I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI)
Vo
l.
1
3
,
No
.
4
,
Dec
em
b
er
2
0
2
4
,
p
p
.
4
4
8
8
~
4
4
9
8
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
1
3
.
i
4
.
p
p
4
4
8
8
-
4
4
9
8
4488
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
Ev
a
lua
tion o
f
dist
ributed
denial o
f
serv
ice att
a
cks
de
tect
io
n i
n
so
ft
wa
re de
fined
networks
Nee
t
hu
S
.
1
,
H
.
V.
Ra
v
is
h Ar
a
dh
y
a
2
1
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
En
g
i
n
e
e
r
i
n
g
,
R
V
C
o
l
l
e
g
e
o
f
En
g
i
n
e
e
r
i
n
g
,
V
i
s
v
e
s
v
a
r
a
y
a
T
e
c
h
n
o
l
o
g
i
c
a
l
U
n
i
v
e
r
s
i
t
y
,
K
a
r
n
a
t
a
k
a
,
I
n
d
i
a
2
D
e
p
a
r
t
me
n
t
o
f
El
e
c
t
r
o
n
i
c
s a
n
d
C
o
m
mu
n
i
c
a
t
i
o
n
En
g
i
n
e
e
r
i
n
g
,
R
V
C
o
l
l
e
g
e
o
f
E
n
g
i
n
e
e
r
i
n
g
,
V
i
s
v
e
s
v
a
r
a
y
a
Te
c
h
n
o
l
o
g
i
c
a
l
U
n
i
v
e
r
si
t
y
,
K
a
r
n
a
t
a
k
a
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
an
13
,
2
0
2
4
R
ev
is
ed
Mar
22
,
2
0
2
4
Acc
ep
ted
Ap
r
17
,
2
0
2
4
S
o
ftwa
re
-
d
e
fin
e
d
n
e
tw
o
rk
i
n
g
(S
D
N)
re
v
o
l
u
ti
o
n
ize
s
n
e
two
rk
in
g
b
y
s
e
p
a
ra
ti
n
g
c
o
n
tro
l
lo
g
ic
a
n
d
d
a
ta
fo
rwa
rd
i
n
g
,
e
n
h
a
n
c
in
g
se
c
u
rit
y
a
g
a
i
n
st
t
h
re
a
ts
li
k
e
d
istri
b
u
ted
d
e
n
ial
o
f
se
rv
ice
(DD
o
S
)
a
tt
a
c
k
s.
Th
e
se
a
tt
a
c
k
s
flo
o
d
c
o
n
tro
l
p
lan
e
b
a
n
d
wi
d
th
,
c
a
u
sin
g
S
DN
n
e
tw
o
rk
fa
il
u
re
s.
Re
c
e
n
t
stu
d
ies
e
m
p
h
a
siz
e
th
e
e
ffica
c
y
o
f
m
a
c
h
in
e
lea
rn
in
g
(
M
L)
a
n
d
sta
ti
stica
l
a
p
p
r
o
a
c
h
e
s
in
i
d
e
n
ti
f
y
in
g
a
n
d
m
it
ig
a
ti
n
g
th
e
se
se
c
u
rit
y
risk
s.
Ho
we
v
e
r,
th
e
re
h
a
s
b
e
e
n
a
lac
k
o
f
f
o
c
u
s
o
n
e
m
p
lo
y
i
n
g
e
n
se
m
b
li
n
g
tec
h
n
iq
u
e
s,
a
m
a
lg
a
m
a
ti
n
g
d
i
v
e
rse
ML
m
o
d
e
ls,
se
lec
ti
n
g
p
e
rti
n
e
n
t
fe
a
tu
re
s,
a
n
d
u
ti
li
z
in
g
o
v
e
rsa
m
p
li
n
g
tec
h
n
iq
u
e
s to
b
a
lan
c
e
c
a
teg
o
rica
l
d
a
ta.
Ou
r
stu
d
y
e
v
a
lu
a
tes
2
0
m
a
c
h
i
n
e
-
lea
rn
in
g
m
o
d
e
ls
,
e
m
p
h
a
siz
in
g
fe
a
tu
re
e
n
g
i
n
e
e
rin
g
a
n
d
a
d
d
re
ss
in
g
c
las
s
imb
a
lan
c
e
u
sin
g
sy
n
th
e
ti
c
m
in
o
r
it
y
o
v
e
rsa
m
p
li
n
g
te
c
h
n
i
q
u
e
(S
M
OTE)
.
Th
e
re
su
l
t
s
in
d
ica
te
th
a
t
e
n
se
m
b
le
m
e
th
o
d
s
su
c
h
a
s
li
g
h
t
g
ra
d
ien
t
b
o
o
stin
g
m
a
c
h
in
e
(LG
BM
)
c
las
sifier,
ra
n
d
o
m
fo
re
st
c
las
sifier,
XG
B
c
las
sifier,
d
e
c
isio
n
tree
c
las
sifier
o
b
tai
n
e
d
n
e
a
r
-
p
e
rfe
c
t
sc
o
re
s
(a
lmo
st
1
0
0
%
)
a
c
ro
ss
a
ll
m
e
tri
c
s,
su
g
g
e
st
i
n
g
p
o
ten
ti
a
l
o
v
e
rfit
ti
n
g
.
C
o
n
v
e
rse
ly
,
m
o
d
e
ls
li
k
e
Ad
a
Bo
o
st
c
las
sifier,
k
-
n
e
ig
h
b
o
rs
c
las
sifier,
a
n
d
su
p
p
o
r
t
v
e
c
to
r
c
las
sifier
(
S
VC
)
e
x
h
i
b
it
e
d
slig
h
tl
y
lo
we
r
(9
9
%
)
b
u
t
re
a
li
stic
p
e
rfo
rm
a
n
c
e
,
u
n
d
e
rsc
o
rin
g
th
e
i
n
tri
c
a
c
y
o
f
a
c
c
u
ra
te
p
re
d
ictio
n
in
c
y
b
e
rse
c
u
rit
y
.
S
im
p
ler
m
o
d
e
ls,
i
n
c
lu
d
in
g
l
o
g
isti
c
r
e
g
re
ss
io
n
,
li
n
e
a
r
d
isc
rimin
a
n
t
a
n
a
ly
sis,
a
n
d
G
a
u
ss
ian
n
a
iv
e
Ba
y
e
s,
d
e
m
o
n
stra
ted
m
o
d
e
ra
te
to
l
o
w
a
c
c
u
ra
c
y
,
a
p
p
r
o
x
ima
tely
a
ro
u
n
d
7
0
%
.
T
h
e
se
fin
d
in
g
s
stre
ss
th
e
imp
e
ra
ti
v
e
n
e
e
d
fo
r
a
n
u
a
n
c
e
d
a
p
p
ro
a
c
h
in
t
h
e
se
lec
ti
o
n
a
n
d
f
in
e
-
tu
n
i
n
g
o
f
ML
m
o
d
e
ls t
o
e
n
s
u
re
e
ffe
c
ti
v
e
D
Do
S
d
e
tec
ti
o
n
i
n
S
DN
e
n
v
ir
o
n
m
e
n
ts.
K
ey
w
o
r
d
s
:
Dis
tr
ib
u
ted
d
en
ial
o
f
s
er
v
ice
attac
k
s
Ma
ch
in
e
lear
n
in
g
Op
en
Flo
w
p
r
o
to
c
o
l
So
f
twar
e
-
d
ef
in
e
d
n
etwo
r
k
s
Sy
n
th
etic
m
in
o
r
ity
o
v
er
s
am
p
lin
g
tech
n
iq
u
e
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
:
Nee
th
u
S
.
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
E
n
g
in
ee
r
in
g
,
R
V
C
o
lleg
e
o
f
E
n
g
i
n
ee
r
in
g
Vis
v
esv
ar
ay
a
T
ec
h
n
o
l
o
g
ical
Un
iv
er
s
ity
My
s
o
r
e
R
o
ad
,
B
en
g
alu
r
u
5
6
0
0
5
9
,
I
n
d
ia
E
m
ail:
n
ee
th
u
s
@
r
v
ce
.
ed
u
.
in
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
lan
d
s
ca
p
e
o
f
s
o
f
twar
e
-
d
r
iv
en
n
etwo
r
k
s
u
n
d
er
g
o
es
a
r
ev
o
lu
tio
n
a
r
y
tr
an
s
f
o
r
m
ati
o
n
th
r
o
u
g
h
s
o
f
twar
e
-
d
ef
in
ed
n
etwo
r
k
in
g
(
SDN)
,
a
n
o
v
el
p
ar
a
d
ig
m
th
at
ef
f
ec
tiv
ely
s
ep
ar
ates
th
e
co
n
tr
o
l
an
d
d
ata
p
la
n
es,
ad
d
r
ess
in
g
is
s
u
es
in
h
er
en
t
in
tr
ad
itio
n
al
n
etwo
r
k
d
esig
n
.
SDN
h
as
em
er
g
ed
as
a
p
ar
ad
ig
m
a
tic
s
h
if
t,
ad
d
r
ess
in
g
lim
itatio
n
s
p
r
esen
t
in
co
n
v
en
t
io
n
al
I
P
-
b
ased
n
etwo
r
k
s
.
T
h
e
ar
ch
itectu
r
e
o
f
SDN
f
ac
ilit
ates
th
e
d
ec
o
u
p
lin
g
o
f
th
e
co
n
tr
o
l
p
lan
e
an
d
d
ata
p
lan
e,
en
ab
lin
g
a
ce
n
tr
alize
d
co
n
tr
o
ller
to
ab
s
tr
ac
t
th
e
u
n
d
er
ly
in
g
n
etwo
r
k
f
r
am
ewo
r
k
f
r
o
m
a
p
p
licatio
n
s
.
T
h
is
ab
s
tr
a
ctio
n
en
s
u
r
es
s
im
p
lifie
d
r
eso
u
r
ce
p
r
o
v
is
io
n
i
n
g
an
d
en
h
an
ce
d
p
r
o
g
r
a
m
m
ab
ilit
y
o
f
n
etwo
r
k
ch
a
r
ac
ter
is
tics
.
C
o
m
p
ar
ed
to
tr
a
d
itio
n
al
n
etwo
r
k
s
,
th
e
lo
g
ical
ce
n
tr
aliza
tio
n
o
f
co
n
tr
o
ller
s
in
SDN
p
r
o
v
id
es
im
p
r
o
v
ed
v
is
ib
ilit
y
o
f
t
h
e
e
n
tire
n
etwo
r
k
,
f
ac
ilit
atin
g
ef
f
icien
t
m
a
n
ag
em
e
n
t
a
n
d
o
p
tim
izatio
n
o
f
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2
2
5
2
-
8
9
3
8
E
va
lu
a
tio
n
o
f d
is
tr
ib
u
ted
d
en
i
a
l o
f ser
vice
a
tta
ck
s
d
etec
tio
n
in
s
o
ftw
a
r
e
d
efin
ed
n
etw
o
r
ks
(
N
ee
th
u
S
.
)
4489
n
etwo
r
k
r
eso
u
r
ce
s
[
1
]
.
Ad
d
itio
n
ally
,
SDN
allo
ws
f
o
r
tailo
r
ed
q
u
ality
o
f
s
er
v
ice
(
Qo
S)
p
r
o
v
i
s
io
n
in
g
to
m
ee
t
th
e
r
eq
u
ir
em
e
n
ts
o
f
v
ar
i
o
u
s
ap
p
lic
atio
n
s
.
As
a
r
esu
lt,
SDN
h
as
f
o
u
n
d
wid
esp
r
ea
d
ad
o
p
tio
n
in
la
r
g
e
-
s
ca
le
n
etwo
r
k
s
an
d
d
ata
ce
n
ter
s
.
Op
e
n
Flo
w
tech
n
o
lo
g
y
[
2
]
e
m
p
o
wer
s
Op
e
n
Flo
w
co
n
tr
o
ller
s
to
m
an
ag
e
m
u
ltip
le
h
ar
d
war
e
o
r
s
o
f
twar
e
en
titi
es.
T
h
e
SDN
ar
ch
itectu
r
e
is
v
er
tically
d
iv
id
ed
in
to
th
r
ee
p
lan
es:
d
ata
p
lan
e,
co
n
tr
o
ller
p
lan
e,
an
d
ap
p
licatio
n
p
la
n
e.
T
h
e
co
n
tr
o
l
p
lan
e
o
r
c
h
estra
tes
n
etwo
r
k
c
o
n
f
ig
u
r
atio
n
an
d
p
o
lic
ies.
I
n
tr
ad
itio
n
al
ap
p
licatio
n
-
s
p
ec
if
ic
in
te
g
r
ate
d
cir
cu
it
(
ASI
C
)
s
witch
es,
wh
er
e
th
e
d
ata
p
la
n
e
an
d
c
o
n
tr
o
l
p
lan
e
c
o
ex
is
t,
ac
co
m
m
o
d
atin
g
f
lo
w
ta
b
le
e
n
tr
ies
an
d
b
u
f
f
er
s
p
ac
e
b
ec
o
m
es
n
ea
r
ly
im
p
o
s
s
ib
le.
T
h
e
co
n
tr
o
ller
is
s
u
es
in
s
tr
u
ctio
n
s
to
th
e
f
o
r
war
d
in
g
elem
en
ts
th
r
o
u
g
h
th
e
So
u
th
b
o
u
n
d
in
ter
f
ac
e.
Ho
wev
er
,
t
h
e
s
am
e
ch
ar
ac
ter
is
tics
th
at
m
ak
e
SDN
n
etwo
r
k
s
ap
p
e
alin
g
also
ex
p
o
s
e
th
em
to
n
ew
s
ec
u
r
ity
r
is
k
s
.
On
e
s
u
ch
r
is
k
i
s
d
is
tr
ib
u
ted
d
e
n
ial
o
f
s
er
v
ice
(
DDo
S)
attac
k
s
,
w
h
ich
ca
n
h
av
e
a
ca
tast
r
o
p
h
ic
im
p
ac
t
o
n
an
SDN
n
etwo
r
k
.
I
f
th
e
n
etwo
r
k
is
in
ad
eq
u
ately
p
r
o
tecte
d
,
DDo
S
attac
k
s
ca
n
o
v
er
wh
e
lm
th
e
c
o
n
tr
o
ller
o
r
Op
e
n
Flo
w
s
witch
b
y
d
ir
ec
tin
g
p
ac
k
ets
to
v
ar
io
u
s
p
o
in
ts
.
N
u
m
er
o
u
s
r
esear
ch
er
s
h
av
e
c
o
n
tr
ib
u
ted
t
o
ad
v
an
ci
n
g
s
tr
ateg
ies
f
o
r
d
etec
tin
g
an
d
m
itig
atin
g
DDo
S
attac
k
s
,
em
p
lo
y
in
g
d
iv
er
s
e
m
eth
o
d
o
lo
g
ies
an
d
f
r
am
e
wo
r
k
s
[
3
]
.
E
x
te
n
s
iv
e
liter
atu
r
e
ex
is
ts
th
at
d
elv
es
in
to
r
esear
ch
c
h
allen
g
es
an
d
p
o
ten
tial
f
u
tu
r
e
d
ir
ec
tio
n
s
f
o
r
DDo
S
attac
k
d
etec
tio
n
an
d
m
i
tig
atio
n
in
SDN
[
1
]
,
[
4
]
–
[
7
]
.
