T
E
L
KO
M
NIK
A
, V
ol
.
17
,
No.
6,
Dec
em
be
r
20
1
9,
p
p.
31
55
~
31
60
IS
S
N: 1
69
3
-
6
93
0
,
accr
ed
ited
F
irst
Gr
ad
e b
y K
em
en
r
istekdikti,
Decr
ee
No:
21
/E
/
K
P
T
/20
18
DOI:
10.12928/TE
LK
OM
N
IK
A
.v
1
7
i
6
.
13059
◼
31
55
Rec
ei
v
ed
M
ay
8
, 2
01
9
;
Re
v
i
s
ed
J
ul
y
2
, 2
0
1
9
;
A
c
c
ep
t
e
d
J
ul
y
18
, 2
01
9
Classif
y
i
ng
con
fid
e
nt
ial
dat
a
us
ing
SVM
fo
r
ef
ficie
nt
cl
o
ud
q
uer
y
proce
ssing
Hu
d
a Ka
d
h
im T
a
yy
eh
*
1
,
Ah
med
S
ab
ah
A
h
med
A
l
-
J
u
maili
2
1
Dep
a
rt
m
e
n
t
o
f
I
n
fo
r
m
a
ti
c
s
Sy
s
te
m
s
M
a
n
a
g
e
m
e
n
t
(ISM
),
Col
l
e
g
e
o
f
Bu
s
i
n
e
s
s
I
n
f
o
rm
a
ti
c
s
,
Uni
v
e
rs
i
ty
o
f
I
n
fo
r
m
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
a
n
d
Co
m
m
u
n
i
c
a
ti
o
n
s
,
Ba
g
h
d
a
d
,
Ira
q
2
Dep
a
rt
m
e
n
t
o
f
Bu
s
i
n
e
s
s
I
n
fo
r
m
a
ti
o
n
T
e
c
h
n
o
l
o
g
y
(BI
T
),
Col
l
e
g
e
o
f
Bu
s
i
n
e
s
s
In
fo
r
m
a
t
i
c
s
,
Uni
v
e
rs
i
ty
o
f
I
n
fo
r
m
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
a
n
d
Co
m
m
u
n
i
c
a
ti
o
n
s
,
Ba
g
h
d
a
d
,
Ira
q
*C
o
rre
s
p
o
n
d
i
n
g
a
u
th
o
r,
e
-
m
a
i
l
:
h
a
l
j
o
b
o
r
i
@uo
i
tc
.e
d
u
.
i
q
1
,
a
s
a
b
a
h
j
@uo
i
tc
.
e
d
u
.i
q
2
Ab
strac
t
Nowa
d
a
y
s
,
o
rg
a
n
i
z
a
ti
o
n
s
a
re
wid
e
l
y
u
s
i
n
g
a
c
l
o
u
d
d
a
ta
b
a
s
e
e
n
g
i
n
e
fro
m
th
e
c
l
o
u
d
s
e
rv
i
c
e
p
ro
v
i
d
e
r
s
.
Pri
v
a
c
y
s
ti
l
l
i
s
t
h
e
m
a
i
n
c
o
n
c
e
rn
fo
r
t
h
e
s
e
o
rg
a
n
i
z
a
ti
o
n
s
whe
re
e
v
e
r
y
o
rg
a
n
i
z
a
ti
o
n
i
s
s
tri
c
tl
y
l
o
o
k
i
n
g
fo
rward
m
o
re
s
e
c
u
re
e
n
v
i
ro
n
m
e
n
t
fo
r
th
e
i
r o
wn
d
a
ta
.
Se
v
e
ra
l
s
t
u
d
i
e
s
h
a
v
e
p
ro
p
o
s
e
d
d
i
ff
e
re
n
t
t
y
p
e
s
o
f
e
n
c
r
y
p
ti
o
n
m
e
t
h
o
d
s
t
o
p
r
o
t
e
c
t
th
e
d
a
t
a
o
v
e
r
th
e
c
l
o
u
d
.
H
o
wev
e
r,
th
e
d
a
i
l
y
tr
a
n
s
a
c
t
i
o
n
s
re
p
re
s
e
n
te
d
b
y
q
u
e
ri
e
s
fo
r
s
u
c
h
d
a
ta
b
a
s
e
s
m
a
k
e
s
e
n
c
r
y
p
t
i
o
n
i
s
i
n
e
ff
i
c
i
e
n
t
s
o
l
u
ti
o
n
.
T
h
e
re
f
o
re
,
re
c
e
n
t
s
t
u
d
i
e
s
p
re
s
e
n
te
d
a
m
e
c
h
a
n
i
s
m
fo
r
c
l
a
s
s
i
fy
i
n
g
th
e
d
a
t
a
p
ri
o
r
to
m
i
g
ra
t
e
i
n
t
o
th
e
c
l
o
u
d
.
Th
i
s
wou
l
d
re
d
u
c
e
th
e
n
e
e
d
o
f
e
n
c
r
y
p
ti
o
n
whi
c
h
e
n
h
a
n
c
e
s
th
e
e
ff
i
c
i
e
n
c
y
.
Ye
t,
m
o
s
t
o
f
th
e
c
l
a
s
s
i
f
i
c
a
ti
o
n
m
e
th
o
d
s
u
s
e
d
i
n
th
e
l
i
t
e
ra
tu
r
e
were
b
a
s
e
d
o
n
s
tri
n
g
-
b
a
s
e
d
m
a
tc
h
i
n
g
a
p
p
ro
a
c
h
.
S
u
c
h
a
p
p
r
o
a
c
h
s
u
ff
e
r
s
o
f
t
h
e
e
x
a
c
t
m
a
tc
h
o
f
te
rm
s
whe
re
th
e
p
a
rti
a
l
m
a
t
c
h
i
n
g
wou
l
d
n
o
t
b
e
c
o
n
s
i
d
e
re
d
.
Th
i
s
p
a
p
e
r
a
i
m
s
t
o
t
a
k
e
th
e
a
d
v
a
n
ta
g
e
o
f
N
-
g
ra
m
re
p
re
s
e
n
ta
ti
o
n
a
l
o
n
g
wit
h
Su
p
p
o
rt
Ve
c
to
r
M
a
c
h
i
n
e
c
l
a
s
s
i
fi
c
a
ti
o
n
.
A
re
a
l
-
ti
m
e
d
a
t
a
wil
l
u
s
e
d
i
n
th
e
e
x
p
e
ri
m
e
n
t.
