
cted
b
y
th
e
au
s
tr
alia
n
ce
n
tr
e
f
o
r
c
y
b
er
s
ec
u
r
it
y
(
AC
C
S)
at
th
e
Un
i
v
er
s
i
t
y
o
f
Ne
w
So
u
th
W
ales.
T
h
e
r
aw
n
et
w
o
r
k
p
ac
k
ets
w
er
e
g
e
n
er
ated
u
s
in
g
th
e
I
XI
A
P
er
f
ec
tSt
o
r
m
to
o
l,
w
h
ic
h
s
i
m
u
lates
a
h
y
b
r
id
e
n
v
ir
o
n
m
e
n
t
o
f
n
o
r
m
al
a
n
d
m
alicio
u
s
ac
ti
v
itie
s
to
m
i
m
ic
r
ea
l
-
w
o
r
ld
tr
af
f
ic
co
n
d
itio
n
s
.
T
h
e
d
ataset
in
c
lu
d
es
b
o
t
h
le
g
iti
m
a
te
tr
af
f
ic
a
n
d
n
i
n
e
ca
te
g
o
r
ies
o
f
attac
k
s
:
Fu
zz
er
s
,
An
al
y
s
i
s
,
B
ac
k
d
o
o
r
s
,
Do
S,
E
x
p
lo
its
,
Gen
er
ic,
R
ec
o
n
n
ais
s
a
n
ce
,
Sh
ellco
d
e,
an
d
W
o
r
m
s
[
30
]
.
T
h
e
Kag
g
le
-
h
o
s
ted
v
er
s
io
n
o
f
th
e
d
atase
t
[3
1
]
p
r
o
v
id
es
th
e
p
r
ep
r
o
ce
s
s
ed
an
d
lab
eled
d
ata
in
C
SV
fo
r
m
at,
co
n
s
is
tin
g
o
f
1
7
5
,
3
4
1
s
a
m
p
les
in
th
e
tr
ain
in
g
s
e
t
an
d
8
2
,
3
3
2
s
a
m
p
les
i
n
t
h
e
te
s
ti
n
g
s
et.
E
ac
h
in
s
ta
n
ce
in
t
h
e
d
atase
t
is
ch
ar
ac
ter
ized
b
y
a
s
et
o
f
4
9
f
ea
t
u
r
es
d
er
iv
e
d
f
r
o
m
r
a
w
p
ac
k
et
ca
p
t
u
r
es,
e
n
co
m
p
as
s
i
n
g
f
lo
w
-
b
ased
,
co
n
ten
t
-
b
ased
,
an
d
ti
m
e
-
b
ased
at
tr
ib
u
tes.
T
h
e
d
ataset
’
s
clas
s
d
is
tr
ib
u
t
io
n
r
ef
lect
s
r
ea
l
-
w
o
r
ld
i
m
b
alan
ce
,
w
it
h
a
h
i
g
h
p
r
o
p
o
r
tio
n
o
f
n
o
r
m
al
tr
a
f
f
ic,
t
h
u
s
o
f
f
er
i
n
g
an
ap
p
r
o
p
r
iate
ch
allen
g
e
f
o
r
in
tr
u
s
i
o
n
d
etec
tio
n
s
y
s
te
m
s
to
ac
h
iev
e
h
ig
h
d
etec
tio
n
ac
c
u
r
ac
y
w
i
th
o
u
t sacr
i
f
ici
n
g
g
e
n
er
aliza
b
ili
t
y
.
4
.
1
.
2
.
CICI
DS2
0
1
9
d
a
t
a
s
et
T
h
e
C
I
C
DDo
S2
0
1
9
d
ataset
w
a
s
cr
ea
ted
b
y
th
e
C
a
n
ad
ian
I
n
s
tit
u
te
f
o
r
C
y
b
er
s
ec
u
r
it
y
(
C
I
C
)
at
th
e
Un
i
v
er
s
it
y
o
f
Ne
w
B
r
u
n
s
w
ic
k
.