Sev
er
al
r
esear
ch
e
r
s
h
av
e
p
r
o
p
o
s
ed
in
n
o
v
ativ
e
ap
p
r
o
ac
h
es
to
DDo
S
attac
k
d
etec
t
io
n
an
d
m
itig
atio
n
with
in
an
SDN
ar
ch
itectu
r
e.
Fo
r
in
s
tan
ce
,
R
ib
eir
o
et
a
l.
[
8
]
in
tr
o
d
u
ce
d
a
tim
e
-
b
ased
m
o
v
in
g
tar
g
et
d
e
f
en
s
e
(
MT
D)
tech
n
iq
u
e
u
tili
zin
g
m
a
ch
in
e
lear
n
in
g
(
ML
)
s
en
s
o
r
s
f
o
r
r
ea
l
-
tim
e
DDo
S a
ttack
d
iag
n
o
s
is
an
d
co
n
tr
o
ller
m
ig
r
atio
n
.
B
h
ay
o
et
a
l.
[
9
]
e
x
p
lo
r
ed
n
aiv
e
B
ay
es
(
NB
)
,
d
ec
is
io
n
tr
ee
(
DT
)
,
an
d
s
u
p
p
o
r
t
v
ec
t
o
r
m
ac
h
in
e
(
SVM)
alg
o
r
ith
m
s
,
ac
h
iev
i
n
g
r
e
m
ar
k
ab
le
ac
cu
r
ac
y
in
DDo
S
attac
k
d
etec
tio
n
s
im
u
latio
n
s
.
C
o
W
atch
[
1
0
]
p
r
esen
ted
a
f
r
am
ewo
r
k
f
o
r
co
o
p
er
ativ
e
p
r
ed
ictio
n
an
d
d
etec
tio
n
o
f
D
Do
S
attac
k
s
in
ed
g
e
co
m
p
u
ti
n
g
(
E
C
)
s
ce
n
ar
io
s
,
lev
er
ag
in
g
SDN'
s
d
i
s
tr
ib
u
ted
ar
ch
itectu
r
e
an
d
em
p
lo
y
in
g
th
e
lo
n
g
s
h
o
r
t
-
ter
m
m
e
m
o
r
y
(
L
STM
)
alg
o
r
ith
m
.
T
h
e
DOCUS
m
o
d
el
[
1
1
]
was
p
r
o
p
o
s
ed
f
o
r
d
is
tin
g
u
is
h
in
g
g
en
u
in
e
DDo
S
attac
k
s
f
r
o
m
h
ar
m
less
f
lash
tr
af
f
ic,
r
ed
u
cin
g
av
er
ag
e
d
etec
tio
n
ti
m
e
s
ig
n
if
ican
tly
in
SDN
-
b
as
ed
n
etwo
r
k
s
.
An
y
a
n
wu
et
a
l.
[
1
2
]
d
e
v
elo
p
e
d
an
in
tr
u
s
io
n
d
etec
tio
n
s
y
s
tem
u
s
in
g
th
e
r
a
d
ial
b
asis
f
u
n
ctio
n
k
e
r
n
el
(
R
B
F)
o
f
SVM
f
o
r
d
etec
ti
n
g
DDo
S a
ttack
s
in
v
eh
icu
lar
a
d
h
o
c
n
etwo
r
k
s
(
V
ANE
T
s
)
with
h
ig
h
ac
c
u
r
ac
y
.
Nau
la
et
a
l
.
[
1
3
]
p
r
o
p
o
s
ed
a
s
ca
lab
le
SDN
-
b
ased
s
ec
u
r
ity
f
r
am
ew
o
r
k
em
p
lo
y
in
g
d
ee
p
r
ein
f
o
r
ce
m
en
t
lear
n
in
g
(
DR
L
)
f
o
r
ef
f
ec
tiv
e
D
Do
S
d
etec
tio
n
a
n
d
m
itig
atio
n
.
Nu
m
er
o
u
s
s
tu
d
ies
ex
p
lo
r
e
th
e
a
p
p
licatio
n
o
f
d
e
ep
lear
n
in
g
tec
h
n
iq
u
es,
s
u
ch
as
r
ec
u
r
r
en
t
n
e
u
r
al
n
etwo
r
k
s
(
R
NN)
,
f
o
r
DDo
S d
etec
tio
n
[
1
4
]
,
[
1
5
]
.
T
h
ese
m
eth
o
d
s
f
in
d
p
r
ev
alen
ce
in
d
ata
c
en
ter
s
[
1
6
]
,
[
1
7
]
an
d
clo
u
d
c
o
m
p
u
tin
g
en
v
ir
o
n
m
en
t
s
[
1
8
]
.
L
o
n
g
an
d
J
in
s
o
n
g
[
19
]
p
r
o
p
o
s
ed
a
d
u
al
ap
p
r
o
ac
h
,
c
o
m
b
in
in
g
in
f
o
r
m
atio
n
en
tr
o
p
y
an
d
a
ML
f
r
am
ew
o
r
k
u
s
in
g
an
s
tack
ed
s
p
a
r
s
e
au
to
-
en
co
d
er
(
SS
AE
)
-
SVM
ar
ch
ite
ctu
r
e
f
o
r
im
p
r
o
v
ed
DDo
S
d
etec
tio
n
.
A
f
r
am
ewo
r
k
in
co
r
p
o
r
atin
g
a
co
m
p
o
s
ite
m
u
ltil
ay
er
p
e
r
ce
p
tr
o
n
an
d
a
n
ef
f
ec
tiv
e
f
ea
tu
r
e
ex
tr
ac
tio
n
tech
n
i
q
u
e
f
o
r
d
etec
tin
g
an
d
id
en
tify
in
g
DDo
S a
s
s
au
lts
was p
r
esen
ted
[2
0
]
.
An
an
o
m
aly
tr
ee
p
r
o
p
o
s
ed
[2
1
]
is
tailo
r
ed
to
d
etec
t
an
d
tr
ac
e
th
e
s
p
ec
if
ic
r
o
u
te
tak
en
b
y
a
DDo
S
attac
k
with
in
th
e
SDN
ar
ch
itectu
r
e,
r
esp
o
n
d
in
g
to
o
b
s
er
v
ed
f
lu
ctu
a
tio
n
s
in
n
etwo
r
k
tr
af
f
ic.
C
u
i
et
a
l.
[2
2
]
illu
s
tr
ated
a
DDo
S
ass
au
lt
d
etec
tio
n
an
d
d
ef
en
s
e
s
y
s
tem
b
ased
o
n
co
g
n
itiv
e
-
in
s
p
ir
ed
co
m
p
u
tin
g
,
wh
i
ch
u
s
es
d
u
al
ad
d
r
ess
en
tr
o
p
y
an
d
s
u
g
g
ests
s
u
itab
le
d
ef
en
s
e
an
d
r
ec
o
v
er
y
p
r
o
ce
d
u
r
es
u
p
o
n
attac
k
d
etec
tio
n
.
I
m
r
a
n
et
a
l.
[2
3
]
p
r
o
v
id
es
an
ex
ten
s
iv
e
an
aly
s
is
o
f
s
ev
er
al
m
itig
atio
n
s
tr
ateg
ies
d
ev
elo
p
ed
t
o
c
o
m
b
at
m
alicio
u
s
tr
af
f
ic
in
th
e
SDN
en
v
ir
o
n
m
en
t,
s
y
s
tem
atica
lly
p
r
o
v
id
in
g
s
o
lu
tio
n
s
b
ased
o
n
h
o
w
th
ey
m
an
ag
e
s
u
ch
m
alicio
u
s
tr
af
f
ic.
Acc
o
r
d
i
n
g
to
Sah
ay
et
a
l.
[2
4
]
,
an
au
t
o
n
o
m
ic
DDo
S
d
ef
en
s
e
f
r
am
ewo
r
k
n
am
ed
an
al
y
tics
r
ea
d
in
ess
an
d
o
p
tim
al
m
atu
r
ity
ad
v
an
ce
m
e
n
t
(
Ar
OM
A)
,
en
c
o
m
p
ass
in
g
b
u
s
in
ess
m
o
n
ito
r
in
g
,
a
n
o
m
aly
d
is
co
v
er
y
,
a
n
d
m
itig
atio
n
,
was
p
r
o
p
o
s
ed
.
Ar
OM
A
d
em
o
n
s
tr
ated
th
e
ca
p
ab
ilit
y
to
f
ac
ilit
ate
co
llab
o
r
atio
n
s
b
etw
ee
n
in
ter
n
et
s
er
v
ice
p
r
o
v
i
d
er
s
(
I
SP
s
)
an
d
th
eir
g
u
ests
in
th
e
r
ea
lm
o
f
DDo
S
m
itig
atio
n
.
C
u
i
et
a
l
.
[2
5
]
in
tr
o
d
u
ce
d
a
co
m
p
r
eh
en
s
iv
e
s
y
s
tem
co
m
p
r
is
in
g
f
o
u
r
cr
u
cial
m
o
d
u
les
-
attac
k
d
is
co
v
er
y
d
etec
to
r
,
attac
k
d
is
co
v
er
y
,
attac
k
tr
ac
eb
ac
k
,
an
d
attac
k
m
itig
atio
n
.
T
h
e
e
x
is
tin
g
liter
atu
r
e
o
n
ly
f
o
cu
s
es
o
n
v
er
y
f
ew
ML
alg
o
r
ith
m
s
,
wh
ile
o
u
r
wo
r
k
co
n
s
id
er
s
p
er
f
o
r
m
an
ce
an
aly
s
is
o
f
all
ML
alg
o
r
ith
m
s
an
d
also
u
s
es
s
y
n
th
etic
m
in
o
r
ity
o
v
e
r
-
s
am
p
lin
g
tech
n
iq
u
e
(
SMOT
E
)
tech
n
iq
u
e
f
o
r
f
ea
tu
r
e
en
g
in
ee
r
i
n
g
an
d
ad
d
r
ess
in
g
class
i
f
m
b
alan
ce
.
Ad
d
itio
n
ally
,
th
e
p
a
p
er
i
m
p
lem
en
ts
a
DDo
S
attac
k
d
is
co
v
er
y
s
y
s
tem
b
ase
d
o
n
n
eu
r
al
n
etwo
r
k
s
f
o
r
d
ir
ec
t
attac
k
id
en
tific
atio
n
an
d
i
n
t
r
o
d
u
ce
s
a
n
attac
k
tr
ac
eb
ac
k
s
y
s
tem
s
tr
ateg
ically
lev
er
ag
in
g
th
e
d
is
tin
ctiv
e
ch
a
r
ac
ter
is
tics
o
f
SDN.
I
n
th
is
w
o
r
k
,
th
e
m
ain
f
o
cu
s
is
DDo
S
attac
k
p
r
e
d
ictio
n
in
SDN
en
v
ir
o
n
m
e
n
ts
.
I
n
th
is
e
n
d
ea
v
o
r
,
th
is
wo
r
k
m
ak
es
c
o
n
tr
ib
u
tio
n
s
o
n
v
ar
io
u
s
f
r
o
n
ts
:
E
x
p
lo
r
ato
r
y
d
ata
a
n
aly
s
is
(
E
D
A)
:
a
th
o
r
o
u
g
h
in
s
p
ec
tio
n
o
f
th
e
d
ataset
co
n
s
is
tin
g
o
f
a
r
an
g
e
o
f
f
ea
t
u
r
es
p
er
tin
en
t
t
o
n
etwo
r
k
tr
af
f
ic.
E
ac
h
f
ea
tu
r
e
is
e
x
am
in
ed
f
o
r
it
s
d
ata
ty
p
e
,
r
a
n
g
e,
d
is
tr
ib
u
tio
n
,
an
d
p
r
esen
ce
o
f
m
is
s
in
g
o
r
n
u
ll v
al
u
es,
an
d
a
n
o
m
alo
u
s
p
o
i
n
ts
.
‒
Data
clea
n
in
g
:
d
ep
en
d
i
n
g
o
n
th
e
n
atu
r
e
an
d
im
p
o
r
tan
ce
o
f
th
e
m
is
s
in
g
v
alu
es,
tech
n
iq
u
es
s
u
ch
as
im
p
u
tatio
n
o
r
ex
clu
s
io
n
ar
e
em
p
l
o
y
ed
t
o
ad
d
r
ess
th
em
.
‒
Statis
t
ical
m
ea
s
u
r
es
s
u
ch
as
ce
n
tr
al
ten
d
en
c
y
(
m
ea
n
,
m
ed
ian
)
an
d
d
is
p
er
s
io
n
(
s
tan
d
ar
d
d
e
v
iatio
n
,
in
ter
q
u
ar
tile
r
an
g
e
)
ar
e
ca
lcu
l
ated
to
p
r
o
v
id
e
a
h
ig
h
-
lev
el
u
n
d
er
s
tan
d
in
g
o
f
t
h
e
d
is
tr
ib
u
tio
n
an
d
s
p
r
ea
d
o
f
th
e
d
ata.
Fo
r
in
s
tan
ce
,
if
th
e
m
ea
n
p
ac
k
et
co
u
n
t
is
n
o
tab
ly
h
ig
h
er
th
an
th
e
m
ed
ian
,
it
s
u
g
g
ests
a
s
k
ewe
d
d
is
tr
ib
u
tio
n
,
p
o
ten
tially
d
u
e
to
th
e
p
r
esen
ce
o
f
o
u
tlier
s
o
r
h
ea
v
y
tr
af
f
ic
e
p
is
o
d
es.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell,
Vo
l.
1
3
,
No
.
4
,
Dec
em
b
er
2
0
2
4
:
4
4
8
8
-
4
4
9
8
4490
‒
Vis
u
aliza
tio
n
is
ca
r
r
ied
o
u
t
u
s
in
g
th
e
f
o
llo
win
g
p
l
o
ts
:
i)
b
o
x
p
lo
ts
to
d
e
p
ict
d
ata
d
is
tr
ib
u
tio
n
,
aid
in
g
o
u
tlier
d
etec
tio
n
,
p
o
ten
tially
in
d
icati
n
g
DDo
S
attac
k
s
in
th
is
s
tu
d
y
;
ii)
s
ca
tter
p
lo
ts
to
v
is
u
ali
ze
r
elatio
n
s
h
ip
s
b
etwe
en
co
n
tin
u
o
u
s
v
ar
iab
les,
h
in
tin
g
at
co
r
r
elate
d
f
ea
tu
r
es,
in
d
icativ
e
o
f
s
p
ec
if
ic
n
etwo
r
k
b
eh
av
i
o
r
;
an
d
b
ar
ch
ar
ts
to
d
is
p
lay
ca
te
g
o
r
ica
l
d
ata
d
is
tr
ib
u
tio
n
s
,
cr
u
cial
f
o
r
u
n
d
er
s
tan
d
i
n
g
th
e
f
r
eq
u
e
n
cy
o
f
v
ar
iab
les
s
u
ch
as p
r
o
to
co
l ty
p
es o
r
p
o
r
t n
u
m
b
er
s
.
‒
Giv
en
th
e
ty
p
ical
class
im
b
alan
ce
in
DDo
S
d
atasets
,
th
e
SMOT
E
was
u
s
ed
t
o
g
en
er
ate
s
y
n
th
etic
s
am
p
les
o
f
th
e
m
in
o
r
ity
class
.
T
h
is
b
al
an
ce
d
th
e
d
ataset,
th
u
s
p
r
ev
en
t
in
g
th
e
m
o
d
els f
r
o
m
b
ein
g
b
ia
s
ed
to
war
d
s
th
e
m
ajo
r
ity
class
.