Aft
e
r
c
o
n
d
u
c
t
i
n
g
th
e
c
l
a
s
s
i
fi
c
a
t
i
o
n
,
t
h
e
Ad
v
a
n
c
e
d
En
c
r
y
p
t
i
o
n
Sta
n
d
a
r
d
a
l
g
o
ri
th
m
wil
l
b
e
u
s
e
d
to
e
n
c
ry
p
t
th
e
c
o
n
fi
d
e
n
t
i
a
l
d
a
ta
.
Res
u
l
t
s
s
h
o
wed
t
h
a
t
th
e
p
ro
p
o
s
e
d
m
e
th
o
d
o
u
t
p
e
rfo
r
m
e
d
t
h
e
b
a
s
e
l
i
n
e
e
n
c
r
y
p
ti
o
n
m
e
th
o
d
.
Th
i
s
e
m
p
h
a
s
i
z
e
s
th
e
u
s
e
fu
l
n
e
s
s
o
f
u
s
i
n
g
th
e
m
a
c
h
i
n
e
l
e
a
rn
i
n
g
te
c
h
n
i
q
u
e
s
fo
r
th
e
p
ro
c
e
s
s
o
f
c
l
a
s
s
i
fy
i
n
g
t
h
e
d
a
ta
b
a
s
e
d
o
n
c
o
n
fi
d
e
n
t
i
a
l
i
ty
.
Key
w
ords
:
a
d
v
a
n
c
e
d
s
ta
n
d
a
rd
e
n
c
r
y
p
ti
o
n
,
c
l
o
u
d
d
a
ta
b
a
s
e
,
c
l
o
u
d
q
u
e
r
y
p
ro
c
e
s
s
i
n
g
,
s
u
p
p
o
rt
v
e
c
t
o
r m
a
c
h
i
n
e
Copy
righ
t
©
2
0
1
9
Uni
v
e
rsi
t
a
s
Ahm
a
d
D
a
hl
a
n.
All
righ
t
s
r
e
s
e
rve
d
.
1.
Int
r
o
d
u
ctio
n
T
he
em
ergenc
e
of
c
l
ou
d
c
om
pu
ti
ng
s
er
v
i
c
es
s
uc
h
as
Har
d
w
ar
e
-
as
-
a
-
S
er
v
i
c
e
(
H
A
A
S
)
,
Inf
r
as
tr
uc
ture
-
as
-
a
-
S
erv
i
c
e
(
IA
A
S
)
an
d
S
of
tware
-
as
-
a
-
S
erv
i
c
e
(
S
A
A
S
)
ha
v
e
c
on
tr
i
bu
t
ed
to
war
d
f
ac
i
l
i
tat
i
n
g
da
i
l
y
-
ba
s
i
s
bu
s
i
ne
s
s
's
tr
an
s
ac
ti
on
s
[1,
2]
.
I
n
the
pa
s
t,
a
n
organ
i
z
at
i
on
i
s
r
eq
ui
r
ed
t
o
prov
i
de
a
f
ul
l
-
s
up
po
r
t
ha
r
d
war
e,
pl
atf
or
m
an
d
s
of
t
w
ar
e
f
or
ac
c
om
pl
i
s
hi
n
g
i
ts
g
oa
l
s
.
T
hi
s
c
om
es
wi
th
hi
gh
ex
p
en
s
i
v
e
c
os
t
of
i
ns
tal
l
at
i
on
,
l
i
c
en
s
i
ng
a
n
d
m
ai
nte
na
nc
e
[3]
.
T
h
eref
ore,
th
e
c
l
o
ud
c
o
m
pu
ti
n
g
s
er
v
i
c
es
ha
v
e
of
f
ered
a
great
o
pp
ort
un
i
t
y
wh
i
c
h
r
e
pres
en
t
ed
b
y
en
a
bl
i
ng
organ
i
z
ati
on
s
a
nd
c
orp
orati
on
s
to
us
e s
pe
c
i
f
i
c
s
erv
i
c
e
pe
r
us
ag
e
[4]
.
T
he
Data
ba
s
e
-
as
-
a
-
S
erv
i
c
e
(
DA
A
S
)
w
as
o
ne
of
the
s
erv
i
c
es
t
ha
t
h
av
e
b
ee
n
of
f
e
r
ed
b
y
s
ev
era
l
c
l
ou
d
s
erv
i
c
e
pro
v
i
de
r
s
[5
]
.
D
A
A
S
f
ac
i
l
i
t
ate
s
the
proc
es
s
of
i
n
i
t
i
at
i
ng
a
da
tab
as
e
f
or
a
s
pe
c
i
f
i
c
organi
z
at
i
o
n
w
he
r
e
s
uc
h
organ
i
z
at
i
o
n
i
s
ab
l
e
to
us
e
s
uc
h
da
tab
as
e
i
nc
l
u
di
n
g
qu
er
y
i
ng
an
d
s
tori
ng
tas
k
s
w
i
th
ou
t
bo
th
erin
g
the
ex
p
en
s
es
of
m
ai
nte
n
an
c
e
a
nd
b
ac
k
up
op
erat
i
on
s
[6
]
.
Ho
w
e
v
er,
t
hi
s
h
as
po
s
ed
a
s
erio
us
c
h
al
l
en
gi
ng
i
s
s
u
e
w
hi
c
h
i
s
t
h
e
pri
v
ac
y
[7
]
.
In
s
om
e
f
i
el
ds
,
i
t
i
s
hi
gh
l
y
r
i
s
k
y
t
o
l
et
t
he
i
r
o
w
n
da
ta
v
ul
ne
r
ab
l
e
t
o
be
v
i
ol
ate
d
b
y
a
n
y
t
hi
r
d
pa
r
t
y
s
uc
h
as
the
m
ed
i
c
al
do
m
ai
n
[8
]
.
T
he
r
ef
ore,
org
an
i
z
a
ti
o
ns
ten
d
to
us
e
an
en
c
r
y
p
ti
on
t
a
s
k
i
n
order
t
o
protec
t th
ei
r
o
w
n
da
ta
.
A
pp
arent
l
y
,
t
hi
s
w
o
ul
d
c
h
al
l
en
g
e
the
us
e
of
c
l
ou
d
c
o
m
pu
ti
n
g
s
erv
i
c
es
es
pe
c
i
al
l
y
f
or
the
D
A
A
S
w
h
ere
th
e
orga
n
i
z
ati
on
s
are
ac
c
om
m
od
ati
n
g
r
eg
u
l
ar
tas
k
s
da
i
l
y
s
uc
h
as
the
s
tori
ng
,
ba
c
k
i
ng
up
a
nd
qu
er
y
i
ng
.