I
t
is
am
o
n
g
t
h
e
last
an
d
th
e
m
o
s
t
co
m
p
lete
d
atasets
esp
e
ciall
y
d
ev
elo
p
ed
to
an
al
y
ze
m
o
d
er
n
DDo
S
attac
k
s
.
T
h
e
tr
af
f
ic
s
et
co
n
tai
n
s
in
ter
m
ed
iar
y
tr
a
f
f
ic
a
n
d
a
w
id
e
s
co
p
e
o
f
r
ec
en
t
DDo
S
v
ar
ian
t
s
:
p
o
r
t m
a
p
p
in
g
(
P
o
r
tMa
p
)
,
n
et
w
o
r
k
b
asic
i
n
p
u
t/o
u
tp
u
t
s
y
s
te
m
(
NetB
I
OS
)
,
li
g
h
t
w
eig
h
t
d
ir
ec
to
r
y
ac
ce
s
s
p
r
o
to
co
l
(
L
DA
P
)
,
Mic
r
o
s
o
f
t
S
QL
s
er
v
er
(
MSS
QL
)
,
u
s
er
d
ata
g
r
a
m
p
r
o
to
co
l
(
UDP
)
,
UDP
lat
en
c
y
a
ttack
(
U
DP
-
L
a
g
)
,
SYN,
n
et
w
o
r
k
ti
m
e
p
r
o
to
co
l
(
N
T
P
)
,
d
o
m
ain
n
a
m
e
s
y
s
te
m
(
DNS
)
,
an
d
s
i
m
p
le
n
et
w
o
r
k
m
a
n
a
g
e
m
e
n
t
p
r
o
to
co
l
(
SNMP
)
f
lo
o
d
attac
k
s
ar
e
i
n
clu
d
ed
.
T
h
ese
attac
k
s
w
er
e
i
m
p
le
m
e
n
ted
w
it
h
o
p
en
-
s
o
u
r
ce
attac
k
p
ac
k
ag
es
i
n
a
te
s
tb
ed
to
e
m
u
l
ate
v
o
l
u
m
e
tr
ic
an
d
r
e
f
lecti
v
e
f
lo
o
d
in
g
s
it
u
atio
n
s
a
i
m
in
g
at
g
en
er
ati
n
g
te
s
t
d
ata
th
at
clo
s
el
y
r
ep
r
esen
t
s
a
r
ea
l
-
w
o
r
ld
s
ce
n
ar
io
[3
2
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
1
6
9
3
-
6930
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
,
Vo
l.
23
,
No
.
5
,
Octo
b
e
r
20
25
:
1
2
1
2
-
1
227
1220
C
I
C
DDo
S2
0
1
9
[3
3
]
d
ataset
i
s
av
ailab
le
in
a
s
tr
u
ct
u
r
ed
f
o
r
m
at
to
b
e
u
s
ed
f
o
r
ML
,
in
clu
d
in
g
m
o
r
e
th
an
8
0
f
lo
w
f
ea
t
u
r
es
ca
p
tu
r
e
d
v
ia
C
I
C
F
lo
w
Me
ter
.
T
h
e
d
at
a
is
d
iv
id
ed
in
b
et
w
ee
n
t
w
o
d
a
y
s
o
f
co
llectio
n
,
in
o
r
d
er
to
k
ee
p
p
er
f
o
r
m
an
ce
tes
t
ed
in
tem
p
o
r
al
d
if
f
er
e
n
ce
s
.
Du
e
to
s
u
ch
r
ich
f
ea
tu
r
e
s
p
ac
e
an
d
h
ig
h
d
iv
er
s
it
y
o
f
att
ac
k
v
ec
to
r
s
,
it i
s
an
attr
ac
ti
v
e
b
aseli
n
e
f
o
r
m
o
d
el
’
s
ad
ap
tab
ilit
y
p
er
f
o
r
m
a
n
ce
an
d
d
etec
ti
o
n
ac
cu
r
ac
y
.