‒
W
e
also
u
tili
ze
d
co
r
r
elatio
n
a
n
aly
s
is
to
id
en
tify
lin
ea
r
r
elat
io
n
s
h
ip
s
b
etwe
en
f
ea
tu
r
es,
w
h
ile
th
e
r
a
n
d
o
m
f
o
r
est
alg
o
r
ith
m
ass
ess
ed
f
ea
tu
r
e
im
p
o
r
ta
n
ce
,
en
ab
lin
g
a
s
elec
tio
n
o
f
th
e
m
o
s
t
r
elev
an
t
f
ea
tu
r
es
f
o
r
th
e
m
o
d
els.
‒
A
d
iv
er
s
e
s
u
ite
o
f
ML
alg
o
r
ith
m
s
was
s
elec
ted
to
class
if
y
wh
eth
er
th
e
tr
af
f
ic
is
n
o
r
m
al
o
r
attac
k
ed
.
Alg
o
r
ith
m
s
s
u
ch
as
g
r
a
d
ien
t
b
o
o
s
tin
g
an
d
en
s
em
b
le
m
o
d
els:
as
lig
h
t
g
r
ad
ie
n
t
b
o
o
s
tin
g
m
a
ch
in
e
(
L
GB
M)
class
if
ier
,
r
an
d
o
m
f
o
r
est
class
i
f
ier
,
XGB
class
if
ier
,
ex
tr
a
t
r
ee
s
class
if
ier
,
an
d
Ad
aBo
o
s
t
cla
s
s
if
ier
.
DT
:
DT
class
if
ier
an
d
ex
tr
a
tr
ee
clas
s
if
i
er
.
Pro
x
im
ity
-
b
ased
alg
o
r
ith
m
s
:
K
-
n
eig
h
b
o
r
s
class
if
ier
.
SVM
:
s
u
p
p
o
r
t
v
ec
to
r
class
if
ier
(
SVC
)
.
Sto
ch
asti
c
g
r
ad
ien
t
d
escen
t:
SGD
class
if
ier
.
L
in
ea
r
m
o
d
els:
l
o
g
is
tic
r
e
g
r
ess
io
n
,
lin
ea
r
d
is
cr
im
in
an
t
an
aly
s
is
,
r
id
g
e
c
lass
if
ier
C
V,
an
d
r
id
g
e
class
if
ier
.
B
ay
esian
m
eth
o
d
s
:
Gau
s
s
ian
NB
.
O
th
e
r
class
if
ier
s
:
p
er
ce
p
tr
o
n
,
n
ea
r
est
ce
n
tr
o
id
,
p
ass
iv
e
a
g
g
r
ess
iv
e
c
lass
if
ier
,
q
u
ad
r
atic
d
is
cr
im
in
an
t
an
aly
s
is
,
an
d
d
u
m
m
y
class
if
ier
.
‒
T
h
e
k
ey
f
in
d
i
n
g
s
o
f
th
e
r
esear
ch
ar
e
to
tu
n
e
ea
ch
m
o
d
el
m
eti
cu
lo
u
s
ly
an
d
v
alid
ate
u
s
in
g
tec
h
n
iq
u
es
s
u
ch
as
cr
o
s
s
-
v
alid
atio
n
to
m
itig
ate
o
v
er
f
itti
n
g
an
d
en
s
u
r
e
th
e
r
o
b
u
s
tn
ess
o
f
th
e
p
r
e
d
ictio
n
s
.
T
h
e
r
em
ain
in
g
s
ec
tio
n
s
o
f
t
h
e
p
ap
er
c
o
v
er
t
h
e
f
o
llo
win
g
to
p
ics:
s
ec
tio
n
2
elab
o
r
ates
o
n
t
h
e
d
atasets
u
tili
ze
d
,
s
o
u
r
ce
d
f
r
o
m
v
ar
io
u
s
d
atab
ases
.
Sectio
n
3
o
u
tlin
es
th
e
s
ch
em
atic
d
iag
r
am
d
et
ailin
g
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
l
o
g
y
f
o
r
p
r
e
d
ictin
g
DDo
S
attac
k
s
in
th
e
SDN
en
v
ir
o
n
m
en
t.
Sectio
n
4
d
el
v
es
in
to
s
ig
n
if
ican
t
p
er
f
o
r
m
an
ce
m
ea
s
u
r
es,
r
esu
lts
,
an
d
c
o
m
p
ar
is
o
n
s
with
s
tate
-
of
-
th
e
-
a
r
t
ap
p
r
o
ac
h
es.
L
astl
y
,
s
ec
tio
n
5
co
n
cl
u
d
es
th
e
p
ap
er
a
n
d
o
u
tlin
es d
ir
ec
tio
n
s
f
o
r
f
u
tu
r
e
r
esear
c
h
.
2.
M
E
T
H
O
D
2
.
1
.
I
np
ut
da
t
a
s
et
Fo
r
th
e
DDo
S
attac
k
d
etec
tio
n
,
at
f
ir
s
t,
SDN
d
ataset
is
co
l
lecte
d
f
r
o
m
[2
6
]
a
n
d
p
r
ep
r
o
c
ess
ed
.
T
h
e
d
a
t
a
s
et
c
o
n
t
a
i
n
s
1
0
4
3
4
6
r
o
w
s
an
d
2
3
c
o
l
u
m
n
s
,
w
h
i
c
h
a
r
e
a
m
i
x
o
f
f
e
a
t
u
r
e
s
a
n
d
l
a
b
e
l
s
,
a
s
u
b
s
ta
n
t
i
a
l
a
n
d
v
a
l
u
a
b
l
e
r
e
s
o
u
r
c
e
f
o
r
r
e
s
e
a
r
c
h
i
n
g
a
n
d
d
ev
e
l
o
p
i
n
g
D
D
o
S
d
e
t
ec
t
i
o
n
m
e
t
h
o
d
s
.
T
h
e
s
e
c
o
l
u
m
n
s
c
o
u
l
d
b
e
d
i
v
i
d
e
d
i
n
t
o
:
‒
Featu
r
es:
th
ese
ar
e
th
e
attr
ib
u
tes
u
s
ed
to
d
escr
ib
e
n
etwo
r
k
tr
a
f
f
ic
p
atter
n
s
an
d
c
h
ar
ac
ter
is
tics
.
T
h
ese
f
ea
tu
r
es
h
elp
ML
m
o
d
els lea
r
n
to
d
is
ti
n
g
u
is
h
b
etwe
en
n
o
r
m
al
an
d
DDo
S tr
af
f
ic.
So
m
e
co
m
m
o
n
f
e
atu
r
es f
o
r
DDo
S
d
etec
tio
n
m
ig
h
t
in
clu
d
e
p
ac
k
et
s
izes,
p
ac
k
et
r
ates,
b
y
te
r
ates,
s
o
u
r
ce
I
P
ad
d
r
ess
es,
d
esti
n
atio
n
I
P
ad
d
r
ess
es
,
an
d
v
ar
i
o
u
s
s
tatis
tical
m
etr
ics
r
elate
d
to
n
etwo
r
k
tr
af
f
ic.
‒
L
ab
els:
th
is
co
lu
m
n
i
n
d
icate
s
wh
eth
er
ea
c
h
n
etwo
r
k
tr
a
f
f
ic
r
ec
o
r
d
f
r
e
p
r
esen
ts
n
o
r
m
al
tr
a
f
f
ic
o
r
a
DDo
S
attac
k
.
T
h
e
lab
el
c
o
lu
m
n
is
ess
en
tial
f
o
r
s
u
p
e
r
v
is
ed
lear
n
in
g
,
wh
er
e
m
o
d
els
lear
n
f
r
o
m
lab
el
ed
d
ata
as
s
h
o
wn
in
Fig
u
r
e
1
.
‒
Data
s
p
lit:
wh
en
u
s
in
g
th
is
d
ataset
f
o
r
ML
,
it'
s
co
m
m
o
n
to
s
p
lit
it
in
to
tr
ain
in
g
,
v
alid
atio
n
,
an
d
test
s
ets
to
ev
alu
ate
m
o
d
el
p
er
f
o
r
m
an
ce
p
r
o
p
er
l
y
.
Her
e
7
5
%
is
u
s
ed
f
o
r
tr
ain
in
g
an
d
2
5
%
o
f
th
e
d
a
taset
is
u
s
ed
f
o
r
test
in
g
an
d
v
alid
atio
n
.
Fig
u
r
e
1
.
Data
s
et
with
v
ar
io
u
s
f
ea
tu
r
es h
av
in
g
lab
eled
i
n
f
o
r
m
atio
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2
2
5
2
-
8
9
3
8
E
va
lu
a
tio
n
o
f d
is
tr
ib
u
ted
d
en
i
a
l o
f ser
vice
a
tta
ck
s
d
etec
tio
n
in
s
o
ftw
a
r
e
d
efin
ed
n
etw
o
r
ks
(
N
ee
th
u
S
.
)
4491
2
.
2
.
F
r
a
m
ewo
r
k
T
h
is
s
tu
d
y
d
elin
ea
tes
a
co
m
p
r
eh
en
s
iv
e
ap
p
r
o
ac
h
to
p
r
ed
ict
S
DN
-
DDo
S
attac
k
s
b
y
em
p
lo
y
i
n
g
a
s
in
g
le
p
h
ased
m
eth
o
d
o
lo
g
ical
f
r
am
ewo
r
k
th
at
in
teg
r
ates
v
ar
io
u
s
ML
alg
o
r
ith
m
s
in
to
lazy
class
if
ier
s
.
T
h
e
m
eth
o
d
o
l
o
g
y
aim
s
to
lev
e
r
ag
e
th
e
s
tr
en
g
th
s
o
f
ea
c
h
alg
o
r
ith
m
ic
ca
teg
o
r
y
t
o
en
h
a
n
ce
p
r
e
d
i
ctiv
e
ac
cu
r
ac
y
a
n
d
g
ain
d
ee
p
e
r
in
s
ig
h
ts
in
to
th
e
c
h
ar
ac
ter
is
tics
o
f
n
etwo
r
k
tr
af
f
i
c
d
ata
th
at
co
u
ld
p
o
ten
tially
s
ig
n
al
th
e
p
r
esen
ce
o
f
DDo
S
attac
k
s
.
Fi
g
u
r
e
2
d
ep
icts
th
e
o
v
er
all
s
tag
es
in
th
e
p
r
o
p
o
s
ed
wo
r
k
.
T
h
e
s
u
b
s
eq
u
en
t
s
ec
tio
n
s
p
r
o
v
id
e
co
n
cise d
escr
ip
tio
n
s
o
f
t
h
e
ele
m
en
ts
co
n
s
titu
tin
g
th
e
m
o
d
els.
Fig
u
r
e
2
.
A
s
ch
em
atic
d
iag
r
a
m
o
f
th
e
m
eth
o
d
o
lo
g
y
2
.
2
.
1
.
E
x
plo
ra
t
o
ry
da
t
a
a
na
l
y
s
is
T
h
e
s
tatis
tical
s
u
m
m
ar
y
o
f
th
e
d
ataset
is
s
h
o
wn
in
T
ab
le
s
1
an
d
2
.
T
h
e
d
ataset
c
o
m
p
r
i
s
es
8
,
7
1
4
o
b
s
er
v
atio
n
s
ac
r
o
s
s
v
ar
io
u
s
v
a
r
iab
les,
s
h
ed
d
in
g
lig
h
t
o
n
n
etwo
r
k
tr
a
f
f
ic
d
y
n
am
ics.
T
h
e
'
s
wit
ch
'
v
ar
iab
le,
lik
ely
r
ep
r
esen
tin
g
n
etwo
r
k
s
witch
id
en
tifie
r
s
,
h
as
a
m
ea
n
v
alu
e
o
f
1
1
,
0
8
9
.
8
3
,
s
u
g
g
esti
n
g
p
o
ten
tial
en
co
d
in
g
o
r
s
ca
lin
g
,
as swit
ch
id
en
tifie
r
s
ar
e
ty
p
ically
ca
teg
o
r
ical.
T
ab
le
1
.
Data
s
et
p
k
t
c
o
u
n
t
b
y
t
e
c
o
u
n
t
dur
d
u
r
_
n
sec
t
o
t
_
d
u
r
p
a
c
k
e
t
i
n
s
p
k
t
p
e
r
f
l
o
w
c
o
u
n
t
1
4
.
0
0
8
,
7
1
4
.
0
0
8
,
7
1
4
.
0
0
8
,
7
1
4
.
0
0
8
,
7
1
4
.
0
0
8
,
7
1
4
.
0
0
8
,
7
1
4
.
0
0
mea
n
7
.
4
7
E+
0
4
7
.
9
1
E+
0
7
1
9
4
.
0
1
5
.
4
8
E+
0
8
1
.
9
5
E+
1
1
1
,
9
2
0
.
4
2
1
.
0
5
E+
0
4
st
d
4
.
0
7
E+
0
4
4
.
3
1
E+
0
7
1
1
7
.
3
5
2
.
3
2
E+
0
8
1
.
1
7
E+
1
1
2
5
4
.
1
3
4
.
0
5
E+
0
3
mi
n
2
8
4
.
0
0
3
.
0
3
E+
0
5
-
7
.
9
0
E+
0
7
8
.
3
7
E+
0
8
5
5
8
.
0
0
0
.
0
0
E+
0
0
2
5
%
3
.
7
3
E+
0
4
3
.
9
6
E+
0
7
1
0
0
.
0
0
3
.
9
1
E+
0
8
1
.
0
1
E+
1
1
1
,
9
3
1
.
0
0
8
.
6
4
E+
0
3
5
0
%
7
.
6
6
E+
0
4
8
.
1
7
E+
0
7
1
9
0
.
0
0
5
.
5
6
E+
0
8
1
.
9
1
E+
1
1
1
,
9
4
3
.
0
0
1
.
3
4
E+
0
4
7
5
%
1
.
1
3
E+
0
5
1
.
2
0
E+
0
8
2
8
0
.
0
0
7
.
2
6
E+
0
8
2
.
8
1
E+
1
1
1
,
9
4
3
.
0
0
1
.
3
5
E+
0
4
max
1
.
3
5
E+
0
5
1
.
4
4
E+
0
8
4
7
3
.
0
0
9
.
1
4
E+
0
8
4
.
7
3
E+
1
1
2
,
2
4
2
.
0
0
1
.
3
7
E+
0
4
T
ab
le
2
.
Featu
r
e
s
elec
tio
n
i
n
d
ataset
b
y
t
e
p
e
r
f
l
o
w
p
k
t
r
a
t
e
t
x
_
b
y
t
e
s
r
x
_
b
y
t
e
s
t
x
_
k
b
p
s
r
x
_
k
b
p
s
t
o
t
_
k
b
p
s
c
o
u
n
t
8
,
7
1
4
.
0
0
8
,
7
1
4
.
0
0
8
,
7
1
4
.
0
0
8
,
7
1
4
.
0
0
8
,
7
1
4
.
0
0
8
,
7
1
4
.
0
0
8
,
7
1
4
.
0
0
mea
n
1
.
1
1
E+
0
7
3
4
9
.
8
6
4
.
8
2
E+
0
7
4
.
8
1
E+
0
7
8
7
3
.
4
3
8
7
3
.
1
1
1
,
7
4
6
.
5
4
st
d
4
.
3
5
E+
0
6
1
3
4
.
9
1
1
.
5
1
E+
0
8
1
.
1
0
E+
0
8
2
,
8
4
8
.
9
7
2
,
2
6
9
.
0
7
3
,
4
2
6
.
3
6
mi
n
0
.