Cons
i
d
erin
g
th
e
en
c
r
y
pt
i
on
an
d
de
c
r
y
pt
i
o
n
r
eq
ui
r
e
d
to
pe
r
f
or
m
the
s
e
tas
k
s
,
a
s
i
gn
i
f
i
c
an
t
ti
m
e
c
on
s
um
i
ng
w
o
ul
d
i
nd
ee
d
oc
c
ur
[9]
.
In
order
to
i
m
prov
e
t
he
ef
f
i
c
i
en
c
y
of
s
uc
h
i
s
s
ue
,
s
ev
eral
r
es
ea
r
c
he
r
s
ha
v
e
prop
os
ed
v
ario
u
s
t
y
p
es
of
en
c
r
y
p
ti
on
m
eth
od
s
th
at
wou
l
d
ha
v
e
ef
f
i
c
i
en
t
pe
r
f
orm
an
c
e
i
n
t
erm
s
o
f
the
ti
m
e
an
d
m
e
m
o
r
y
c
o
ns
um
pti
on
[1
0
-
12]
.
H
o
w
e
v
er,
us
i
ng
l
i
g
ht
en
c
r
y
p
ti
on
w
o
ul
d
l
ea
d
to
s
ev
eral
p
ote
nti
al
att
em
pts
t
o
v
i
ol
ate
t
he
da
ta.
T
he
r
ef
ore,
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
M
NIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
be
r
20
19
:
31
55
-
31
60
3156
s
o
m
e
au
th
ors
ha
v
e
r
ec
e
n
tl
y
ex
am
i
ne
d
a
tr
ad
e
-
of
f
m
e
c
ha
ni
s
m
i
n
w
hi
c
h
th
e
da
ta
i
s
b
ei
n
g
c
l
as
s
i
f
i
ed
f
i
r
s
tl
y
i
n
t
erm
s
of
c
on
f
i
de
nti
al
i
t
y
an
d
t
he
n
b
a
s
ed
on
i
ts
s
tat
us
a
pa
r
t
i
a
l
en
c
r
y
pt
i
o
n
wi
l
l
tak
e
a
p
l
ac
e.
H
o
w
e
v
er,
m
o
s
t
of
the
c
l
as
s
i
f
i
c
ati
on
m
eth
od
s
us
ed
t
o
c
at
eg
or
i
z
e
t
he
c
on
f
i
de
nt
i
a
l
da
ta
wer
e re
l
y
i
n
g o
n s
tr
i
ng
-
ba
s
ed
m
atc
hi
ng
ap
proac
h.
V
ari
ou
s
r
es
ea
r
c
h
s
tu
di
es
ha
v
e
ex
am
i
ne
d
th
e
ef
f
i
c
i
en
c
y
of
c
l
ou
d
d
ata
s
t
ora
ge
f
or
ex
am
pl
e,
W
an
g
et
a
l
.
[1
3]
pres
en
te
d
a
r
a
nk
i
ng
a
pp
r
o
ac
h
f
or
i
m
prov
i
ng
the
ef
f
i
c
i
en
c
y
of
s
ea
r
c
h
wi
th
i
n
da
t
a
s
tor
ed
i
n
th
e
c
l
ou
d.
In
f
ac
t,
th
e
s
ea
r
c
h
w
i
t
hi
n
an
e
nc
r
y
pt
ed
da
t
a
i
s
to
o
c
om
pl
i
c
ate
d
an
d
m
a
y
y
i
e
l
d
i
na
c
c
urat
e
r
es
ul
ts
.
T
he
r
ef
ore,
the
au
t
h
ors
ha
v
e
too
k
the
a
dv
an
t
a
ge
of
s
tat
i
s
t
i
c
al
tec
hn
i
qu
es
s
uc
h
as
term
fr
eq
ue
nc
y
a
nd
m
utu
al
i
nf
orm
ati
o
n
an
d
i
n
order
to
r
a
nk
t
he
do
c
um
en
ts
wi
th
i
n
th
e
c
l
ou
d
da
ta
.
In
t
hi
s
r
eg
ard,
t
he
s
ea
r
c
h
qu
e
r
y
t
y
p
ed
b
y
the
us
er
w
i
l
l
b
e
ex
am
i
ne
d
i
n
term
s
of
the
term
f
r
eq
ue
nc
y
an
d
m
utu
al
i
nf
orm
ati
on
i
n
ord
er
to
r
etri
ev
e
th
e
m
os
t
r
el
e
v
an
t
do
c
um
en
ts
.
In
a
dd
i
ti
o
n,
R
en
et
al
.
[10
]
propos
e
d
a
n
ef
f
i
c
i
en
t
qu
er
y
proc
es
s
i
ng
o
v
er
t
he
c
l
ou
d
ba
s
ed
on
a
k
-
ne
ares
t
ne
i
g
hb
o
u
r
c
l
as
s
i
f
i
c
ati
on
m
eth
od
.
T
he
propos
ed
m
eth
od
a
i
m
s
to
i
nd
ex
the
do
c
um
en
ts
w
i
t
hi
n
t
he
c
l
ou
d
s
torag
e
i
n
ord
er
to
e
ff
i
c
i
en
tl
y
an
d
ef
f
ec
ti
v
el
y
r
e
t
r
i
ev
e
t
he
m
os
t
r
el
e
v
an
t
da
ta.
Me
a
n
w
h
i
l
e,
t
he
a
uth
ors
ha
v
e
us
ed
t
he
r
an
do
m
pe
r
turba
ti
on
a
pp
r
o
a
c
h
i
n
order
t
o
i
ns
ure o
pt
i
m
al
c
on
f
i
de
nt
i
a
l
i
t
y
.
A
pa
r
t
f
r
om
the
en
c
r
y
pt
i
o
n
,
s
om
e
r
es
ea
r
c
he
r
s
ha
v
e
att
em
pte
d
to
c
l
as
s
i
f
y
th
e
da
t
a
ba
s
ed
o
n
the
c
on
f
i
d
en
t
i
a
l
i
t
y
prio
r
to
m
i
grate
th
e
d
ata
i
nto
th
e
c
l
ou
d.
G
r
ae
p
el
et
a
l
.
[1
4]
pres
en
te
d
a
c
l
as
s
i
f
i
c
at
i
on
m
eth
od
f
or
c
ate
go
r
i
z
i
n
g
t
h
e
d
ata
prio
r
to
the
m
i
grati
on
.