4
.
2
.
Usef
uln
ess
o
f
t
he
d
a
t
a
s
et
T
h
is
r
esear
ch
f
o
c
u
s
e
s
o
n
t
h
e
d
etec
tio
n
an
d
cla
s
s
i
f
icatio
n
o
f
D
Do
S
attac
k
s
—
c
y
b
er
t
h
r
ea
ts
c
h
ar
ac
ter
ized
b
y
o
v
er
w
h
el
m
i
n
g
a
tar
g
e
t
s
y
s
t
e
m
u
s
i
n
g
m
u
ltip
le
co
m
p
r
o
m
is
e
d
d
ev
ices
ac
ti
n
g
in
co
o
r
d
in
atio
n
.
T
h
is
d
if
f
er
s
f
r
o
m
Do
S a
ttack
s
,
w
h
ic
h
o
r
ig
i
n
ate
f
r
o
m
a
s
i
n
g
le
attac
k
i
n
g
d
ev
ice
t
o
ex
h
au
s
t t
h
e
r
eso
u
r
ce
s
o
f
a
tar
g
et.
W
h
ile
C
I
C
DDo
S2
0
1
9
_
Kag
g
l
e
d
ir
ec
tly
a
lig
n
s
w
it
h
th
e
s
co
p
e
o
f
th
i
s
s
t
u
d
y
b
y
p
r
o
v
id
in
g
la
b
eled
d
ata
f
o
r
v
ar
io
u
s
r
ea
l
-
w
o
r
ld
DDo
S
attac
k
t
y
p
es,
UNSW
-
NB
1
5
_
Kag
g
le
al
s
o
p
la
y
s
a
v
al
u
ab
le
r
o
le.
Desp
ite
b
ein
g
f
o
cu
s
ed
o
n
Do
S a
n
d
o
th
er
tr
ad
itio
n
a
l a
ttack
s
,
it o
f
f
er
s
r
ich
,
d
iv
er
s
e
f
ea
t
u
r
es a
n
d
co
m
p
lex
m
u
lti
-
clas
s
s
ce
n
ar
io
s
th
at
ca
n
e
n
h
a
n
ce
t
h
e
m
o
d
el
’
s
a
b
ilit
y
to
lear
n
g
en
er
alize
d
atta
ck
p
atter
n
s
.
T
h
is
i
s
p
ar
ticu
lar
l
y
i
m
p
o
r
ta
n
t f
o
r
p
r
e
-
tr
ain
i
n
g
s
ta
g
es,
e
n
ab
lin
g
t
h
e
d
etec
tio
n
f
r
a
m
e
w
o
r
k
to
d
i
f
f
er
en
tia
te
b
et
w
ee
n
b
en
i
g
n
a
n
d
m
alicio
u
s
b
e
h
av
io
r
s
u
n
d
er
d
if
f
er
e
n
t
attac
k
s
u
r
f
ac
e
s
.
Mo
r
eo
v
er
,
u
s
in
g
b
o
th
d
ata
s
ets
i
m
p
r
o
v
es
t
h
e
s
y
s
te
m
’
s
cr
o
s
s
-
d
o
m
a
in
ad
ap
tab
ilit
y
an
d
r
esil
ien
ce
to
e
v
o
lv
i
n
g
atta
ck
tech
n
iq
u
e
s
,
s
tr
e
n
g
t
h
e
n
i
n
g
it
s
ap
p
licab
ilit
y
i
n
r
ea
l
-
w
o
r
ld
S
DN
en
v
ir
o
n
m
en
t
s
.
4
.
3
.
P
re
pro
ce
s
s
ing
T
h
ese
in
clu
d
e
r
ec
o
g
n
izin
g
im
p
o
r
tan
t
co
m
p
o
n
en
ts
th
at
ar
e
im
p
o
r
tan
t
f
o
r
later
an
aly
s
is
,
n
o
r
m
aliz
i
n
g
f
ea
tu
r
es
to
u
n
iq
u
e
s
ca
les
f
o
r
s
tab
ilit
y
,
an
d
clea
n
in
g
r
aw
d
ata
to
r
em
o
v
e
n
o
is
e
an
d
r
ed
u
n
d
an
cy
.