0
0
E+
0
0
-
2
.
8
5
E+
0
3
9
.
2
6
E+
0
2
-
-
2
5
%
9
.
2
1
E+
0
6
2
8
8
.
0
0
3
.
5
9
E+
0
3
1
.
4
7
E+
0
3
-
-
5
0
%
1
.
4
3
E+
0
7
4
4
6
.
0
0
3
.
8
4
E+
0
3
3
.
5
4
E+
0
3
-
-
7
5
%
1
.
4
4
E+
0
7
4
5
1
.
0
0
4
.
2
5
E+
0
3
6
.
1
6
E+
0
6
-
-
2
.
5
7
E+
0
3
max
1
.
4
6
E+
0
7
4
5
6
.
0
0
1
.
2
7
E+
0
9
9
.
9
1
E+
0
8
2
.
0
6
E+
0
4
1
.
6
6
E+
0
4
2
.
0
6
E+
0
4
'
Pk
tco
u
n
t'
an
d
'
b
y
teco
u
n
t'
s
ig
n
if
y
p
ac
k
et
an
d
b
y
te
co
u
n
ts
p
er
n
etwo
r
k
ev
e
n
t,
with
'
p
k
tco
u
n
t'
av
er
ag
in
g
at
2
.
3
2
,
in
d
icatin
g
co
n
s
is
ten
c
y
ac
r
o
s
s
o
b
s
er
v
atio
n
s
.
Me
an
wh
ile,
'
b
y
teco
u
n
t'
av
er
ag
es
7
9
,
0
6
0
,
7
7
6
.
5
4
b
y
tes,
r
ef
lectin
g
v
a
r
y
in
g
ev
en
t
s
izes.
Du
r
atio
n
v
a
r
iab
les
'
d
u
r
'
,
'
d
u
r
_
n
s
ec
'
,
an
d
'
to
t_
d
u
r
'
d
ep
ict
ev
en
t
len
g
th
s
in
s
ec
o
n
d
s
,
n
an
o
s
ec
o
n
d
s
,
an
d
to
tal
d
u
r
atio
n
,
with
m
ea
n
v
alu
es in
d
icatin
g
a
wid
e
r
an
g
e
o
f
e
v
en
t d
u
r
atio
n
s
.
'
Flo
w
s
'
av
er
ag
e
3
.
3
2
p
e
r
ev
en
t,
in
d
icatin
g
s
im
u
ltan
eo
u
s
co
m
m
u
n
icatio
n
s
ess
io
n
s
,
with
co
n
s
is
ten
t
f
lo
w
n
u
m
b
er
s
.
'
Pack
etin
s
'
,
p
o
s
s
ib
ly
in
d
icatin
g
in
co
m
i
n
g
p
ac
k
ets,
av
er
a
g
es
1
,
9
2
0
.
4
2
,
h
i
n
tin
g
at
v
ar
iab
ilit
y
in
in
co
m
i
n
g
p
ac
k
et
f
r
eq
u
en
cy
.
'
Pk
tp
er
f
lo
w'
an
d
'
b
y
tep
er
f
lo
w'
s
h
o
w
p
ac
k
et
an
d
b
y
te
co
u
n
ts
p
er
f
l
o
w,
with
d
iv
er
s
e
av
e
r
ag
e
s
an
d
h
ig
h
s
tan
d
ar
d
d
ev
iatio
n
s
,
cr
u
cial
f
o
r
tr
af
f
ic
p
r
o
f
ilin
g
.
'
Pk
tr
ate'
av
er
ag
es
3
4
9
.
8
6
,
with
a
s
ig
n
if
ican
t
s
tan
d
ar
d
d
e
v
iatio
n
,
s
u
g
g
esti
n
g
a
wid
e
r
a
n
g
e
in
p
ac
k
et
tr
an
s
m
is
s
io
n
r
ates,
p
o
ten
tially
r
e
f
lectin
g
i
d
le
p
e
r
io
d
s
o
r
k
ee
p
-
aliv
e
co
n
n
ec
tio
n
s
.
2
.
2
.
2
.
Vis
ua
liza
t
io
n
T
h
e
f
o
llo
win
g
v
is
u
aliza
tio
n
s
wer
e
o
b
s
er
v
e
d
h
e
r
e.
T
h
e
ch
ar
t
is
a
p
air
s
p
lo
t
m
atr
ix
t
h
at
v
is
u
alize
s
th
e
r
elatio
n
s
h
ip
s
b
etwe
en
s
ev
er
al
ca
teg
o
r
ical
v
a
r
iab
les:
'
s
w
itch
'
,
'
f
lo
ws
'
,
'
p
air
f
lo
w
'
,
an
d
'
p
o
r
t_
n
o
'
,
alo
n
g
with
th
eir
d
is
tr
ib
u
tio
n
with
r
esp
ec
t
to
tw
o
lab
els
(
0
an
d
1
,
w
h
ich
c
o
u
l
d
r
ep
r
esen
t
d
if
f
e
r
en
t
class
es
s
u
ch
as
'
n
o
r
m
al'
an
d
'
attac
k
'
in
th
e
co
n
tex
t
o
f
n
etwo
r
k
tr
af
f
ic)
.
L
o
o
k
i
n
g
at
th
e
d
iag
o
n
al,
we
s
ee
d
en
s
ity
p
lo
t
s
f
o
r
ea
ch
v
ar
iab
le,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell,
Vo
l.
1
3
,
No
.
4
,
Dec
em
b
er
2
0
2
4
:
4
4
8
8
-
4
4
9
8
4492
co
lo
r
ed
d
if
f
er
en
tly
f
o
r
ea
ch
la
b
el,
in
d
icatin
g
th
e
d
is
tr
ib
u
tio
n
o
f
ea
ch
v
ar
ia
b
le
with
in
th
e
two
class
es.
I
t
ap
p
ea
r
s
th
at
th
e
v
ar
iab
les
h
a
v
e
d
is
cr
ete
v
alu
es
wh
ich
m
ig
h
t
r
ep
r
esen
t
c
ateg
o
r
ical
o
r
b
in
n
e
d
n
u
m
er
ical
d
ata.
Fo
r
in
s
tan
ce
,
'
s
witch
'
s
h
o
ws
a
b
im
o
d
al
d
is
tr
ib
u
tio
n
f
o
r
o
n
e
lab
el
an
d
a
m
o
r
e
u
n
if
o
r
m
d
is
tr
ib
u
tio
n
f
o
r
th
e
o
th
er
,
s
u
g
g
esti
n
g
th
at
ce
r
tain
s
witch
v
alu
es a
r
e
m
o
r
e
p
r
ev
alen
t in
o
n
e
class
th
an
in
th
e
o
th
er
.
I
n
th
e
o
f
f
-
d
iag
o
n
al
p
lo
ts
,
wh
ich
s
h
o
w
th
e
r
elatio
n
s
h
ip
b
etwe
en
two
v
ar
iab
les,
we
s
ee
d
is
tin
ct
g
r
o
u
p
in
g
s
o
f
p
o
in
ts
.
T
h
ese
s
c
atter
p
lo
ts
r
e
v
ea
l
h
o
w
t
h
e
ca
te
g
o
r
ies
o
f
ea
c
h
v
ar
iab
le
a
r
e
d
is
tr
ib
u
ted
c
o
n
ce
r
n
in
g
ea
ch
o
th
er
.
Fo
r
ex
am
p
le,
th
e
'
s
witch
'
v
s
.
'
f
lo
ws
'
p
lo
t
s
h
o
ws
th
at
ce
r
tain
s
w
itch
es
ar
e
ass
o
ciate
d
with
s
p
ec
if
ic
f
lo
w
co
u
n
ts
,
an
d
th
is
ass
o
ciatio
n
d
if
f
e
r
s
b
y
lab
el,
w
h
ich
co
u
ld
b
e
in
d
icativ
e
o
f
ce
r
tain
p
atter
n
s
o
f
n
etwo
r
k
b
e
h
a
v
i
o
r
t
h
a
t
a
r
e
c
h
a
r
a
ct
e
r
is
t
i
c
o
f
n
o
r
m
a
l
v
e
r
s
u
s
at
t
a
c
k
t
r
a
f
f
ic
.
T
h
e
p
l
o
ts
w
i
t
h
'
p
a
i
r
f
l
o
w'
a
g
a
i
n
s
t
o
t
h
e
r
v
a
r
i
a
b
l
es
e
x
h
i
b
i
t
a
r
e
la
t
i
v
el
y
t
i
g
h
t
cl
u
s
te
r
i
n
g
o
f
p
o
i
n
t
s
a
t
l
o
w
e
r
v
al
u
e
s
o
f
'
p
a
i
r
f
l
o
w'
,
es
p
e
c
i
al
l
y
f
o
r
o
n
e
l
a
b
e
l
,
p
e
r
h
a
p
s
i
n
d
i
c
a
t
i
n
g
t
h
at
l
o
w
'
p
ai
r
f
l
o
w'
v
a
l
u
e
s
a
r
e
s
t
r
o
n
g
l
y
as
s
o
c
ia
t
e
d
wi
t
h
a
p
a
r
t
i
c
u
la
r
c
l
as
s
a
s
d
e
p
i
ct
ed
i
n
F
i
g
u
r
e
3
.
Fin
ally
,
th
e
'
p
o
r
t_
n
o
'
v
ar
iab
le
s
h
o
ws
m
u
ltip
le
p
ea
k
s
in
d
is
tr
i
b
u
tio
n
,
s
u
g
g
esti
n
g
th
at
th
e
r
e
a
r
e
co
m
m
o
n
p
o
r
ts
th
at
a
r
e
f
r
eq
u
e
n
tly
u
s
ed
in
b
o
th
class
es,
b
u
t
th
er
e'
s
a
v
is
ib
le
d
is
tin
ctio
n
in
th
e
d
is
tr
ib
u
tio
n
o
f
'
p
o
r
t_
n
o
'
b
etwe
en
th
e
lab
els.
T
h
e
'
s
witc
h
'
v
alu
e
co
u
n
t
in
d
icate
s
th
e
n
u
m
b
er
o
f
n
etwo
r
k
ev
en
ts
o
r
co
n
n
ec
tio
n
s
h
an
d
le
d
b
y
ea
ch
s
witch
id
en
tifie
r
.
W
ith
s
witch
3
h
an
d
lin
g
th
e
m
o
s
t
ev
en
ts
(
4
,
0
7
7
)
,
f
o
llo
wed
b
y
s
wi
tch
2
(
3
,
3
5
8
)
,
an
d
s
witch
1
(
1
,
2
8
0
)
,
th
er
e'
s
a
clea
r
in
d
icatio
n
th
at
s
o
m
e
s
witch
e
s
ar
e
m
o
r
e
h
ea
v
i
ly
u
tili
ze
d
th
a
n
o
th
er
s
.
T
h
is
co
u
ld
b
e
d
u
e
to
th
e
n
etwo
r
k
to
p
o
l
o
g
y
,
wh
er
e
ce
r
tain
s
witch
es
s
er
v
e
as
m
ajo
r
h
u
b
s
f
o
r
t
r
af
f
ic.
T
h
e
d
if
f
er
en
ce
s
in
co
u
n
ts
m
ig
h
t
also
r
ef
lect
th
e
s
tr
ateg
ic
p
o
s
itio
n
in
g
o
f
s
witch
es
with
in
th
e
n
etwo
r
k
,
o
r
v
ar
y
in
g
lev
els
o
f
p
r
o
tecti
on
ag
ain
s
t
DDo
S
attac
k
s
,
as
s
u
m
in
g
th
e
s
witch
es
ar
e
p
ar
t
o
f
an
SDN
en
v
ir
o
n
m
en
t
b
ein
g
m
o
n
ito
r
ed
f
o
r
s
ec
u
r
ity
p
u
r
p
o
s
es.
T
h
e
'
f
lo
ws
'
co
u
n
t
p
r
o
v
id
es
in
s
ig
h
ts
in
to
h
o
w
m
an
y
co
n
cu
r
r
e
n
t
n
etwo
r
k
f
l
o
ws
ar
e
co
m
m
o
n
with
in
th
e
ev
en
ts
ca
p
tu
r
ed
.
T
h
e
m
ajo
r
ity
o
f
ev
en
ts
h
a
v
e
b
etwe
en
2
an
d
4
f
lo
ws,
with
2
.
0
0
f
l
o
ws
b
ein
g
th
e
m
o
s
t
f
r
eq
u
e
n
t
(
2
,
5
6
9
tim
es).
T
h
is
s
u
g
g
ests
th
at
d
u
al
-
f
lo
w
co
m
m
u
n
icati
o
n
is
a
co
m
m
o
n
p
atter
n
,
p
o
s
s
ib
ly
in
d
icativ
e
o
f
b
id
ir
ec
tio
n
al
tr
af
f
ic
b
etwe
en
h
o
s
ts
.
T
h
e
p
r
esen
ce
o
f
a
s
in
g
le
n
o
t
a
n
u
m
b
er
(
NaN
)
v
alu
e
i
n
d
ic
ates
an
ev
en
t
wh
er
e
th
e
f
lo
w
co
u
n
t w
as n
o
t r
ec
o
r
d
ed
o
r
is
m
is
s
in
g
,
wh
ich
m
a
y
r
e
q
u
ir
e
ad
d
itio
n
al
d
ata
clea
n
i
n
g
o
r
im
p
u
tatio
n
.
T
h
e
'
p
air
f
lo
w'
co
u
n
t,
with
0
.
0
0
o
cc
u
r
r
in
g
8
,
7
1
4
tim
es,
s
u
g
g
ests
th
at
all
r
ec
o
r
d
ed
ev
e
n
ts
a
r
e
o
f
o
n
e
ty
p
e,
p
o
ten
tially
r
e
p
r
esen
tin
g
u
n
p
air
ed
f
lo
ws.
T
h
is
co
u
ld
s
ig
n
if
y
u
n
i
d
ir
ec
tio
n
al
c
o
m
m
u
n
icatio
n
o
r
non
-
r
esp
o
n
s
iv
e
tr
af
f
ic
,
wh
ich
m
ig
h
t
b
e
ch
a
r
ac
ter
is
tic
o
f
ce
r
t
ain
ty
p
es
o
f
n
etwo
r
k
b
e
h
av
io
r
o
r
ap
p
licatio
n
s
.
T
h
e
s
in
g
le
NaN
v
alu
e
in
d
icate
s
o
n
e
in
s
tan
ce
o
f
m
is
s
in
g
d
ata,
wh
i
ch
,
d
ep
e
n
d
in
g
o
n
t
h
e
co
n
te
x
t,
m
ig
h
t
b
e
n
e
g
lig
ib
le
o
r
co
u
ld
war
r
an
t f
u
r
th
er
in
v
est
ig
a
tio
n
to
en
s
u
r
e
d
ata
in
teg
r
ity
.
Fo
r
th
e
'
p
r
o
to
co
l'
v
ar
iab
l
e,
ev
er
y
s
in
g
le
c
o
u
n
te
d
ev
en
t
(
8
,
7
1
4
in
to
tal)
u
s
ed
th
e
u
s
er
d
atag
r
am
p
r
o
to
co
l
(
UDP
)
p
r
o
to
c
o
l.
T
h
is
u
n
if
o
r
m
ity
im
p
lies
th
at
th
e
d
ataset
m
ig
h
t
b
e
f
ilter
ed
t
o
ca
p
tu
r
e
o
n
ly
UDP
tr
af
f
ic,
o
r
th
at
UDP
is
o
v
er
wh
elm
in
g
ly
t
h
e
p
r
o
to
co
l
o
f
ch
o
ice
in
th
e
n
etwo
r
k
e
n
v
ir
o
n
m
en
t
u
n
d
er
s
tu
d
y
.