S
i
m
i
l
arl
y
,
Z
ardar
i
et
a
l
.
[1
5]
prop
os
e
d
a
c
l
as
s
i
f
i
c
ati
o
n
tec
hn
i
qu
e
f
or
di
s
ti
ng
u
i
s
hi
ng
th
e
c
on
f
i
de
nt
i
a
l
i
t
y
of
the
d
ata
.
T
he
i
r
te
c
hn
i
q
ue
was
i
nt
en
d
ed
t
o p
r
o
v
i
de
s
e
v
e
r
al
c
l
as
s
es
f
or the
c
o
nf
i
de
nt
i
al
i
t
y
.
Rec
en
t
l
y
,
A
l
b
ad
r
i
&
S
u
l
a
i
m
an
[16]
ha
v
e
ex
am
i
ne
d
the
c
l
as
s
i
f
i
c
ati
on
of
da
ta
be
f
ore
m
i
grati
ng
i
t
i
nto
the
c
l
ou
d
wi
th
a
r
u
l
e
-
ba
s
e
d
c
l
as
s
i
f
i
c
ati
o
n
tec
h
ni
qu
e.
T
he
au
t
h
ors
ha
v
e
us
e
d
a
r
ea
l
-
t
i
m
e
da
ta
of
s
tud
en
t
s
an
d
m
an
ua
l
l
y
an
n
ota
t
e
e
ac
h
da
t
a
i
ns
t
an
c
e
i
nto
t
he
i
r
c
on
f
i
de
nt
i
a
l
i
t
y
s
tat
us
.
Cons
eq
ue
nti
al
l
y
,
t
h
e
au
th
ors
ha
v
e
de
v
el
op
e
d
a
s
et
of
r
ul
es
i
n
order
to
di
s
ti
ng
u
i
s
h
the
d
ata
.
T
he
c
l
as
s
i
f
i
c
at
i
on
was
bu
i
l
t
b
as
ed
o
n
s
tr
i
n
g
-
b
as
ed
m
atc
hi
ng
am
on
g
th
e
da
ta
att
r
i
b
ute
s
.
B
as
ed
on
s
uc
h c
l
as
s
i
f
i
c
at
i
o
n,
a
pa
r
t
i
al
en
c
r
y
p
ti
on
ha
s
be
en
pe
r
f
orm
ed
f
or the
c
on
f
i
de
nti
al
da
t
a
i
n
order
to
r
e
du
c
e t
he
l
o
ad
of
qu
er
y
proc
es
s
i
n
g.
Z
ar
da
r
i
et
a
l
.
[17]
ha
v
e
ex
am
i
ne
d
the
r
ol
e
of
m
ac
hi
ne
l
ea
r
ni
ng
i
n
term
s
of
c
l
as
s
i
f
y
i
ng
c
on
f
i
de
nt
i
a
l
da
t
a
us
i
ng
K
N
N
c
l
as
s
i
f
i
c
ati
o
n
m
eth
od
.
T
he
p
r
op
os
ed
c
l
as
s
i
f
i
c
at
i
on
m
eth
o
d
ai
m
ed
to
i
de
nti
f
y
whi
c
h
da
ta
ne
ed
s
to
be
en
c
r
y
p
ted
ba
s
ed
on
i
ts
c
on
f
ed
i
n
ti
a
l
i
t
y
.
T
he
c
l
as
s
i
f
i
c
ati
on
was
r
el
y
i
n
g
on
s
tr
i
ng
-
ba
s
ed
s
i
m
i
l
arit
y
of
da
ta
at
tr
i
b
ute
s
.
Re
nu
et
a
l
.
[18]
ha
v
e
prop
os
e
d
a
bi
n
ar
y
tr
e
e
c
l
as
s
i
f
i
c
ati
on
m
eth
od
f
or
protec
ti
ng
c
o
nf
i
de
n
ti
a
l
da
t
a.
T
he
pro
po
s
ed
m
eth
od
i
s
b
as
ed
on
pre
-
de
f
i
ne
d
d
i
c
ti
o
na
r
y
al
on
g
w
i
t
h
a
s
tr
i
n
g
-
ba
s
e
d
m
atc
hi
n
g.
S
uc
h
d
i
c
ti
o
na
r
y
c
o
nta
i
ns
c
on
f
i
de
nti
a
l
da
ta
a
nd
th
e
s
tr
i
ng
-
b
as
e
d
m
atc
hi
ng
wi
l
l
c
om
pa
r
e
the
ne
w
d
ata
or
u
ns
e
en
da
t
a
w
i
t
h
the
pre
de
f
i
ne
d
on
es
.
F
r
om
the
l
i
t
erature
,
on
e
c
ou
l
d
no
t
i
c
e
tha
t
m
os
t
of
the
c
l
as
s
i
f
i
c
ati
on
m
eth
od
s
us
ed
f
or
c
ate
g
ori
z
i
ng
c
on
f
i
de
nti
al
da
ta
w
ere
r
el
y
i
ng
on
s
tr
i
ng
-
ba
s
e
d
m
atc
hi
n
g.
T
a
k
i
ng
the
a
dv
an
t
ag
e
of
oth
er
da
t
a
r
ep
r
es
e
nta
t
i
on
s
uc
h
as
N
-
gram
an
d
uti
l
i
z
i
ng
the
f
r
eq
ue
nc
y
of
term
s
woul
d
f
ac
i
l
i
tat
e
to
war
d
i
m
prov
i
ng
the
c
l
as
s
i
f
i
c
ati
o
n.
T
he
r
ef
ore,
thi
s
s
tud
y
ai
m
s
to
prop
os
e
a
bag
-
of
-
wor
d
(
B
o
W
)
r
ep
r
es
en
tat
i
o
n
or
so
-
c
al
l
e
d
N
-
gram
al
on
g
wi
th
S
up
po
r
t
V
ec
tor
Ma
c
hi
ne
(
S
V
M) c
l
as
s
i
f
i
er
f
or the
pro
c
es
s
of
c
on
f
i
de
nti
a
l
i
t
y
c
l
as
s
i
f
i
c
ati
o
n.
2
.
Rese
ar
ch M
eth
o
d
T
he
propos
e
d
m
eth
od
of
t
hi
s
s
tu
d
y
c
on
s
i
s
ts
of
f
i
v
e
s
tep
s
as
s
ho
w
n
i
n
F
i
gu
r
e
1
.
F
i
r
s
t
s
tep
i
s
r
e
l
ate
d
to
the
da
t
a
s
et
us
ed
i
n
t
he
ex
p
erim
en
t
whi
c
h
wi
l
l
b
e
us
e
d
f
or
the
c
l
as
s
i
f
i
c
ati
o
n.