Su
ch
a
s
tep
o
p
tim
izes
d
ata
f
r
o
m
p
ar
ticu
lar
s
ce
n
ar
io
s
an
d
lay
s
th
e
f
o
u
n
d
atio
n
f
o
r
th
e
m
o
d
el
to
lear
n
ef
f
icien
tly
.
Z
-
s
co
r
e
n
o
r
m
aliza
tio
n
w
as
ap
p
lied
to
s
tan
d
ar
d
ize
th
e
d
ataset,
en
s
u
r
in
g
th
at
f
ea
tu
r
es
h
ad
a
m
ea
n
o
f
ze
r
o
an
d
a
s
tan
d
ar
d
d
ev
iatio
n
o
f
o
n
e
.
Z
-
s
co
r
e
n
o
r
m
aliza
tio
n
in
v
o
lv
es
tr
an
s
f
o
r
m
i
n
g
v
alu
es
b
y
s
h
ar
in
g
b
y
s
tan
d
ar
d
d
ev
iatio
n
an
d
s
u
b
tr
ac
tin
g
th
e
m
ea
n
[3
4
]
.
T
h
e
z
-
s
co
r
e
n
o
r
m
aliza
tio
n
f
o
r
m
u
la
is
:
′
=
−
̅
(
)
(
2
)
w
h
er
e
is
t
h
e
v
al
u
e
o
f
r
o
w
E
o
f
th
e
i
-
t
h
co
lu
m
n
an
d
= z
-
s
co
r
e
n
o
r
m
alize
d
v
al
u
es.
(
)
=
√
1
(
−
1
)
∑
(
−
̅
)
2
=
1
(
3
)
an
d
m
ea
n
v
al
u
e,
̅
=
1
∑
=
1
(
4
)
Data
n
o
r
m
aliza
tio
n
ca
n
en
h
an
ce
m
o
d
el
p
er
f
o
r
m
an
ce
,
av
o
id
o
v
er
f
itti
n
g
,
an
d
p
r
o
m
o
te
co
n
v
er
g
en
ce
.
Scalin
g
th
e
d
ata
to
a
r
an
g
e
b
etw
ee
n
0
an
d
1
is
th
e
m
eth
o
d
em
p
lo
y
ed
in
d
ata
n
o
r
m
aliza
tio
n
.
A
ll
f
ea
tu
r
e
v
alu
e
s
ar
e
th
en
co
n
v
er
ted
to
a
co
m
m
o
n
s
ca
le.
4
.
4
.
Sp
litt
ing
t
he
d
a
t
a
s
et
Usi
n
g
an
8
0
%
tr
ain
in
g
an
d
2
0
%
test
in
g
r
ate,
th
e
s
tr
u
ctu
r
ed
d
ata
s
et
is
d
iv
id
ed
in
to
tr
ain
in
g
an
d
test
i
n
g
s
u
b
s
ets.
T
h
is
w
as
ac
h
iev
ed
b
y
im
p
lem
en
tin
g
th
e
d
ef
au
lt
tr
ain
-
test
s
p
lit
f
u
n
ctio
n
av
ailab
le
in
P
y
th
o
n
’
s
m
o
d
e
l
s
elec
tio
n
m
o
d
u
le,
en
s
u
r
in
g
an
ap
p
r
o
p
r
iate
d
iv
is
io
n
o
f
d
ata
f
o
r
m
o
d
el
ev
alu
atio
n
.
T
o
s
af
eg
u
ar
d
ea
ch
ass
au
l
t
lev
el
p
r
o
p
o
r
tio
n
ately
r
ep
r
esen
ted
in
a
b
asic c
o
llectio
n
o
f
d
ata,
th
e
o
r
d
er
n
o
t o
n
ly
s
h
u
f
f
les
a
s
et
o
f
d
ata
b
u
t a
ls
o
cr
ea
tes a
s
tr
atif
ied
s
tr
ateg
y
.