UDP'
s
co
n
n
ec
tio
n
less
n
atu
r
e
m
ak
es
it
a
f
r
eq
u
en
t
ch
o
ice
f
o
r
s
tr
ea
m
in
g
,
g
am
in
g
,
o
r
ce
r
tain
ty
p
es
o
f
DDo
S
attac
k
s
.
T
h
e
s
in
g
le
N
aN
v
alu
e
s
u
g
g
ests
an
i
n
s
tan
ce
wh
er
e
p
r
o
t
o
co
l
in
f
o
r
m
atio
n
is
m
is
s
in
g
.
T
h
e
'
p
o
r
t_
n
o
'
v
ar
iab
le
c
o
u
n
t
i
n
d
icate
s
th
e
n
u
m
b
e
r
o
f
n
etwo
r
k
ev
e
n
ts
ass
o
ciate
d
with
d
if
f
er
en
t
p
o
r
t
n
u
m
b
er
s
.
Po
r
ts
1
an
d
2
ar
e
n
ea
r
ly
eq
u
ally
r
ep
r
esen
ted
(
2
,
3
9
7
an
d
2
,
3
9
2
e
v
en
ts
,
r
esp
ec
tiv
ely
)
,
f
o
llo
wed
b
y
p
o
r
t
3
(
2
,
1
8
0
)
an
d
p
o
r
t
4
(
1
,
7
4
5
)
.
T
h
is
d
is
tr
ib
u
tio
n
co
u
l
d
r
ef
lect
th
e
u
s
ag
e
o
f
t
h
ese
p
o
r
ts
f
o
r
d
if
f
er
en
t
s
er
v
ices
o
r
ap
p
licatio
n
s
with
in
th
e
n
etwo
r
k
.
T
h
e
s
in
g
u
lar
NaN
v
alu
e
in
d
icate
s
o
n
e
ev
en
t
wh
er
e
th
e
p
o
r
t
n
u
m
b
e
r
was
n
o
t
r
ec
o
r
d
e
d
,
wh
ich
m
ay
n
ee
d
to
b
e
ad
d
r
ess
ed
d
u
r
i
n
g
d
ata
p
r
ep
r
o
ce
s
s
in
g
.
Ov
er
all,
t
h
is
p
ai
r
s
p
lo
t
m
atr
ix
is
a
p
o
wer
f
u
l
ex
p
lo
r
ato
r
y
to
o
l
th
at
allo
ws
u
s
to
s
ee
th
e
r
elatio
n
s
h
ip
s
b
etwe
en
d
if
f
er
en
t
p
air
s
o
f
v
ar
iab
les
an
d
h
o
w
th
ese
r
elatio
n
s
h
ip
s
m
ay
d
if
f
er
b
y
class
.
T
h
is
ca
n
p
r
o
v
i
d
e
v
al
u
ab
le
in
s
ig
h
ts
f
o
r
f
u
r
t
h
er
a
n
al
y
s
is
,
s
u
ch
as
f
ea
tu
r
e
s
elec
tio
n
o
r
an
o
m
aly
d
etec
tio
n
in
th
e
co
n
tex
t o
f
n
etwo
r
k
s
ec
u
r
ity
as d
ep
icted
i
n
Fig
u
r
e
3
.
Data
p
r
ep
r
o
ce
s
s
in
g
p
lay
ed
a
cr
itical
r
o
le
in
m
ain
tai
n
in
g
d
ataset
in
teg
r
ity
b
y
a
d
d
r
ess
in
g
m
is
s
in
g
o
r
n
u
l
l
v
alu
es.
T
h
is
en
s
u
r
ed
t
h
at
th
e
p
r
ed
ictiv
e
m
o
d
els
r
ec
eiv
ed
c
lean
an
d
co
m
p
lete
d
ata,
a
c
r
u
cial
f
ac
to
r
f
o
r
th
e
r
eliab
ilit
y
o
f
s
u
b
s
eq
u
en
t
an
a
ly
s
es.
On
e
-
h
o
t
en
co
d
in
g
wa
s
ap
p
lied
to
h
an
d
le
ca
teg
o
r
i
ca
l
v
ar
iab
les.
T
h
is
tr
an
s
f
o
r
m
atio
n
c
o
n
v
er
te
d
ca
teg
o
r
ical
d
ata
in
to
a
n
u
m
er
ic
al
f
o
r
m
at,
f
ac
ilit
atin
g
b
etter
u
n
d
er
s
tan
d
i
n
g
an
d
ev
alu
atio
n
o
f
in
p
u
t f
ea
tu
r
es b
y
ML
alg
o
r
ith
m
s
.
Featu
r
e
en
g
in
ee
r
i
n
g
,
a
p
iv
o
tal
s
tep
,
in
v
o
lv
ed
ap
p
l
y
in
g
d
o
m
ai
n
k
n
o
wled
g
e
to
c
r
ea
te
n
ew
f
ea
t
u
r
es
aim
ed
at
en
h
an
cin
g
m
o
d
el
p
er
f
o
r
m
an
ce
.
T
ec
h
n
i
q
u
es
s
u
ch
as
co
r
r
elatio
n
an
aly
s
is
id
en
tifie
d
l
in
ea
r
r
elatio
n
s
h
ip
s
b
etwe
en
f
ea
tu
r
es,
wh
ile
th
e
r
a
n
d
o
m
f
o
r
est
alg
o
r
ith
m
ass
ess
e
d
f
ea
tu
r
e
im
p
o
r
tan
ce
.
T
h
is
p
r
o
ce
s
s
allo
wed
f
o
r
t
h
e
s
elec
tio
n
o
f
t
h
e
m
o
s
t
r
elev
an
t
f
ea
tu
r
es,
r
ed
u
ci
n
g
d
im
en
s
io
n
ality
,
an
d
d
ir
ec
tin
g
th
e
m
o
d
els
to
war
d
s
th
e
m
o
s
t
p
r
ed
ictiv
e
attr
ib
u
tes.
T
o
a
d
d
r
e
s
s
clas
s
im
b
alan
ce
in
h
er
en
t
in
DDo
S
d
atasets
,
th
e
SM
OT
E
wa
s
em
p
lo
y
ed
.
T
h
is
tech
n
iq
u
e
g
e
n
er
ated
s
y
n
th
etic
s
am
p
les
o
f
th
e
m
in
o
r
ity
class
,
ef
f
ec
tiv
ely
b
alan
cin
g
th
e
d
ata
s
et
an
d
p
r
ev
en
tin
g
b
ias
to
war
d
s
th
e
m
ajo
r
ity
cla
s
s
.
A
d
iv
er
s
e
ar
r
ay
o
f
ML
al
g
o
r
ith
m
s
was
s
elec
ted
f
o
r
th
e
s
u
b
s
eq
u
en
t
p
h
ase,
aim
in
g
to
p
r
o
v
id
e
a
co
m
p
r
e
h
e
n
s
iv
e
p
er
s
p
ec
tiv
e
o
n
th
e
d
ataset.
T
h
e
s
elec
tio
n
en
co
m
p
ass
ed
v
ar
io
u
s
alg
o
r
ith
m
s
,
r
ef
lectin
g
d
if
f
er
en
t
ap
p
r
o
ac
h
e
s
an
d
m
et
h
o
d
o
lo
g
ies
in
o
r
d
er
t
o
ca
p
tu
r
e
th
e
co
m
p
lex
ity
a
n
d
n
u
an
ce
s
o
f
th
e
d
ata
.
I
n
s
u
m
m
a
r
y
,
th
is
co
m
p
r
e
h
en
s
iv
e
ap
p
r
o
ac
h
,
f
r
o
m
e
x
p
lo
r
ato
r
y
an
al
y
s
is
th
r
o
u
g
h
p
r
ep
r
o
c
ess
in
g
an
d
f
ea
tu
r
e
en
g
in
ee
r
in
g
to
th
e
s
elec
tio
n
o
f
d
iv
er
s
e
ML
m
o
d
els,
was
u
n
d
er
tak
e
n
to
en
s
u
r
e
a
r
o
b
u
s
t
an
d
well
-
in
f
o
r
m
ed
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2
2
5
2
-
8
9
3
8
E
va
lu
a
tio
n
o
f d
is
tr
ib
u
ted
d
en
i
a
l o
f ser
vice
a
tta
ck
s
d
etec
tio
n
in
s
o
ftw
a
r
e
d
efin
ed
n
etw
o
r
ks
(
N
ee
th
u
S
.
)
4493
f
o
u
n
d
atio
n
f
o
r
s
u
b
s
eq
u
en
t
p
h
ases
o
f
an
aly
s
is
an
d
m
o
d
el
d
e
v
elo
p
m
en
t.
E
ac
h
s
tep
was
s
tr
ateg
ically
ch
o
s
en
to
ad
d
r
ess
s
p
ec
if
ic
ch
allen
g
es a
n
d
en
h
a
n
ce
th
e
o
v
er
all
r
eliab
ilit
y
an
d
p
r
ed
ictiv
e
ca
p
ab
ilit
y
o
f
th
e
m
o
d
els.
Fig
u
r
e
3
.
Plo
t m
atr
i
x
v
is
u
alizin
g
th
e
r
elatio
n
s
h
ip
s
b
etwe
en
s
ev
er
al
ca
teg
o
r
ical
v
ar
iab
les:
'
s
witch
'
,
'
f
lo
ws
'
,
'
p
air
f
lo
w'
,
an
d
'
p
o
r
t_
n
o
'
,
alo
n
g
with
th
eir
d
is
tr
ib
u
tio
n
with
r
e
s
p
ec
t to
n
o
r
m
al
tr
a
f
f
ic
an
d
atta
ck
tr
af
f
ic
2
.
2
.
3
.
M
o
del dev
elo
pm
ent
Var
io
u
s
en
s
em
b
le
m
eth
o
d
s
,
in
clu
d
in
g
L
GB
M
class
if
ier
,
r
an
d
o
m
f
o
r
est cla
s
s
if
ier
,
XGB cla
s
s
if
ier
,
DT
class
if
ier
,
an
d
e
x
tr
a
t
r
ee
s
class
if
ier
,
as
well
as
p
o
p
u
lar
ML
class
if
ier
s
lik
e
Ad
aBo
o
s
t
class
if
ier
,
K
-
n
eig
h
b
o
r
s
class
if
ier
,
an
d
SVC
,
ar
e
b
ein
g
ex
p
lo
r
ed
f
o
r
tr
ain
i
n
g
m
o
d
els
to
d
etec
t
DDo
S
attac
k
s
.
T
h
e
d
ataset
is
d
iv
id
ed
in
to
7
5
%
f
o
r
tr
ain
in
g
an
d
2
5
%
f
o
r
test
in
g
an
d
v
alid
atio
n
.
E
n
s
e
m
b
le
tech
n
iq
u
es,
p
iv
o
tal
in
ML
,
in
v
o
l
v
e
g
r
a
d
ien
t
b
o
o
s
tin
g
,
wh
ic
h
elev
ates w
ea
k
lear
n
er
s
in
to
r
o
b
u
s
t p
r
ed
ictio
n
m
o
d
els.
T
h
e
L
GB
M
class
if
ier
,
ab
b
r
ev
iated
f
r
o
m
lig
h
t
g
r
a
d
ien
t
b
o
o
s
tin
g
m
ac
h
in
e
class
if
ier
,
u
tili
ze
s
DT
alg
o
r
ith
m
s
f
o
r
class
if
icatio
n
an
d
r
an
k
i
n
g
.
I
t
in
co
r
p
o
r
ates
g
r
ad
ien
t
-
b
ased
o
n
e
-
s
id
e
s
am
p
lin
g
(
GOSS)
an
d
ex
clu
s
iv
e
f
ea
tu
r
e
b
u
n
d
lin
g
(
E
FB
)
tech
n
iq
u
es
to
en
h
a
n
ce
e
f
f
icien
cy
with
lar
g
e
-
s
ca
le
d
ata,
b
o
o
s
tin
g
s
p
ee
d
a
n
d
r
ed
u
cin
g
m
e
m
o
r
y
u
s
ag
e.
R
an
d
o
m
f
o
r
est,
an
o
th
e
r
en
s
em
b
le
m
eth
o
d
,
ex
ce
ls
in
r
eg
r
ess
io
n
an
d
class
if
icatio
n
ta
s
k
s
b
y
clu
s
ter
in
g
DT
in
to
an
en
s
em
b
le,
co
m
b
atin
g
o
v
er
f
itti
n
g
a
n
d
lack
o
f
r
esil
ien
ce
.
XGBo
o
s
t,
o
r
e
x
tr
e
m
e
g
r
ad
ien
t
b
o
o
s
tin
g
,
is
r
en
o
wn
e
d
f
o
r
its
h
ig
h
p
er
f
o
r
m
a
n
ce
an
d
s
tr
ate
g
ies
to
m
itig
ate
o
v
er
f
itti
n
g
,
esp
ec
ially
ad
ep
t
at
h
an
d
lin
g
lar
g
e
d
atasets
an
d
p
ar
allel
p
r
o
ce
s
s
in
g
,
co
n
tr
ib
u
tin
g
to
its
wid
e
ad
o
p
tio
n
in
ML
.
E
x
tr
a
tr
ee
s
class
if
ier
,
ak
in
to
r
a
n
d
o
m
f
o
r
es
t,
cr
ea
tes
m
u
ltip
le
tr
ee
s
an
d
em
p
l
o
y
s
r
a
n
d
o
m
f
ea
tu
r
e
s
u
b
s
ets f
o
r
n
o
d
e
s
p
litt
in
g
.
Ho
wev
er
,
it
d
iv
er
g
es
b
y
f
o
r
g
o
in
g
b
o
o
t
s
tr
ap
p
in
g
an
d
o
p
tin
g
f
o
r
r
an
d
o
m
s
p
lits
,
em
p
h
asizin
g
its
u
n
iq
u
e
ap
p
r
o
ac
h
o
f
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell,
Vo
l.
1
3
,
No
.
4
,
Dec
em
b
er
2
0
2
4
:
4
4
8
8
-
4
4
9
8
4494
in
tr
o
d
u
cin
g
r
an
d
o
m
n
ess
in
to
t
h
e
tr
ee
-
b
u
ild
in
g
p
r
o
ce
s
s
.
Ad
a
B
o
o
s
t
co
llab
o
r
ates
m
u
ltip
le
we
ak
class
if
ier
s
to
f
o
r
m
a
r
o
b
u
s
t
en
s
em
b
le,
ad
ju
s
tin
g
th
eir
r
o
les
in
cr
em
en
tally
to
im
p
r
o
v
e
ac
cu
r
ac
y
,
p
a
r
ticu
lar
ly
in
ch
allen
g
in
g
s
ce
n
ar
io
s
.
DT
class
if
ier
s
ex
tr
a
ct
d
ec
is
io
n
-
m
ak
i
n
g
in
s
ig
h
ts
f
r
o
m
in
p
u
t
d
ata,
s
er
v
in
g
as
f
o
u
n
d
atio
n
al
to
o
ls
in
ML
.