Nex
t
s
tep
i
s
r
el
at
ed
to
t
he
ad
j
us
tm
en
t
r
eq
ui
r
ed
f
or
m
ak
i
ng
the
da
ta
s
u
i
ta
bl
e
f
or
th
e
c
l
as
s
i
f
i
c
ati
o
n
tas
k
.
S
uc
h
ad
j
us
tm
en
t
i
s
k
no
w
n
as
N
-
gram
r
ep
r
es
en
t
ati
o
n.
A
f
ter
tha
t,
th
e
c
l
as
s
i
f
i
c
ati
on
s
t
e
p
w
i
l
l
tak
e
pl
ac
e
b
y
c
ate
go
r
i
z
i
ng
the
d
ata
i
nto
c
o
nf
i
de
n
ti
a
l
a
nd
no
n
-
c
on
f
i
de
nti
al
.
B
as
e
d
on
the
r
es
ul
ts
of
s
uc
h
c
l
as
s
i
f
i
c
ati
on
,
a
n
e
nc
r
y
pti
on
pr
oc
es
s
wi
l
l
be
p
erf
or
m
ed
up
on
th
e
c
on
f
i
de
nti
a
l
d
ata
us
i
ng
A
d
v
an
c
e
d
E
nc
r
y
p
ted
S
tan
d
ard
(
A
E
S
)
.
F
i
na
l
l
y
,
an
e
v
al
ua
t
i
on
f
or
the
en
c
r
y
p
te
d
d
a
ta
wi
l
l
be
do
n
e
us
i
ng
qu
er
y
proc
es
s
i
ng
.
Ne
x
t s
ub
-
s
ec
ti
on
s
wi
l
l
ta
c
k
l
e e
ac
h s
tep
i
nd
ep
e
nd
en
t
l
y
.
2
.1.
Dat
a
T
he
da
ta
us
ed
i
n
thi
s
s
t
ud
y
i
s
the
on
e
tha
t
h
as
be
en
i
ntro
du
c
ed
b
y
A
l
b
ad
r
i
and
S
ul
ai
m
an
[1
6]
whi
c
h
c
o
n
s
i
s
ts
of
Uni
v
ers
i
t
y
s
tud
en
t's
i
nf
orm
ati
o
n
.
S
uc
h
da
t
a
c
on
ta
i
ns
r
e
gu
l
ar
i
nf
orm
ati
on
ab
o
ut
s
tud
en
t
s
s
uc
h
as
the
i
r
ba
s
i
c
i
nf
or
m
ati
on
a
nd
i
nf
or
m
ati
on
r
el
at
ed
to
the
i
r
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
Cl
as
s
i
fy
i
ng
c
on
fi
d
en
t
i
a
l
d
at
a u
s
i
n
g S
V
M f
or e
ffi
c
i
en
t c
l
ou
d
qu
ery
... (
H
ud
a
K
a
dh
i
m
T
ay
y
eh
)
3157
c
ou
r
s
es
,
grad
i
ng
an
d
p
a
y
m
en
ts
.
T
he
da
t
a
h
as
be
en
m
an
ua
l
l
y
an
n
ota
t
ed
ba
s
e
d
on
f
ou
r
c
l
as
s
l
ab
el
s
i
nc
l
u
di
ng
(
i
)
s
e
ns
i
t
i
v
e
da
t
a,
(
i
i
)
Co
nf
i
de
nti
al
da
t
a,
(
i
i
i
)
I
nte
r
na
l
da
ta
,
an
d
(
i
v
)
pu
bl
i
c
da
t
a.
S
en
s
i
ti
v
e
da
t
a
i
s
r
el
ate
d
t
o
th
e
b
as
i
c
i
nf
or
m
ati
on
s
uc
h
as
d
ate
of
bi
r
t
h
a
nd
m
oth
er
na
m
e.
Conf
i
de
nti
al
d
ata
i
s
the
m
os
t
r
es
tr
i
c
ted
i
nf
orm
ati
on
w
h
i
c
h
i
s
r
el
a
te
d
to
t
he
p
a
y
m
e
nts
,
s
tud
e
nt's
grades
an
d
ot
he
r
i
nf
orm
ati
on
t
ha
t
i
s
no
t
to
l
erat
ed
to
be
v
i
ol
ate
d
.
T
he
i
nt
erna
l
d
ata
i
s
t
he
da
t
a
tha
t
i
s
be
i
n
g
pe
r
m
i
tte
d
t
o
u
s
e
b
y
the
s
taf
f
of
the
un
i
v
e
r
s
i
t
y
s
uc
h
as
the
progr
es
s
r
ep
orts
of
the
s
tud
en
ts
on
l
y
.
F
i
na
l
l
y
,
th
e
p
u
bl
i
c
d
ata
w
h
i
c
h
i
s
t
he
no
r
m
al
i
nf
or
m
ati
on
th
at
i
s
n
ot
priv
ate
s
uc
h
as
the
na
m
e
of
the
s
tud
en
t.
I
n
f
ac
t,
e
ac
h
c
l
as
s
l
a
be
l
r
e
qu
i
r
es
s
pe
c
i
f
i
c
e
nc
r
y
pt
i
on
m
e
c
ha
ni
s
m
i
n
whi
c
h
the
m
os
t
c
on
f
i
de
nt
i
al
ne
e
ds
s
op
hi
s
ti
c
a
ted
en
c
r
y
pti
o
n
a
nd
v
i
c
e
v
ers
a.
T
ab
l
e
1
s
ho
w
s
t
he
de
ta
i
l
s
of
th
e d
a
tas
et.
2
.2
.
N
-
g
r
am Rep
r
es
ent
atio
n
In
ord
er
to
en
ab
l
e
the
S
V
M
c
l
as
s
i
f
i
c
at
i
on
,
i
t
i
s
ne
c
es
s
ar
y
to
turn
th
e
d
ata
i
n
to
v
ec
tors
.
F
or
thi
s
pu
r
po
s
e,
t
he
N
-
g
r
am
r
ep
r
es
en
tat
i
on
ha
s
b
ee
n
us
ed
.
S
uc
h
r
ep
r
es
en
t
ati
o
n
ai
m
s
to
proc
es
s
al
l
t
he
term
s
tha
t
ha
v
e
be
en
oc
c
urr
ed
wi
t
hi
n
th
e
d
ata
s
et
[1
9]
.