4
.
5
.
B
a
la
ncing
t
ec
hn
iqu
es
T
w
o
d
is
tin
ct
tr
ain
in
g
d
ata
p
r
ep
ar
atio
n
m
eth
o
d
s
w
er
e
ap
p
lied
,
s
p
ec
if
ically
tailo
r
ed
f
o
r
n
o
m
in
al
d
ata.
T
h
ese
m
eth
o
d
s
in
clu
d
ed
u
n
d
er
s
am
p
lin
g
,
o
v
er
s
am
p
lin
g
,
an
d
a
co
m
b
in
atio
n
o
f
v
ar
io
u
s
r
esam
p
lin
g
s
tr
ateg
ies
.
T
h
e
p
r
im
ar
y
o
b
j
ec
tiv
e
w
as
to
b
alan
ce
th
e
d
ata
s
ets
an
d
en
h
an
ce
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
d
ev
elo
p
ed
m
o
d
els
.
SMOT
E
-
ed
ited
n
ea
r
est
n
eig
h
b
o
r
s
(
SMOT
E
-
E
NN
)
ap
p
ea
r
s
as
SMOT
E
an
d
E
NN
m
eth
o
d
s
’
f
u
s
io
n
,
ef
f
icien
tly
d
ea
lin
g
w
ith
im
b
alan
ce
d
d
ata
class
if
icatio
n
co
n
ce
r
n
s
.
B
y
in
teg
r
atin
g
SMOT
E
’
s
m
in
o
r
ity
-
l
e
v
e
l
o
v
er
s
am
p
lin
g
w
ith
E
NN
’
s
m
aj
o
r
ity
-
class
u
n
d
er
s
am
p
lin
g
,
th
e
tech
n
iq
u
e
h
ar
m
o
n
izes
th
e
s
h
ar
e
o
f
th
e
d
ataset.
Su
ch
a
m
eth
o
d
w
as
f
o
r
m
u
lated
b
y
[3
5
]
as a
s
tr
o
n
g
s
tr
ateg
y
to
co
n
tr
o
l c
lass
im
b
alan
ce
.
L
ik
ely
,
SMOT
E
-
T
o
m
ek
lin
k
s
(
S
MO
T
E
-
T
OM
E
K
)
is
th
e
o
th
er
am
alg
am
atio
n
m
eth
o
d
d
ev
elo
p
ed
in
im
b
alan
ce
d
d
ata
class
if
icatio
n
s
ce
n
ar
io
s
.
T
h
is
is
a
p
o
ten
t
ap
p
r
o
ac
h
r
en
o
w
n
ed
f
o
r
its
ef
f
icac
y
in
co
n
s
id
er
in
g
im
b
alan
ce
d
s
ets
o
f
d
ata,
s
p
ec
if
ically
w
h
ile
m
ee
tin
g
n
o
is
y
/o
v
er
lap
p
in
g
s
am
p
les.
B
y
co
m
b
in
i
n
g
SMOT
E
an
d
T
o
m
ek
li
n
k
s
,
s
u
ch
a
m
eth
o
d
ef
f
icien
tly
n
av
ig
ates
im
b
alan
ce
d
s
ce
n
ar
io
s
f
o
r
d
ev
elo
p
in
g
m
o
d
el
p
er
f
o
r
m
an
ce
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KOM
NI
K
A
T
elec
o
m
m
u
n
C
o
m
p
u
t E
l
C
o
n
tr
o
l
DDo
S
a
tta
ck
d
etec
tio
n
u
s
in
g
o
p
tima
l scr
u
tin
y
b
o
o
s
ted
g
r
a
p
h
co
n
v
o
lu
tio
n
a
l …
(
Hu
d
a
Mo
h
a
mme
d
I
b
a
d
i
)
1221
4
.
6
.