T
h
e
K
-
n
eig
h
b
o
r
s
class
if
ier
,
a
p
ar
t
o
f
th
e
k
-
n
ea
r
est
n
eig
h
b
o
r
s
(
KNN)
f
am
ily
,
p
r
ed
icts
b
as
ed
o
n
th
e
m
ajo
r
ity
class
o
f
its
k
-
n
ea
r
est
n
eig
h
b
o
r
s
in
th
e
f
ea
t
u
r
e
s
et,
o
f
f
er
in
g
f
lex
ib
ilit
y
in
in
s
tan
ce
-
b
ased
lear
n
in
g
.
Su
p
p
o
r
t
v
ec
to
r
m
ac
h
in
es
(
SVM)
co
n
s
tr
u
ct
m
o
d
els
t
o
d
is
t
in
g
u
is
h
b
etwe
en
d
if
f
e
r
en
t
class
in
s
tan
ce
s
,
ass
u
m
in
g
lin
ea
r
s
ep
ar
ab
ilit
y
.
T
h
e
s
to
ch
a
s
tic
g
r
ad
ien
t
d
escen
t
(
SGD)
c
l
ass
if
ier
is
a
lin
ea
r
class
if
ier
o
p
tim
ized
u
s
in
g
SGD,
s
u
itab
le
f
o
r
s
ce
n
ar
io
s
with
s
u
b
s
tan
tial
d
ata
s
izes.
L
o
g
is
tic
r
eg
r
ess
io
n
p
r
ed
icts
b
i
n
ar
y
o
u
tco
m
es
u
s
in
g
co
n
tin
u
o
u
s
in
d
ep
en
d
en
t
v
a
r
iab
les,
wh
ile
l
in
ea
r
d
is
cr
im
in
an
t
a
n
aly
s
is
id
en
tifie
s
lin
ea
r
f
ea
tu
r
e
co
m
b
i
n
atio
n
s
to
ef
f
ec
tiv
ely
d
is
tin
g
u
is
h
class
es.
T
h
e
r
id
g
e
class
if
ier
,
tailo
r
ed
f
o
r
m
u
lti
-
cl
ass
class
if
icatio
n
task
s
,
in
co
r
p
o
r
ates
co
n
ce
p
ts
f
r
o
m
tr
ad
itio
n
al
class
if
icatio
n
m
eth
o
d
s
an
d
r
id
g
e
r
eg
r
ess
io
n
to
m
in
im
ize
o
v
er
f
itti
n
g
r
is
k
s
.
Gau
s
s
ian
NB
ex
ce
ls
in
class
if
icatio
n
task
s
w
ith
n
o
r
m
ally
d
is
tr
ib
u
ted
,
co
n
tin
u
o
u
s
d
ata,
ass
u
m
in
g
co
n
d
itio
n
al
i
n
d
ep
en
d
en
ce
am
o
n
g
f
ea
tu
r
es
with
in
ea
ch
class
.
Qu
ad
r
atic
d
is
cr
im
in
an
t
an
aly
s
is
a
llo
ws
d
is
tin
ct
co
v
ar
ian
ce
m
atr
ices
f
o
r
ea
ch
class
,
en
ab
lin
g
a
f
lex
ib
le
ca
p
t
u
r
e
o
f
n
o
n
-
lin
ea
r
d
ec
is
io
n
b
o
u
n
d
ar
ie
s
.
A
d
u
m
m
y
class
if
ier
s
er
v
es
as
a
b
asic
b
aselin
e
m
o
d
el,
co
n
tr
asti
n
g
with
m
o
r
e
co
m
p
lex
alg
o
r
ith
m
s
a
n
d
e
s
tab
lis
h
in
g
a
f
u
n
d
am
en
tal
r
e
f
er
en
ce
p
o
in
t
f
o
r
p
er
f
o
r
m
an
ce
ev
al
u
atio
n
.
2
.
2
.
4
.
P
er
f
o
r
m
a
nce
m
et
rics
T
h
e
ev
alu
atio
n
o
f
m
o
d
els
in
v
o
lv
ed
th
e
u
s
e
o
f
d
iv
e
r
s
e
m
etr
ics
ap
p
r
o
p
r
iate
f
o
r
class
if
icatio
n
task
s
,
en
co
m
p
ass
in
g
ac
cu
r
ac
y
,
b
ala
n
ce
d
ac
cu
r
ac
y
,
r
ec
eiv
er
o
p
e
r
atin
g
ch
ar
ac
ter
is
tic
(
R
OC
)
-
ar
ea
u
n
d
er
th
e
cu
r
v
e
(
AUC)
,
an
d
F1
-
s
co
r
e.
T
h
ese
m
etr
ics
co
llectiv
ely
o
f
f
er
e
d
a
co
m
p
r
eh
e
n
s
iv
e
ass
ess
m
en
t
o
f
th
e
p
er
f
o
r
m
a
n
ce
o
f
ea
ch
m
o
d
el.
Acc
u
r
ac
y
is
d
eter
m
in
ed
as
th
e
r
atio
o
f
co
r
r
ec
t
p
r
ed
ictio
n
s
to
th
e
to
tal
n
u
m
b
er
o
f
p
r
ed
ictio
n
s
.
I
t
ca
n
b
e
co
m
p
u
ted
u
s
in
g
(
1
)
.
A
m
et
r
ic
co
m
m
o
n
ly
u
s
ed
in
ML
to
ass
es
s
th
e
p
er
f
o
r
m
an
ce
o
f
a
cl
ass
if
i
ca
tio
n
m
o
d
el,
p
ar
ticu
lar
ly
in
s
ce
n
ar
io
s
wit
h
im
b
alan
ce
d
class
d
is
tr
ib
u
tio
n
s
.
B
alan
ce
d
ac
c
u
r
ac
y
ad
d
r
ess
es
th
is
is
s
u
e
b
y
co
m
p
u
tin
g
th
e
a
v
er
ag
e
o
f
s
en
s
itiv
ity
(
tr
u
e
p
o
s
itiv
e
r
ate
)
ac
r
o
s
s
all
class
e
s
as
d
ep
icted
in
(
2
)
.
T
h
e
F1
-
s
co
r
e
,
d
ep
icted
in
(
3
)
r
ep
r
esen
ts
th
e
h
ar
m
o
n
ic
av
e
r
ag
e
o
f
r
ec
all
an
d
p
r
ec
is
io
n
,
tak
in
g
i
n
to
ac
co
u
n
t
b
o
th
f
alse
p
o
s
itiv
es
an
d
f
alse n
eg
ativ
es.
T
h
is
m
ak
es it p
ar
ticu
lar
ly
ef
f
ec
tiv
e
w
h
e
n
d
ea
lin
g
with
im
b
alan
ce
d
d
at
asets
.
=
(
+
)
(
+
+
+
)
(
1
)
=
1
+
2
+
⋯
.
(
2
)
1
=
2
×
(
×
)
(
+
)
(
3
)
T
h
e
AUC
i
s
a
m
ea
s
u
r
e
ca
lc
u
lated
f
r
o
m
th
e
R
OC
cu
r
v
e,
o
f
f
er
in
g
a
s
in
g
u
lar
s
ca
lar
v
alu
e
th
at
en
ca
p
s
u
lates
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
class
if
ier
ac
r
o
s
s
v
ar
io
u
s
th
r
esh
o
ld
s
.
An
id
ea
l
class
if
ier
ac
h
iev
es
an
AUC
o
f
1
.
0
,
wh
er
ea
s
a
r
an
d
o
m
cl
ass
if
ier
y
ield
s
an
AUC
o
f
0
.
5
.
E
f
f
icien
t
m
o
d
els
with
b
r
ie
f
tr
ain
in
g
tim
es
ar
e
d
esira
b
le
b
u
t
s
elec
tin
g
th
e
m
s
h
o
u
ld
n
o
t
co
m
p
r
o
m
is
e
p
r
ed
ic
tiv
e
p
er
f
o
r
m
an
ce
.
T
h
is
p
lo
t
is
ess
en
tial
f
o
r
ass
ess
in
g
th
e
tr
ad
e
-
o
f
f
s
b
etwe
en
m
o
d
el
a
cc
u
r
ac
y
a
n
d
th
e
c
o
m
p
u
tatio
n
al
r
eso
u
r
ce
s
n
ee
d
e
d
f
o
r
b
o
th
tr
ai
n
in
g
an
d
in
f
e
r
en
ce
.
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
ass
ess
m
en
t
o
f
ML
m
o
d
els
f
o
r
DDo
S
attac
k
d
etec
tio
n
in
SDN
r
ev
ea
ls
d
iv
e
r
s
e
p
e
r
f
o
r
m
a
n
ce
ch
ar
ac
ter
is
tics
.
T
ab
le
3
d
ep
ict
s
th
e
co
m
p
r
eh
en
s
iv
e
p
er
f
o
r
m
an
ce
ev
alu
atio
n
o
f
v
ar
i
o
u
s
M
L
alg
o
r
ith
m
s
in
th
e
d
etec
tio
n
o
f
DDo
S
Attack
s
.
T
h
e
L
GB
M
c
lass
if
ier
,
DT
class
if
i
er
,
r
an
d
o
m
f
o
r
est
class
if
ier
,
XGB
clas
s
if
ier
,
ex
tr
a
tr
ee
s
clas
s
if
ier
h
as
ac
h
iev
ed
p
er
f
ec
t
s
co
r
es
ac
r
o
s
s
all
m
etr
ics,
in
clu
d
in
g
ac
cu
r
ac
y
,
b
alan
ce
d
ac
cu
r
ac
y
,
R
OC
-
AU
C
,
an
d
F1
s
co
r
e
,
all
at
1
.
0
0
(
1
0
0
%).
T
h
is
s
u
g
g
ests
th
at
th
e
m
o
d
el
h
as
lik
ely
o
v
e
r
f
it
to
th
e
tr
ain
in
g
d
ata,
esp
ec
ial
ly
in
th
e
co
n
tex
t
o
f
a
co
m
p
lex
p
r
o
b
lem
lik
e
p
r
ed
ictin
g
DDo
S
attac
k
s
in
SD
N.
T
h
e
co
m
p
u
tatio
n
tim
e
o
f
1
.
6
8
,
0
.
5
2
,
6
.
3
8
,
1
.
7
,
an
d
3
.
4
1
s
ec
o
n
d
s
r
esp
ec
tiv
ely
in
d
icate
s
h
ig
h
ef
f
icien
cy
in
m
o
d
el
tr
ain
in
g
,
h
ig
h
lig
h
ts
s
im
p
licity
,
s
p
ee
d
,
an
d
ex
tr
em
ely
f
ast
tr
ain
in
g
.
Des
p
ite
ef
f
icien
t
tr
ain
in
g
tim
es,
th
ese
m
o
d
els
m
ay
n
o
t
r
ef
lect
r
ea
l
-
wo
r
ld
p
er
f
o
r
m
an
ce
ac
cu
r
ately
.
S
u
m
m
a
r
y
:
h
o
w
e
v
e
r
,
d
e
c
is
i
o
n
t
r
e
e
cl
as
s
i
f
ie
r
s
g
i
v
e
t
h
e
b
es
t
p
e
r
f
o
r
m
a
n
c
e
w
i
t
h
a
f
as
t
c
o
m
p
u
t
a
t
i
o
n
t
i
m
e
o
f
0
.
5
2
s
e
c
o
n
d
s
.
Mo
d
els
s
u
ch
as
ex
tr
a
tr
ee
class
if
ier
,
Ad
aBo
o
s
t
cla
s
s
if
i
er
an
d
K
-
n
eig
h
b
o
r
s
class
if
i
er
ac
h
iev
e
co
m
m
en
d
a
b
le
ac
cu
r
ac
y
o
f
0
.
9
9
(
9
9
%)
with
v
ar
y
in
g
co
m
p
u
t
atio
n
tim
es,
s
u
ch
as
3
.
4
1
,
0
.
1
8
,
an
d
4
.
6
6
s
ec
o
n
d
s
r
esp
ec
tiv
ely
,
s
h
o
wca
s
in
g
th
eir
ef
f
icac
y
an
d
s
ca
lab
ilit
y
.
I
n
co
n
tr
ast,
m
o
d
els
lik
e
SVC
an
d
SGD
class
if
ier
p
r
esen
t
s
lig
h
tly
lo
wer
s
co
r
es,
in
d
icatin
g
a
f
o
cu
s
o
n
g
en
er
aliza
tio
n
o
v
er
m
e
m
o
r
izatio
n
.
L
o
g
is
tic
r
eg
r
ess
io
n
s
tr
ik
es
a
b
alan
ce
b
etwe
en
s
im
p
licity
an
d
ef
f
icien
cy
,
d
e
m
o
n
s
tr
atin
g
m
o
d
er
ate
ac
cu
r
ac
y
as
s
h
o
wn
in
T
ab
le
3
.
Su
m
m
ar
y
:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2
2
5
2
-
8
9
3
8
E
va
lu
a
tio
n
o
f d
is
tr
ib
u
ted
d
en
i
a
l o
f ser
vice
a
tta
ck
s
d
etec
tio
n
in
s
o
ftw
a
r
e
d
efin
ed
n
etw
o
r
ks
(
N
ee
th
u
S
.
)
4495
h
o
wev
er
,
ex
tr
atr
ee
class
if
ier
s
g
iv
e
th
e
b
est
p
er
f
o
r
m
an
ce
o
f
9
9
%
ac
cu
r
ac
y
with
a
f
ast
co
m
p
u
tatio
n
tim
e
o
f
0
.
1
8
s
ec
o
n
d
s
.
L
in
ea
r
d
is
cr
im
in
an
t
an
aly
s
is
,
r
id
g
e
class
if
ier
C
V,
an
d
r
id
g
e
class
if
ier
ex
h
ib
it
lo
wer
ac
cu
r
ac
y
ar
o
u
n
d
7
2
.
3
3
%
s
u
g
g
esti
n
g
p
o
ten
tial
lim
itatio
n
s
in
ca
p
tu
r
in
g
co
m
p
lex
p
atter
n
s
.
Gau
s
s
ian
NB
,
p
er
ce
p
tr
o
n
,
n
ea
r
est
ce
n
tr
o
id
,
p
ass
iv
e
ag
g
r
ess
iv
e
cl
ass
if
ier
,
an
d
q
u
ad
r
atic
d
is
cr
im
in
an
t
an
aly
s
is
s
h
o
w
v
a
r
y
in
g
d
eg
r
ee
s
o
f
ac
cu
r
ac
y
o
f
6
3
%,
ea
ch
with
im
p
licatio
n
s
f
o
r
th
eir
s
u
itab
ilit
y
in
D
Do
S
d
etec
tio
n
s
ce
n
ar
io
s
.
T
h
e
d
u
m
m
y
class
if
ier
s
er
v
es
as
a
b
aselin
e
w
ith
a
m
in
im
al
ac
cu
r
ac
y
o
f
5
7
%,
r
ef
lectin
g
its
s
im
p
lis
tic
p
r
ed
ictio
n
ap
p
r
o
a
ch
b
ased
o
n
s
im
p
le
r
u
les.
T
h
is
m
o
d
el
p
r
o
v
id
es
a
b
en
ch
m
ar
k
f
o
r
e
v
alu
atin
g
t
h
e
p
er
f
o
r
m
a
n
ce
o
f
m
o
r
e
s
o
p
h
is
ticated
alg
o
r
ith
m
s
.
T
h
e
s
wif
t
co
m
p
u
tatio
n
tim
e
o
f
th
e
d
u
m
m
y
class
if
ier
alig
n
s
with
its
in
h
er
en
t
s
im
p
licity
.