T
he
n,
th
e
di
s
ti
nc
t
oc
c
urr
en
c
e
of
ter
m
s
w
i
l
l
be
m
ai
nta
i
ne
d.
In
o
the
r
wor
d
s
,
the
r
ed
u
nd
a
nt
term
s
w
i
l
l
be
di
s
c
ard
ed
.
T
hi
s
i
s
to
i
ns
ure
tha
t
al
l
t
he
un
i
qu
e
term
s
are
be
i
ng
c
on
s
i
de
r
e
d.
A
f
ter
tha
t
,
th
e
un
n
ec
es
s
ar
y
term
s
w
i
l
l
be
r
em
ov
ed
s
uc
h
as
the
s
to
p
w
ords
.
T
hi
s
i
s
d
ue
to
the
i
r
i
ns
i
gn
i
f
i
c
an
t
i
m
p
ac
t
i
n
t
erm
s
o
f
de
term
i
ni
ng
t
he
c
l
as
s
l
ab
el
.
Henc
e,
al
l
t
he
s
e
term
s
w
i
l
l
be
us
e
d
as
c
ol
um
ns
or
att
r
i
bu
tes
w
h
ere
the
d
ata
i
ns
ta
nc
e
wi
l
l
b
e
ex
am
i
ne
i
n
t
erm
s
o
f
the
s
e
att
r
i
bu
tes
ba
s
ed
on
the
oc
c
urr
en
c
e.
T
ab
l
e
2
de
p
i
c
ts
an
ex
am
pl
e
of
th
i
s
r
ep
r
es
en
t
ati
on
.
A
s
s
ho
wn
i
n
T
ab
l
e
2,
ea
c
h
d
ata
i
ns
t
an
c
e
wi
l
l
be
ex
am
i
ne
d
i
n
t
erm
s
o
f
the
t
er
m
s
tha
t
l
oc
ate
d
i
n
t
he
c
o
l
um
ns
.
S
u
c
h
ex
am
i
na
ti
o
n
r
ef
ers
to
whet
h
er
the
i
ns
t
an
c
e
c
o
nta
i
n
th
i
s
term
o
r
no
t.
Con
tai
ni
ng
the
t
erm
w
i
l
l
b
e
r
e
pres
en
t
ed
as
‘
1
’
,
wh
i
l
e
th
e
ab
s
en
c
e
wi
l
l
be
r
ep
r
e
s
en
ted
as
‘
0
’
.
No
w
e
ac
h d
a
ta
i
ns
tan
c
e
w
i
l
l
be
r
ep
r
es
en
ted
as
a
v
ec
to
r
l
i
k
e t
he
f
ol
l
o
wi
ng
v
ec
tor:
00
00
10
0
00
0
0
i
n
thi
s
r
eg
ard,
ea
c
h
v
ec
t
or
w
i
l
l
c
on
tai
ns
a
v
a
l
ue
of
‘
1’
whi
c
h
r
ef
ers
to
the
oc
c
urr
en
c
e
of
a c
orr
es
po
nd
i
n
g t
erm
[20
, 2
1]
.
T
ab
l
e
1.
D
ata
s
et
Det
ai
l
s
D
e
s
c
r
ipt
ion
Qu
a
n
t
it
y
N
o
.
o
f
t
a
b
le
s
35
N
o
.
o
f
f
ield
s
362
N
o
.
o
f
c
la
s
s
lab
e
l
s
4
T
ab
l
e
2
. N
-
gram
Rep
r
es
en
t
ati
o
n
D
a
t
a
Ter
m
1
Ter
m
2
Ter
m
n
C
las
s
I
n
s
t
a
n
c
e
1
1
0
0
S
e
n
s
i
t
iv
e
I
n
s
t
a
n
c
e
2
0
1
0
I
n
t
e
r
n
a
l
I
n
s
t
a
n
c
e
3
0
0
0
P
u
b
li
c
I
n
s
t
a
n
c
e
4
0
0
1
P
u
b
li
c
I
n
s
t
a
n
c
e
5
0
0
0
P
u
b
li
c
2
.3
.
Clas
s
if
ica
t
ion
usin
g
S
V
M
T
hi
s
al
g
orit
hm
i
s
i
nt
en
d
ed
to
c
r
ea
t
e
a
tr
ai
n
i
n
g
m
od
el
ba
s
ed
on
ex
am
pl
es
of
the
da
t
a.
T
hi
s
m
ea
ns
tha
t
the
da
t
a
i
ns
tan
c
es
w
i
t
hi
n
the
da
t
as
e
t
tha
t
i
s
be
i
ng
gi
v
en
a
c
l
as
s
l
ab
e
l
w
i
l
l
be
us
ed
f
or
the
tr
a
i
n
i
ng
[
22
]
.
S
uc
h
m
od
el
w
i
l
l
m
ak
e
the
al
go
r
i
thm
i
s
ab
l
e
to
pred
i
c
t
th
e
c
l
as
s
l
a
be
l
of
ea
c
h
da
ta
i
ns
tan
c
e.
T
he
predi
c
t
i
on
of
thi
s
al
go
r
i
t
hm
i
s
ba
s
ed
o
n
a
m
argi
n
whi
c
h
k
no
wn
as
H
y
p
erpl
an
e.
In
th
e
v
ec
tor
s
pa
c
e
w
he
r
e
the
da
ta
i
s
s
ee
n
as
v
ec
tor,
S
V
M
w
i
l
l
ai
m
to
as
s
i
gn
an
ac
c
urate
H
y
pe
r
p
l
a
ne
t
ha
t
i
s
di
v
i
d
i
ng
the
da
t
a
i
nto
t
wo
c
l
as
s
es
[23
,
24
]
.
F
i
gu
r
e
2
de
pi
c
ts
s
uc
h
di
v
i
s
i
on
b
y
th
e
H
y
p
erp
l
an
e.
T
he
w
a
y
of
c
o
m
pu
ti
n
g
s
uc
h
H
y
p
erpl
an
e
c
a
n
be
i
l
l
us
tr
ate
d
b
as
ed
o
n
the
f
ol
l
o
wi
ng
eq
u
at
i
on
.
(
⃗
)
=
(
(
⃗
×
⃗
⃗
⃗
)
+
)
=
{
+
1
:
(
⃗
×
⃗
⃗
⃗
)
+
>
0
−
1
:
ℎ
(
1)
2
.4
.