P
er
f
o
rm
a
nce
m
et
rics
T
h
e
p
r
esen
ted
p
er
f
o
r
m
an
ce
o
f
th
e
in
tr
u
s
io
n
m
o
d
el
w
as
ev
alu
ated
u
s
in
g
th
e
v
ar
iab
les
o
f
r
ec
all,
ac
cu
r
ac
y
,
F
-
s
co
r
e,
an
d
p
r
ec
is
io
n
.
T
h
e
class
if
icatio
n
m
etr
ics o
v
er
v
iew
is
p
r
esen
ted
h
er
e.
A
p
p
r
o
p
r
iately
g
r
o
u
p
ed
d
ata
r
ate
o
u
t
o
f
w
h
o
le
g
r
o
u
p
ed
d
ata
is
h
o
w
ap
p
r
o
p
r
iately
s
o
m
eth
in
g
is
g
r
o
u
p
ed
.
T
h
e
p
r
ec
is
io
n
d
eter
m
in
es
th
e
ac
cu
r
ac
y
o
f
o
p
tim
is
tic
p
r
ed
ictio
n
s
;
a
lo
w
er
f
alse
p
o
s
itiv
e
r
atio
d
en
o
tes
m
o
r
e
p
r
ec
is
io
n
.
T
h
e
p
er
ce
n
tag
e
o
f
s
am
p
les
th
at
ar
e
ac
cu
r
ately
class
if
ied
as
p
o
s
itiv
e
is
k
n
o
w
n
as
r
ec
all.
R
ec
all
an
d
p
r
ec
is
io
n
ar
e
co
m
b
in
ed
in
th
e
F
-
s
co
r
e
to
cr
ea
te
a
f
air
ass
ess
m
en
t
m
etr
ic.
T
h
is
is
f
ea
s
ib
le
to
d
escr
ib
e
th
at
as
a
p
r
e
cisi
o
n
an
d
r
ec
all
m
ed
i
u
m
.
W
h
o
le
m
etr
ics
ar
e
g
iv
en
th
e
co
n
f
u
s
io
n
m
atr
ix
.
T
h
is
m
atr
ix
r
o
w
s
an
d
co
lu
m
n
s
ar
e
lab
eled
w
ith
ce
r
tain
lev
el
s
(
g
r
o
u
n
d
tr
u
th
)
an
d
p
r
ed
icted
lev
els
in
tu
r
n
.
T
ab
le
3
s
h
o
w
s
th
e
n
o
r
m
al
co
n
f
u
s
io
n
m
atr
ix
f
o
r
a
b
in
ar
y
class
if
icatio
n
is
s
u
e.
T
ab
le
3
.
C
o
n
f
u
s
io
n
m
atr
i
x
f
o
r
a
b
in
ar
y
clas
s
if
icatio
n
A
c
t
u
a
l
/
p
r
e
d
i
c
t
e
d
P
r
e
d
i
c
t
e
d
p
o
s
i
t
i
v
e
(
a
t
t
a
c
k
)
P
r
e
d
i
c
t
e
d
n
e
g
a
t
i
v
e
(
n
o
r
mal
)
A
c
t
u
a
l
p
o
si
t
i
v
e
(
a
t
t
a
c
k
)
TP
FN
A
c
t
u
a
l
n
e
g
a
t
i
v
e
(
n
o
r
mal
)
FP
TN
A
d
d
itio
n
ally
,
tr
u
e
p
o
s
itiv
e,
tr
u
e
n
eg
ativ
e,
f
alse
p
o
s
itiv
e,
an
d
f
alse
n
eg
ativ
e
ar
e
d
en
o
ted
as
,
,
,
an
d
,
r
esp
ec
tiv
ely
.
T
r
u
e
p
o
s
itiv
e
(
TP
)
r
ef
er
s
to
attac
k
d
ata
v
o
lu
m
e
in
n
etw
o
r
k
tr
af
f
ic
th
at
is
lab
eled
as a
n
attac
k
.