E
a
ch
m
o
d
el'
s
u
n
iq
u
e
s
tr
en
g
th
s
an
d
wea
k
n
ess
es
u
n
d
er
s
co
r
e
th
e
cr
itical
im
p
o
r
ta
n
ce
o
f
s
elec
tin
g
an
ap
p
r
o
p
r
iate
alg
o
r
ith
m
tailo
r
e
d
to
s
p
ec
if
ic
task
r
eq
u
ir
e
m
en
ts
an
d
d
ataset
ch
ar
ac
ter
is
tics
.
Strik
in
g
a
b
alan
ce
b
etwe
en
ac
cu
r
a
cy
,
ef
f
icien
c
y
,
a
n
d
g
en
er
aliza
tio
n
is
c
r
u
cial
wh
e
n
im
p
lem
en
tin
g
ML
m
o
d
els
f
o
r
D
Do
S
d
etec
tio
n
in
th
e
d
y
n
am
i
c
lan
d
s
ca
p
e
o
f
SDN.
Fu
r
th
er
ex
p
lo
r
atio
n
,
in
clu
d
i
n
g
tech
n
iq
u
es
lik
e
cr
o
s
s
-
v
alid
atio
n
an
d
test
in
g
o
n
u
n
s
ee
n
d
at
a,
is
im
p
er
ativ
e
to
en
s
u
r
e
th
e
r
eliab
ilit
y
a
n
d
p
r
ac
tical
ap
p
licab
ilit
y
o
f
th
ese
m
o
d
els in
r
ea
l
-
wo
r
ld
cy
b
er
s
ec
u
r
ity
s
ce
n
a
r
io
s
.
Fro
m
th
e
T
ab
le
3
it
is
al
s
o
o
b
s
er
v
ed
th
at
b
ased
o
n
ac
cu
r
ac
y
an
d
tim
e
tak
en
f
o
r
co
m
p
u
tatio
n
s
h
o
ws
th
at
DT
is
th
e
b
est
(
1
0
0
%
-
0
.
5
2
s
ec
o
n
d
s
)
,
ex
tr
a
tr
ee
class
if
ier
(
9
9
%
-
0
.
1
8
s
ec
o
n
d
s
)
an
d
Gau
s
s
ian
NB
-
lo
w
ac
cu
r
ac
y
with
lo
w
c
o
m
p
u
tatio
n
tim
e
o
f
0
.
1
2
s
ec
o
n
d
s
.
T
h
e
b
ar
p
lo
t
f
o
r
ac
cu
r
ac
y
as
s
h
o
wn
in
Fig
u
r
e
4
,
lik
ely
s
h
o
ws
th
at
m
o
d
els
s
u
ch
as
L
GB
M
clas
s
if
ier
,
DT
clas
s
if
ier
,
r
an
d
o
m
f
o
r
est
class
if
ier
,
an
d
XGB
cla
s
s
if
ier
ar
e
p
er
f
o
r
m
in
g
with
h
ig
h
ac
cu
r
ac
y
,
p
o
ten
tially
r
ea
c
h
in
g
1
.
0
0
.
T
h
is
ca
n
b
e
in
d
icativ
e
o
f
v
er
y
e
f
f
e
ctiv
e
m
o
d
els,
b
u
t
in
r
ea
l
-
wo
r
ld
s
ce
n
a
r
io
s
,
s
u
ch
p
er
f
ec
tio
n
o
f
ten
f
lag
s
a
n
ee
d
f
o
r
s
cr
u
tin
y
,
p
o
s
s
ib
ly
h
in
tin
g
at
o
v
er
f
itti
n
g
.
T
ab
le
3
.
Per
f
o
r
m
an
ce
co
m
p
a
r
is
o
n
o
f
ML
alg
o
r
ith
m
s
in
d
etec
tio
n
o
f
DDo
S a
ttack
s
M
o
d
e
l
A
c
c
u
r
a
c
y
B
a
l
a
n
c
e
d
a
c
c
u
r
a
c
y
R
O
C
-
AUC
F1
-
sc
o
r
e
Ti
me
t
a
k
e
n
LG
B
M
c
l
a
ss
i
f
i
e
r
1
1
1
1
1
.
6
8
D
T
c
l
a
ssi
f
i
e
r
1
1
1
1
0
.
5
2
R
a
n
d
o
m f
o
r
e
s
t
c
l
a
ss
i
f
i
e
r
1
1
1
1
6
.
3
8
X
G
B
c
l
a
ssi
f
i
e
r
1
1
1
1
1
.
7
Ex
t
r
a
Tr
e
e
s c
l
a
ss
i
f
i
e
r
1
1
1
1
3
.
4
1
Ex
t
r
a
Tr
e
e
c
l
a
ss
i
f
i
e
r
0
.
9
9
0
.
9
9
0
.
9
9
0
.
9
9
0
.
1
8
A
d
a
B
o
o
st
c
l
a
ss
i
f
i
e
r
0
.
9
9
0
.
9
9
0
.
9
9
0
.
9
9
4
.
6
6
K
N
e
i
g
h
b
o
r
s
C
l
a
ssi
f
i
e
r
0
.
9
9
0
.
9
9
0
.
9
9
0
.
9
9
5
.
4
9
S
V
C
0
.
9
5
0
.
9
6
0
.
9
6
0
.
9
6
5
8
.
1
4
S
G
D
c
l
a
ss
i
f
i
e
r
0
.
8
1
0
.
8
1
0
.
8
1
0
.
8
1
0
.
5
1
Lo
g
i
s
t
i
c
r
e
g
r
e
ssi
o
n
0
.
7
9
0
.
7
8
0
.
7
8
0
.
7
9
0
.
8
4
Li
n
e
a
r
d
i
scri
m
i
n
a
n
t
a
n
a
l
y
s
i
s
0
.
7
3
0
.
7
3
0
.
7
3
0
.
7
3
0
.
3
2
R
i
d
g
e
c
l
a
ss
i
f
i
e
r
C
V
0
.
7
2
0
.
7
2
0
.
7
2
0
.
7
2
0
.
3
R
i
d
g
e
c
l
a
ss
i
f
i
e
r
0
.
7
2
0
.
7
2
0
.
7
2
0
.
7
2
0
.
1
5
G
a
u
ss
i
a
n
N
B
0
.
6
8
0
.
6
9
0
.
6
9
0
.
6
8
0
.
1
2
P
e
r
c
e
p
t
r
o
n
0
.
6
6
0
.
6
8
0
.
6
8
0
.
6
6
0
.
3
1
N
e
a
r
e
st
c
e
n
t
r
o
i
d
0
.
6
5
0
.
6
5
0
.
6
5
0
.
6
5
0
.
1
7
P
a
ssi
v
e
a
g
g
r
e
ss
i
v
e
c
l
a
ssi
f
i
e
r
0
.
5
8
0
.
6
1
0
.
6
1
0
.
5
7
0
.
4
5
Q
u
a
d
r
a
t
i
c
d
i
scri
mi
n
a
n
t
a
n
a
l
y
s
i
s
0
.
5
8
0
.
5
1
0
.
5
1
0
.
4
4
0
.
2
8
D
u
mm
y
c
l
a
ss
i
f
i
e
r
0
.
5
7
0
.
5
0
.
5
0
.
4
1
0
.
1
2
Fig
u
r
e
4
.
Per
f
o
r
m
an
c
e
co
m
p
ar
is
o
n
o
f
v
a
r
io
u
s
m
o
d
els with
r
esp
ec
t to
th
e
ac
cu
r
ac
y
an
d
b
ala
n
ce
d
ac
cu
r
ac
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell,
Vo
l.
1
3
,
No
.
4
,
Dec
em
b
er
2
0
2
4
:
4
4
8
8
-
4
4
9
8
4496
Mo
d
els
wit
h
s
li
g
h
tl
y
less
ac
cu
r
a
cy
,
s
u
c
h
as
A
d
aB
o
o
s
t
cl
ass
if
ier
an
d
K
-
n
ei
g
h
b
o
r
s
cl
ass
i
f
ie
r
,
m
a
y
o
f
f
e
r
m
o
r
e
r
e
alis
t
ic
p
r
e
d
i
cti
v
e
p
e
r
f
o
r
m
an
ce
s
a
n
d
p
o
t
en
tia
ll
y
b
e
tte
r
g
e
n
e
r
al
iza
ti
o
n
t
o
u
n
s
ee
n
d
ata
.
B
al
an
ce
d
ac
c
u
r
ac
y
as
d
e
p
ict
e
d
i
n
Fi
g
u
r
e
4
,
is
c
r
u
ci
al
in
d
atas
ets
w
it
h
i
m
b
al
an
ce
d
cl
ass
es
.
A
b
ar
p
l
o
t
ill
u
s
t
r
ati
n
g
h
ig
h
b
a
la
n
c
e
d
ac
c
u
r
a
cy
f
o
r
t
h
e
s
am
e
t
o
p
-
p
e
r
f
o
r
m
i
n
g
m
o
d
e
ls
s
u
g
g
es
ts
t
h
at
t
h
es
e
m
o
d
els
a
r
e
ef
f
e
cti
v
e
ac
r
o
s
s
b
o
t
h
m
aj
o
r
it
y
a
n
d
m
i
n
o
r
i
ty
c
lass
es.
H
o
we
v
er
,
t
h
e
s
a
m
e
c
au
ti
o
n
r
e
g
a
r
d
i
n
g
o
v
e
r
f
it
tin
g
a
p
p
li
es
h
e
r
e.
F
o
r
m
o
d
els
wit
h
l
o
w
e
r
b
a
la
n
c
ed
ac
c
u
r
a
cy
,
it
m
ay
i
n
d
i
ca
t
e
th
at
t
h
es
e
m
o
d
els
s
t
r
u
g
g
le
m
o
r
e
w
it
h
cl
ass
i
m
b
al
a
n
ce
,
a
f
f
ec
ti
n
g
t
h
eir
p
e
r
f
o
r
m
a
n
c
e
i
n
a
r
e
al
-
wo
r
l
d
s
et
ti
n
g
.
T
h
e
R
OC
A
UC
b
ar
i
n
Fi
g
u
r
e
5
li
k
el
y
m
ir
r
o
r
s
a
cc
u
r
a
cy
a
n
d
b
ala
n
c
e
d
ac
c
u
r
ac
y
t
r
e
n
d
s
.
Fig
u
r
e
5
.
Per
f
o
r
m
an
c
e
co
m
p
ar
is
o
n
o
f
v
a
r
io
u
s
m
o
d
els with
r
esp
ec
t to
th
e
R
OC
-
AUC an
d
F
1
s
co
r
e
Hig
h
R
OC
AU
C
v
alu
es
n
ea
r
1
.
0
0
s
u
g
g
est
ef
f
ec
tiv
e
class
d
is
tin
ctio
n
,
th
o
u
g
h
ex
ce
s
s
iv
e
v
alu
es
m
ay
im
p
ly
o
v
er
f
itti
n
g
.
L
o
wer
v
alu
es
in
d
icate
less
d
is
cr
im
in
ativ
e
p
o
wer
,
im
p
r
o
v
a
b
le
with
f
ea
tu
r
e
en
g
in
ee
r
in
g
o
r
m
o
d
el
tu
n
in
g
.
T
h
e
F1
Sco
r
e
b
ar
co
m
b
in
es p
r
e
cisi
o
n
an
d
r
ec
all,
r
ev
ea
lin
g
a
m
o
d
el'
s
tr
u
e
-
p
o
s
itiv
e
id
en
tific
atio
n
b
alan
ce
.
Hig
h
F1
s
co
r
es
ar
e
c
r
u
cial
f
o
r
DDo
S
d
etec
tio
n
,
w
h
ile
lo
wer
s
co
r
es
r
eq
u
ir
e
p
r
ec
is
io
n
-
r
ec
all
b
alan
c
e
r
ef
in
em
en
t.
T
h
e
tim
e
tak
e
n
b
a
r
p
lo
t
,
as
d
ep
icted
in
Fig
u
r
e
6
,
lik
ely
s
h
o
ws
s
ig
n
if
ican
t
v
ar
ia
t
io
n
ac
r
o
s
s
m
o
d
els.
Mo
d
els
lik
e
SVC
th
at
tak
e
lo
n
g
er
to
t
r
ain
m
ay
n
o
t
b
e
s
u
it
ab
le
f
o
r
e
n
v
ir
o
n
m
en
ts
wh
er
e
q
u
ick
r
etr
ai
n
in
g
is
n
ec
ess
ar
y
.
C
o
n
v
er
s
ely
,
m
o
d
el
s
with
s
h
o
r
t
tr
ain
in
g
tim
es
ar
e
d
esira
b
le
f
o
r
th
eir
ef
f
icien
c
y
b
u
t
s
h
o
u
ld
n
o
t
b
e
ch
o
s
en
at
th
e
c
o
s
t
o
f
p
r
e
d
ictiv
e
p
er
f
o
r
m
a
n
ce
.
T
h
is
p
lo
t
is
cr
u
cial
f
o
r
u
n
d
er
s
tan
d
in
g
th
e
tr
ad
e
-
o
f
f
s
b
etwe
en
m
o
d
el
ac
cu
r
ac
y
an
d
th
e
c
o
m
p
u
tatio
n
al
r
eso
u
r
ce
s
r
e
q
u
ir
ed
f
o
r
m
o
d
el
tr
ai
n
in
g
a
n
d
in
f
e
r
en
ce
.
Fig
u
r
e
6
.
Per
f
o
r
m
an
c
e
co
m
p
ar
is
o
n
o
f
v
a
r
io
u
s
m
o
d
els with
r
esp
ec
t to
th
e
co
m
p
u
tatio
n
tim
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2
2
5
2
-
8
9
3
8
E
va
lu
a
tio
n
o
f d
is
tr
ib
u
ted
d
en
i
a
l o
f ser
vice
a
tta
ck
s
d
etec
tio
n
in
s
o
ftw
a
r
e
d
efin
ed
n
etw
o
r
ks
(
N
ee
th
u
S
.
)
4497
As
it
ca
n
b
e
s
ee
n
f
r
o
m
T
ab
le
4
,
th
e
p
r
o
p
o
s
ed
wo
r
k
p
er
f
o
r
m
s
b
etter
th
an
m
o
s
t
o
f
th
e
s
tate
-
of
-
th
e
-
ar
t
m
eth
o
d
s
.
T
h
e
p
r
esen
t
s
tu
d
y
h
a
s
p
r
o
v
id
ed
b
est
r
esu
lts
in
ter
m
s
o
f
ac
cu
r
ac
y
,
b
alan
ce
d
ac
cu
r
a
cy
,
R
OC
-
AU
C
an
d
F1
s
co
r
e
b
y
u
s
in
g
SMOT
E
f
o
r
f
ea
tu
r
e
e
n
g
in
ee
r
i
n
g
an
d
al
l
p
o
p
u
lar
e
n
s
em
b
lin
g
tech
n
iq
u
es
av
ailab
le.
T
h
is
p
er
f
o
r
m
an
ce
s
p
ec
tr
u
m
em
p
h
a
s
izes
th
e
ch
allen
g
e
o
f
ac
cu
r
at
ely
p
r
ed
ictin
g
DDo
S
attac
k
s
a
n
d
u
n
d
er
s
co
r
es
th
e
im
p
o
r
tan
ce
o
f
s
elec
tin
g
a
s
u
itab
le
m
o
d
el
alig
n
e
d
with
th
e
s
p
ec
if
ic
ch
ar
ac
ter
is
tics
o
f
cy
b
e
r
s
ec
u
r
ity
d
ata.