A
E
S
E
n
c
r
y
p
t
ion
A
f
ter
c
l
as
s
i
f
y
i
n
g
th
e
c
on
f
i
d
en
ti
al
da
t
a
us
i
ng
t
he
pro
po
s
ed
S
V
M,
an
en
c
r
y
pt
i
o
n
ta
s
k
w
i
l
l
tak
e
a
p
l
ac
e.
F
or
t
hi
s
pu
r
po
s
e,
t
he
A
d
v
a
nc
e
E
nc
r
y
pti
o
n
S
ta
nd
ard
wi
l
l
b
e
us
ed
t
o
e
nc
r
y
pt
the
c
on
f
i
de
nt
i
a
l
d
ata
.
A
E
S
ha
s
be
en
wi
d
el
y
us
ed
f
or e
nc
r
y
pti
on
pu
r
p
os
es
r
eg
ard
i
ng
to
i
ts
v
ar
i
ou
s
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
M
NIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
be
r
20
19
:
31
55
-
31
60
3158
k
e
y
l
en
g
ths
s
uc
h
as
1
28
bi
ts
,
19
2
bi
ts
a
nd
25
6
bi
ts
[
2
5]
.
B
as
i
c
al
l
y
,
th
e
af
orem
en
ti
on
ed
l
en
gth
s
wi
l
l
b
e
us
e
d
f
or
the
thre
e
c
l
as
s
es
Int
erna
l
,
S
e
ns
i
t
i
v
e
a
nd
C
on
f
i
de
nti
al
r
e
s
p
ec
ti
v
e
l
y
.
T
hi
s
i
s
s
i
nc
e
the
f
ou
r
th
c
l
as
s
l
a
be
l
w
i
l
l
no
t b
e
en
c
r
y
pt
ed
.
2
.5
.
E
v
aluat
ion
A
f
ter
en
c
r
y
p
ti
ng
the
c
on
f
i
de
nt
i
a
l
d
ata
,
a
q
ue
r
y
pro
c
es
s
i
ng
wi
l
l
tak
e
a
pl
ac
e
w
he
r
e
m
ul
ti
pl
e
t
y
p
es
of
q
ue
r
i
es
a
r
e
be
i
ng
us
ed
.
T
he
e
v
a
l
u
at
i
on
of
qu
er
y
proc
es
s
i
ng
w
i
l
l
b
e
b
as
ed
on
the
t
i
m
e c
on
s
u
m
ed
to
r
etri
e
v
e
or to e
x
ec
u
te
q
ue
r
i
es
.
F
i
gu
r
e
1.
T
he
pro
po
s
ed
m
eth
od
s
t
ep
s
F
i
g
ure
2.
H
y
p
erpl
an
e s
e
pa
r
ati
o
n
3
. Re
sult
s
a
nd
A
n
al
y
s
is
In
t
hi
s
s
ec
t
i
on
,
t
he
r
es
ul
ts
of
ti
m
e
c
on
s
um
pti
on
wi
l
l
b
e
d
ep
i
c
t
ed
.
A
s
m
en
ti
on
e
d
ea
r
l
i
er,
ba
s
ed
o
n
th
e
r
es
u
l
ts
of
c
l
as
s
i
f
i
c
ati
on
,
the
da
t
a
w
i
l
l
b
e
en
c
r
y
p
ted
i
n
ac
c
ordanc
e
to
ea
c
h
c
l
as
s
l
ab
el
.
F
or
t
hi
s
pu
r
po
s
e,
th
r
ee
t
y
pe
s
of
qu
er
y
wi
l
l
be
us
ed
i
n
t
h
e
ex
p
erim
en
ts
i
nc
l
u
di
ng
A
d
d,
S
el
ec
t
an
d
D
el
e
te
qu
erie
s
.
F
or
ea
c
h
t
y
p
e,
50
q
ue
r
i
es
wi
l
l
be
us
e
d
f
or
the
ev
a
l
ua
ti
on
.
In
ad
di
t
i
o
n,
the
ba
s
e
l
i
ne
e
nc
r
y
pt
i
o
n
of
[16]
wi
l
l
be
us
e
d
to
c
om
pa
r
e
i
ts
pe
r
f
or
m
an
c
e
ag
ai
ns
t
the
prop
os
ed
m
eth
od
.
F
i
g
ure
s
3
,
4
,
a
nd
5
w
i
l
l
s
ho
w th
e res
u
l
ts
of
ba
s
el
i
ne
al
on
g
wi
th
th
e
propos
ed
m
eth
od
f
or
ea
c
h t
y
pe
of
qu
er
i
es
.
F
or
the
A
dd
q
ue
r
y
r
es
u
l
ts
as
s
ho
wn
i
n
F
i
g
ure
3
,
ap
p
r
ox
i
m
ate
l
y
bo
t
h
t
he
ba
s
e
l
i
n
e
a
nd
the
propos
e
d
m
eth
od
ha
s
s
i
m
i
l
ar
pe
r
f
orm
an
c
e.
Ho
wev
er,
th
e
pro
po
s
e
d
m
eth
o
d
ha
s
s
l
i
gh
t
l
y
be
tte
r
p
erf
or
m
an
c
e
b
y
c
o
n
s
u
m
i
ng
l
es
s
ti
m
e.
In
ge
ne
r
al
,
th
e
ad
d
qu
er
y
i
s
the
m
os
t
qu
er
y
t
ha
t
c
on
s
um
e
ti
m
e
c
o
m
pa
r
ed
to
oth
er
q
ue
r
i
es
.
F
or
the
s
el
ec
t
qu
er
y
r
es
ul
ts
s
ho
wn
i
n
F
i
gu
r
e
4
,
the
pr
o
po
s
ed
m
eth
od
h
as
ou
tp
erf
or
m
ed
the
ba
s
e
l
i
n
e
b
y
c
on
s
um
i
ng
l
es
s
er
ti
m
e
.
S
i
m
i
l
arl
y
,
f
or
the
r
es
ul
ts
of
de
l
ete
q
u
er
y
s
ho
w
n
i
n
F
i
gu
r
e
5
,
the
prop
os
ed
m
eth
od
s
h
o
w
ed
s
up
er
i
or
pe
r
f
or
m
an
c
e c
o
m
pa
r
ed
to
t
he
b
as
el
i
ne
.
T
he
r
ea
s
on
be
h
i
n
d
t
he
s
up
erio
r
i
t
y
of
the
pro
po
s
ed
m
eth
od
l
i
es
on
the
ef
f
ec
ti
v
en
e
s
s
o
f
S
V
M
c
l
as
s
i
f
i
c
ati
on
m
eth
o
d
w
hi
c
h
f
ac
i
l
i
t
ate
d
to
w
ar
d
c
l
as
s
i
f
y
i
ng
the
da
ta
i
n
ac
c
ordanc
e
to
the
c
on
f
i
d
en
t
i
al
i
t
y
.