T
r
u
e
ne
g
ativ
e
(
TN
)
s
h
o
w
s
n
etw
o
r
k
d
ata
v
o
lu
m
e
in
th
e
n
etw
o
r
k
tr
af
f
ic
th
at
is
co
n
s
id
er
ed
n
o
r
m
al.
Fals
e
p
o
s
itiv
e
(
FP
)
is
n
o
r
m
al
in
s
tan
ce
s
in
n
etw
o
r
k
tr
af
f
ic
th
at
ar
e
m
is
class
if
ied
as
attac
k
in
s
tan
ce
s
.
Fals
e
n
eg
ativ
e
(
FN
)
r
elate
s
to
attac
k
in
s
tan
ce
s
in
n
etw
o
r
k
tr
af
f
ic
T
h
at
a
r
e
m
is
class
if
ied
as n
o
r
m
al.
T
h
e
class
if
icatio
n
m
etr
ics
ass
ess
in
g
an
I
DS
s
y
s
tem
ar
e
d
escr
ib
ed
:
C
lass
if
icatio
n
r
ate
o
r
a
cc
u
r
ac
y
(
C
R
)
:
th
is
is
th
e
ac
cu
r
ac
y
r
ate
o
n
d
iag
n
o
s
in
g
u
n
u
s
u
al/u
s
u
al
m
an
n
er
.
T
h
is
is
u
s
u
ally
ap
p
lied
w
h
ile
a
s
et
o
f
d
ata
is
b
alan
ce
d
to
p
r
e
v
en
t a
g
o
o
d
m
o
d
el
p
er
f
o
r
m
an
ce
f
r
o
m
f
alse
s
en
s
e.
=
+
+
+
+
(
5
)
R
ec
all
(
R
)
o
r
tr
u
e
p
o
s
itiv
e
r
ate
(
T
P
R
)
o
r
Sen
s
itiv
ity
:
th
is
ca
u
s
ed
b
y
s
h
ar
in
g
n
u
m
b
er
o
f
ac
cu
r
ately
esti
m
at
e
d
attac
k
s
b
y
en
tire
attac
k
s
’
n
u
m
b
er
.
T
h
is
is
cl
ass
if
ier
ab
ilit
y
to
d
iag
n
o
s
e
w
h
o
le
p
o
s
itiv
e
ca
s
es (
attac
k
s
)
.
=
+
(
6
)
P
r
ec
is
io
n
(
P
)
:
s
h
o
w
s
attac
k
d
iag
n
o
s
is
co
n
f
id
en
ce
.
T
h
is
is
ef
f
ec
tiv
e
b
u
t
th
at
is
n
o
t
r
ec
o
m
m
en
d
ed
to
ap
p
ly
A
cc
u
r
ac
y
d
u
e
to
th
e
s
et
o
f
d
ata
is
n
o
t
b
alan
ce
d
.
=
+
(
7
)
F1
-
s
co
r
e:
th
is
is
th
e
h
ar
m
o
n
ic
m
ea
n
am
o
n
g
p
r
ec
is
io
n
an
d
r
ec
all
(
T
P
R
)
,
A
ch
iev
in
g
a
v
alu
e
clo
s
e
to
th
e
lo
w
er
o
f
th
e
tw
o
class
d
is
tr
ib
u
tio
n
s
,
p
ar
ticu
lar
ly
in
ca
s
es
o
f
class
im
b
alan
ce
,
in
d
icate
s
im
p
r
o
v
ed
class
if
icati
o
n
b
alan
ce
.
A
v
alu
e
ap
p
r
o
ac
h
in
g
1
s
u
g
g
ests
th
at
th
e
class
if
ier
is
p
er
f
o
r
m
in
g
o
p
tim
ally
in
d
is
tin
g
u
is
h
in
g
b
etw
e
e
n
th
e
class
es.
1
−
=
2
+
(
8
)
4
.
7
.