T
h
e
v
ar
y
in
g
o
u
tc
o
m
es
s
tr
ess
th
e
s
ig
n
if
ican
ce
o
f
c
o
m
p
r
e
h
en
d
in
g
th
e
u
n
d
er
ly
i
n
g
d
ata
an
d
p
r
o
b
lem
co
n
tex
t
wh
en
ch
o
o
s
in
g
a
n
d
f
in
e
-
tu
n
in
g
ML
alg
o
r
ith
m
s
f
o
r
cy
b
e
r
s
ec
u
r
ity
a
p
p
licatio
n
s
.
T
ab
le
4
.
C
o
m
p
a
r
is
o
n
o
f
p
r
o
p
o
s
ed
wo
r
k
with
s
tate
o
f
ar
t
A
u
t
h
o
r
(
y
e
a
r
)
D
a
t
a
s
e
t
u
s
e
d
M
e
t
h
o
d
o
l
o
g
y
P
e
r
f
o
r
ma
n
c
e
me
t
r
i
c
s
N
o
.
o
f
f
e
a
t
u
r
e
s
c
o
n
si
d
e
r
e
d
C
l
a
s
si
f
i
e
r
s
A
c
c
u
r
a
c
y
(
%)
R
O
C
-
AUC
(
%)
F1
-
sc
o
r
e
(
%)
A
l
g
h
o
s
o
n
.
a
n
d
A
b
b
a
s (2
0
2
1
)
[
3
]
C
I
C
D
D
o
S
2
0
1
9
d
a
t
a
se
t
20
R
F
,
LG
B
,
C
a
t
b
o
o
st
,
C
N
N
99
u
n
k
n
o
w
n
u
n
k
n
o
w
n
Lo
n
g
a
n
d
Ji
n
s
o
n
g
(
2
0
2
2
)
[
19
]
D
A
R
P
A
R
d
a
t
a
se
t
18
S
S
A
E
-
S
V
M
98
u
n
k
n
o
w
n
u
n
k
n
o
w
n
P
o
l
a
t
et
al
.
[
1
4
]
S
C
A
D
A
d
a
t
a
s
e
t
10
R
N
N
,
LST
M
,
a
n
d
G
R
U
9
7
.
6
2
u
n
k
n
o
w
n
u
n
k
n
o
w
n
V
a
d
h
i
l
et
al
.
[
6
]
C
I
C
I
D
S
-
2
0
1
7
)
v
a
r
i
a
b
l
e
n
u
m
b
e
r
o
f
f
e
a
t
u
r
e
s
D
T,
R
F
,
L
R
,
G
N
B
,
A
B
& E
n
s
e
mb
l
e
9
9
.
5
0
u
n
k
n
o
w
n
9
9
.
4
4
B
h
a
y
o
et
al
.
(
2
0
2
3
)
[
9
]
S
i
mu
l
a
t
e
d
t
r
a
f
f
i
c
05
N
B
,
S
V
M
,
a
n
d
D
T
9
8
.
1
u
n
k
n
o
w
n
u
n
k
n
o
w
n
P
r
e
sen
t
s
t
u
d
y
[
2
7
]
16
En
se
mb
l
e
t
e
c
h
n
i
q
u
e
s,
M
L
w
i
t
h
S
M
O
TE
1
0
0
1
0
0
1
0
0
4.
CO
NCLU
SI
O
N
I
n
th
is
wo
r
k
,
we
ap
p
lied
v
ar
io
u
s
wid
ely
u
s
ed
ML
tec
h
n
iq
u
e
s
to
id
en
tif
y
DDo
S
attac
k
s
wi
th
in
SDN.
Ou
r
s
tu
d
y
ex
ten
s
iv
ely
ev
alu
at
es
th
e
p
er
f
o
r
m
an
ce
o
f
th
ese
p
o
p
u
lar
ML
tech
n
iq
u
es,
d
is
tin
g
u
is
h
in
g
its
elf
f
r
o
m
ex
is
tin
g
wo
r
k
s
t
h
at
f
o
cu
s
o
n
o
n
ly
a
s
u
b
s
et
o
f
th
ese
m
et
h
o
d
s
.
No
tab
ly
,
we
o
b
s
er
v
e
d
th
at
alth
o
u
g
h
ce
r
t
ain
m
o
d
els
ex
h
ib
it
o
u
ts
tan
d
in
g
p
er
f
o
r
m
a
n
ce
ac
r
o
s
s
m
u
ltip
le
m
etr
ics,
co
n
s
is
ten
t
h
ig
h
s
co
r
es
m
ay
in
d
icate
p
o
ten
tial
o
v
er
f
itti
n
g
,
p
ar
ticu
lar
l
y
in
in
tr
icate
d
o
m
ain
s
lik
e
cy
b
er
s
ec
u
r
i
ty
.
T
h
e
ML
m
o
d
els,
in
cl
u
d
in
g
L
GB
M
class
if
ier
,
DT
class
if
ier
,
r
an
d
o
m
f
o
r
est
class
if
ier
,
XGB
class
if
ier
,
ex
tr
a
tr
ee
s
class
if
ier
,
Ad
aBo
o
s
t
class
if
ier
,
an
d
K
-
n
eig
h
b
o
r
s
class
if
ier
,
d
em
o
n
s
tr
ated
n
o
tewo
r
th
y
ef
f
ec
tiv
e
n
ess
in
v
ar
io
u
s
asp
ec
ts
o
f
D
Do
S
d
etec
tio
n
an
d
m
itig
atio
n
.
T
o
e
n
s
u
r
e
p
r
ac
tica
l
ap
p
licab
ilit
y
,
it
is
cr
u
cial
to
s
u
b
ject
th
ese
m
o
d
els
to
f
u
r
t
h
e
r
s
cr
u
tin
y
th
r
o
u
g
h
tech
n
iq
u
es
s
u
ch
as
cr
o
s
s
-
v
alid
atio
n
an
d
tes
tin
g
o
n
u
n
s
ee
n
d
ata.
Fu
tu
r
e
en
d
ea
v
o
r
s
s
h
o
u
l
d
ex
p
lo
r
e
en
h
an
cin
g
m
o
d
els
with
lo
wer
p
e
r
f
o
r
m
a
n
ce
u
s
in
g
ad
v
an
ce
d
f
ea
tu
r
e
s
elec
tio
n
,
h
y
p
er
p
ar
a
m
eter
tu
n
in
g
,
o
r
en
s
em
b
le
m
eth
o
d
s
.
I
m
p
lem
en
tin
g
ML
m
o
d
els
f
o
r
DDo
S
d
etec
tio
n
a
n
d
m
itig
atio
n
tr
an
s
ce
n
d
s
b
ein
g
an
o
p
tio
n
;
it
s
tan
d
s
as a
n
ec
ess
ity
in
th
e
co
n
tem
p
o
r
ar
y
cy
b
er
s
ec
u
r
ity
lan
d
s
ca
p
e.
RE
F
E
R
E
NC
E
S
[
1
]
B
.
A
l
h
i
j
a
w
i
,
S
.
A
l
m
a
j
a
l
i
,
H
.
E
l
g
a
l
a
,
H
.
B
.
S
a
l
a
m
e
h
,
a
n
d
M
.
A
y
y
a
s
h
,
“
A
su
r
v
e
y
o
n
D
o
S
/
D
D
o
S
mi
t
i
g
a
t
i
o
n
t
e
c
h
n
i
q
u
e
s
i
n
S
D
N
s:
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
c
o
m
p
a
r
i
so
n
,
so
l
u
t
i
o
n
s,
t
e
st
i
n
g
t
o
o
l
s
a
n
d
d
a
t
a
set
s
,
”
C
o
m
p
u
t
e
r
s
a
n
d
E
l
e
c
t
ri
c
a
l
En
g
i
n
e
e
r
i
n
g
,
v
o
l
.
9
9
,
A
p
r
.
2
0
2
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
mp
e
l
e
c
e
n
g
.
2
0
2
2
.
1
0
7
7
0
6
.
[
2
]
K
.
G
i
o
t
i
s
,
C
.
A
r
g
y
r
o
p
o
u
l
o
s,
G
.
A
n
d
r
o
u
l
i
d
a
k
i
s,
D
.
K
a
l
o
g
e
r
a
s,
a
n
d
V
.
M
a
g
l
a
r
i
s,
“
C
o
mb
i
n
i
n
g
O
p
e
n
F
l
o
w
a
n
d
sF
l
o
w
f
o
r
a
n
e
f
f
e
c
t
i
v
e
a
n
d
sca
l
a
b
l
e
a
n
o
m
a
l
y
d
e
t
e
c
t
i
o
n
a
n
d
mi
t
i
g
a
t
i
o
n
m
e
c
h
a
n
i
sm
o
n
S
D
N
e
n
v
i
r
o
n
m
e
n
t
s,”
C
o
m
p
u
t
e
r
N
e
t
w
o
rks
,
v
o
l
.
6
2
,
p
p
.
1
2
2
–
1
3
6
,
2
0
1
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
b
j
p
.
2
0
1
3
.
1
0
.
0
1
4
.
[
3
]
E.
S
.
A
l
g
h
o
s
o
n
a
n
d
O
.
A
b
b
a
ss,
“
D
e
t
e
c
t
i
n
g
d
i
st
r
i
b
u
t
e
d
d
e
n
i
a
l
o
f
ser
v
i
c
e
a
t
t
a
c
k
s
u
si
n
g
ma
c
h
i
n
e
l
e
a
r
n
i
n
g
m
o
d
e
l
s,”
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
A
d
v
a
n
c
e
d
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
Ap
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
1
2
,
n
o
.
1
2
,
p
p
.
6
1
6
–
6
2
2
,
2
0
2
1
,
d
o
i
:
1
0
.
1
4
5
6
9
/
I
JA
C
S
A
.
2
0
2
1
.
0
1
2
1
2
7
7
.
[
4
]
Y
.
C
u
i
e
t
a
l
.
,
“
T
o
w
a
r
d
s
D
D
o
S
d
e
t
e
c
t
i
o
n
m
e
c
h
a
n
i
sms
i
n
so
f
t
w
a
r
e
-
d
e
f
i
n
e
d
n
e
t
w
o
r
k
i
n
g
,
”
J
o
u
r
n
a
l
o
f
N
e
t
w
o
rk
a
n
d
C
o
m
p
u
t
e
r
Ap
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
1
9
0
,
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
j
n
c
a
.
2
0
2
1
.
1
0
3
1
5
6
.
[
5
]
J.
S
i
n
g
h
a
n
d
S
.
B
e
h
a
l
,
“
D
e
t
e
c
t
i
o
n
a
n
d
mi
t
i
g
a
t
i
o
n
o
f
D
D
o
S
a
t
t
a
c
k
s
i
n
S
D
N
:
A
c
o
m
p
r
e
h
e
n
si
v
e
r
e
v
i
e
w
,
r
e
sea
r
c
h
c
h
a
l
l
e
n
g
e
s
a
n
d
f
u
t
u
r
e
d
i
r
e
c
t
i
o
n
s,
”
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
Re
v
i
e
w
,
v
o
l
.
3
7
,
2
0
2
0
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
sr
e
v
.
2
0
2
0
.
1
0
0
2
7
9
.
[
6
]
F
.
A
.
V
a
d
h
i
l
,
M
.
L.
S
a
l
i
h
i
,
a
n
d
M
.
F
.
N
a
n
n
e
,
“
M
a
c
h
i
n
e
l
e
a
r
n
i
n
g
-
b
a
se
d
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
sy
s
t
e
m f
o
r
d
e
t
e
c
t
i
n
g
w
e
b
a
t
t
a
c
k
s,
”
I
AE
S
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
1
3
,
n
o
.
1
,
p
p
.
7
1
1
–
7
2
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
5
9
1
/
i
j
a
i
.
v
1
3
.
i
1
.
p
p
7
1
1
-
7
2
1
.
[
7
]
B
.
W
a
n
g
,
Y
.
Zh
e
n
g
,
W
.
Lo
u
,
a
n
d
Y
.
T
.
H
o
u
,
“
D
D
o
S
a
t
t
a
c
k
p
r
o
t
e
c
t
i
o
n
i
n
t
h
e
e
r
a
o
f
c
l
o
u
d
c
o
m
p
u
t
i
n
g
a
n
d
so
f
t
w
a
r
e
-
d
e
f
i
n
e
d
n
e
t
w
o
r
k
i
n
g
,
”
C
o
m
p
u
t
e
r
N
e
t
w
o
rks
,
v
o
l
.
8
1
,
p
p
.
3
0
8
–
3
1
9
,
2
0
1
5
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
n
e
t
.
2
0
1
5
.
0
2
.
0
2
6
.
[
8
]
M
.
A
.
R
i
b
e
i
r
o
,
M
.
S
.
P
.
F
o
n
se
c
a
,
a
n
d
J.
D.
S
a
n
t
i
,
“
D
e
t
e
c
t
i
n
g
a
n
d
mi
t
i
g
a
t
i
n
g
D
D
o
S
a
t
t
a
c
k
s
w
i
t
h
m
o
v
i
n
g
t
a
r
g
e
t
d
e
f
e
n
s
e
a
p
p
r
o
a
c
h
b
a
se
d
o
n
a
u
t
o
m
a
t
e
d
f
l
o
w
c
l
a
ss
i
f
i
c
a
t
i
o
n
i
n
S
D
N
n
e
t
w
o
r
k
s
,
”
C
o
m
p
u
t
e
rs
a
n
d
S
e
c
u
r
i
t
y
,
v
o
l
.
1
3
4
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
s
e
.
2
0
2
3
.
1
0
3
4
6
2
.
[
9
]
J.
B
h
a
y
o
,
S
.
A
.
S
h
a
h
,
S
.
H
a
m
e
e
d
,
A
.
A
h
me
d
,
J
.
N
a
s
i
r
,
a
n
d
D
.
D
r
a
h
e
i
m
,
“
T
o
w
a
r
d
s
a
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
-
b
a
s
e
d
f
r
a
mew
o
r
k
f
o
r
D
D
O
S
a
t
t
a
c
k
d
e
t
e
c
t
i
o
n
i
n
s
o
f
t
w
a
r
e
-
d
e
f
i
n
e
d
I
o
T
(
S
D
-
I
o
T)
n
e
t
w
o
r
k
s,
”
En
g
i
n
e
e
r
i
n
g
Ap
p
l
i
c
a
t
i
o
n
s
o
f
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
1
2
3
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
e
n
g
a
p
p
a
i
.
2
0
2
3
.
1
0
6
4
3
2
.
[
1
0
]
H
.
Zh
o
u
,
Y
.
Z
h
e
n
g
,
X
.
J
i
a
,
a
n
d
J.
S
h
u
,
“
C
o
l
l
a
b
o
r
a
t
i
v
e
p
r
e
d
i
c
t
i
o
n
a
n
d
d
e
t
e
c
t
i
o
n
o
f
D
D
o
S
a
t
t
a
c
k
s
i
n
e
d
g
e
c
o
m
p
u
t
i
n
g
:
A
d
e
e
p
l
e
a
r
n
i
n
g
-
b
a
s
e
d
a
p
p
r
o
a
c
h
w
i
t
h
d
i
st
r
i
b
u
t
e
d
S
D
N
,
”
C
o
m
p
u
t
e
r
N
e
t
w
o
rk
s
,
v
o
l
.
2
2
5
,
2
0
2
3
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
mn
e
t
.
2
0
2
3
.
1
0
9
6
4
2
.
Evaluation Warning : The document was created with Spire.PDF for Python.