T
ra
ini
ng
a
nd
t
esting
T
h
e
d
ataset
is
p
ar
titi
o
n
ed
in
to
8
0
%
f
o
r
tr
ain
in
g
an
d
2
0
%
f
o
r
t
esti
n
g
f
o
r
th
e
ev
al
u
atio
n
o
f
t
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
.
T
o
r
e
d
u
ce
th
e
m
o
d
el
er
r
o
r
,
th
e
tr
ain
in
g
p
r
o
ce
s
s
i
s
p
er
f
o
r
m
ed
f
o
r
1
0
0
ep
o
ch
s
,
u
s
i
n
g
a
lear
n
i
n
g
r
ate
o
f
0
.
0
0
1
f
o
r
q
u
ick
co
n
v
er
g
e
n
ce
.
Fig
u
r
e
s
2
an
d
3
s
h
o
w
tr
en
d
s
o
f
lo
s
s
an
d
ac
cu
r
ac
y
d
u
r
in
g
b
o
th
tr
ain
i
n
g
an
d
test
in
g
s
tag
e
s
i
n
t
w
o
d
ataset
s
.
T
h
e
lo
s
s
tr
e
n
d
s
ar
e
d
ep
icted
in
Fi
g
u
r
es
2
(
a)
an
d
3
(
a)
,
w
h
ile
th
e
a
cc
u
r
ac
y
tr
e
n
d
s
ar
e
s
h
o
w
n
in
Fig
u
r
es
2
(
b
)
an
d
3
(
b
)
.
T
h
e
lo
s
s
v
al
u
es
w
er
e
b
et
w
ee
n
0
.
1
0
an
d
0
.
5
0
.
T
h
is
ev
alu
atio
n
w
as
o
n
t
h
e
b
en
ch
m
ar
k
d
ataset
s
b
u
t
t
h
e
p
er
f
o
r
m
an
ce
o
f
th
e
p
r
o
p
o
s
ed
m
o
d
el
s
u
g
g
est
s
its
ca
p
ab
ilit
y
in
id
en
ti
f
y
i
n
g
at
tack
s
n
o
t
s
p
ec
if
icall
y
test
ed
i
n
th
e
e
v
alu
a
tio
n
.
I
n
ad
d
itio
n
,
a
s
s
h
o
w
n
i
n
F
i
g
u
r
e
s
1
an
d
2
,
th
e
v
a
lu
e
o
f
tr
ai
n
i
n
g
lo
s
s
s
tar
ts
h
ig
h
a
n
d
th
e
n
d
ec
r
ea
s
es
s
lo
w
l
y
i
n
t
h
e
tr
ain
in
g
p
r
o
ce
s
s
.
A
s
i
g
n
i
f
ica
n
t
s
lo
w
i
n
g
d
o
w
n
in
t
h
e
r
ed
u
ctio
n
o
f
er
r
o
r
s
is
u
s
u
a
ll
y
o
b
s
er
v
ed
a
f
te
r
ab
o
u
t 2
0
e
p
o
ch
s
.
T
h
e
ac
cu
r
ac
y
o
f
tr
ain
i
n
g
an
d
t
esti
n
g
s
tead
il
y
i
n
cr
ea
s
ed
as th
e
n
u
m
b
er
o
f
ep
o
ch
s
in
cr
ea
s
ed
an
d
f
i
n
all
y
m
ai
n
t
ai
n
at
t
h
e
h
i
g
h
er
le
v
el
s
as
p
er
f
or
m
ed
in
Fi
g
u
r
e
2
(
a)
i
n
d
icate
s
t
h
e
m
o
d
el
h
as
a
g
o
o
d
g
e
n
e
r
aliza
tio
n
p
r
o
p
er
t
y
in
tr
an
s
f
er
r
in
g
th
e
lear
n
i
n
g
m
o
d
el
as
a
m
ap
p
in
g
f
r
o
m
th
e
tr
ain
i
n
g
d
ata
to
u
n
s
ee
n
s
a
m
p
les
.
T
h
e
co
r
r
esp
o
n
d
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.