T
hi
s
ha
s
i
nc
r
ea
s
e
th
e
nu
m
be
r
of
da
ta
tha
t
m
a
y
ne
ed
l
es
s
er
en
c
r
y
pti
on
l
en
gth
or
ev
en
m
i
gh
t
n
ee
d
an
en
c
r
y
p
ti
o
n
at
al
l
.
I
n
f
ac
t,
the
wa
y
of
r
ep
r
es
en
t
i
ng
the
te
r
m
s
ba
s
ed
on
N
-
gram
ha
s
de
m
on
s
tr
ate
d
be
t
ter
pe
r
f
orm
an
c
e
of
c
l
a
s
s
i
f
i
c
ati
on
c
om
pa
r
ed
to
th
e
s
tr
i
n
g
-
ba
s
e
d
m
atc
hi
ng
r
ep
r
es
en
t
ati
on
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NIK
A
IS
S
N: 1
69
3
-
6
93
0
◼
Cl
as
s
i
fy
i
ng
c
on
fi
d
en
t
i
a
l
d
at
a u
s
i
n
g S
V
M f
or e
ffi
c
i
en
t c
l
ou
d
qu
ery
... (
H
ud
a
K
a
dh
i
m
T
ay
y
eh
)
3159
F
i
g
ure
3.
R
es
ul
ts
of
ad
d
q
u
er
y
f
or the
propos
ed
m
eth
o
d a
n
d t
h
e b
as
e
l
i
ne
F
i
g
ure
4.
R
es
ul
ts
of
s
el
ec
t
qu
er
y
f
or t
he
pro
po
s
e
d m
eth
od
an
d
the
ba
s
e
l
i
ne
F
i
g
ure
5.
R
es
ul
ts
of
de
l
et
e
qu
er
y
f
or t
he
pro
po
s
e
d m
eth
od
an
d
the
ba
s
e
l
i
ne
4.
Co
n
clusion
T
hi
s
pa
pe
r
ha
s
prop
os
ed
a
n
i
m
prov
ed
c
l
as
s
i
f
i
c
at
i
on
m
eth
o
d
f
or
c
ate
go
r
i
z
i
ng
c
on
f
i
de
nt
i
a
l
da
ta
us
i
ng
S
V
M
.
W
i
th
the
us
e
of
N
-
gra
m
r
ep
r
es
en
tat
i
on
,
S
V
M
ha
s
s
ho
wed
be
tte
r
c
l
as
s
i
f
i
c
ati
on
ac
c
urac
y
.
B
as
ed
on
s
uc
h
c
l
as
s
i
f
i
c
ati
o
n
r
es
ul
ts
,
an
en
c
r
y
pt
i
o
n
ha
s
be
e
n
c
o
nd
u
c
te
d
us
i
ng
the
A
E
S
a
l
go
r
i
thm
.
Res
u
l
ts
of
en
c
r
y
pt
i
on
s
ho
wed
th
at
the
pr
op
os
e
d
m
eth
od
h
as
ou
t
pe
r
f
orm
ed
the
ba
s
e
l
i
ne
en
c
r
y
p
ti
on
ba
s
ed
on
t
he
ef
f
i
c
i
en
c
y
of
qu
er
y
proc
es
s
i
ng
.
T
hi
s
e
m
ph
as
i
z
es
the
us
ef
ul
ne
s
s
of
N
-
gram
r
ep
r
es
en
tat
i
on
c
om
pa
r
ed
to
t
he
s
tr
i
n
g
-
ba
s
e
d
m
atc
hi
ng
when
c
ate
go
r
i
z
i
ng
c
on
f
i
de
nti
al
da
ta
. For
f
utu
r
e
s
tud
i
es
,
ad
dr
es
s
i
ng
di
f
f
erent e
nc
r
y
p
ti
on
m
eth
od
s
woul
d
y
i
e
l
d
be
t
ter p
erf
or
m
an
c
e.
Evaluation Warning : The document was created with Spire.PDF for Python.
◼
IS
S
N: 16
93
-
6
93
0
T
E
L
KO
M
NIK
A
V
ol
.
17
,
No
.
6,
D
ec
em
be
r
20
19
:
31
55
-
31
60
3160
Ref
er
en
ce
s
[1
]
Kru
tz
RL,
V
i
n
e
s
RD.
C
l
o
u
d
s
e
c
u
ri
ty
:
A
c
o
m
p
r
e
h
e
n
s
i
v
e
g
u
i
d
e
to
s
e
c
u
r
e
c
l
o
u
d
c
o
m
p
u
ti
n
g
:
W
i
l
e
y
Pu
b
l
i
s
h
i
n
g
.
2
0
1
0
.
[2
]
Be
h
re
n
d
TS
,
W
i
e
b
e
EN
,
L
o
n
d
o
n
J
E
.
,
J
o
h
n
s
o
n
EC.
Clo
u
d
c
o
m
p
u
t
i
n
g
a
d
o
p
ti
o
n
a
n
d
u
s
a
g
e
i
n
c
o
m
m
u
n
i
ty
c
o
l
l
e
g
e
s
.
Be
h
a
v
i
o
u
r &
In
fo
rm
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
.
2
0
1
1
;
30
:
2
3
1
-
2
4
0
.
[3
]
Dil
l
o
n
T
,
W
u
C
,
Cha
n
g
E.
Clo
u
d
c
o
m
p
u
ti
n
g
:
i
s
s
u
e
s
a
n
d
c
h
a
l
l
e
n
g
e
s
.
Ad
v
a
n
c
e
d
In
f
o
rm
a
ti
o
n
Net
w
o
rk
i
n
g
a
n
d
Ap
p
l
i
c
a
t
i
o
n
s
(AI
NA), 2
0
1
0
2
4
th
I
EEE
I
n
te
r
n
a
t
i
o
n
a
l
Con
f
e
re
n
c
e
.
2
0
1
0
:
27
-
33
.
[4
]
Zh
a
n
g
Q
,
Che
n
g
L
,
B
o
u
ta
b
a
R
.
Clo
u
d
c
o
m
p
u
ti
n
g
:
s
ta
te
-
of
-
t
h
e
-
a
rt
a
n
d
re
s
e
a
r
c
h
c
h
a
l
l
e
n
g
e
s
.
J
o
u
rn
a
l
o
f
i
n
te
rn
e
t
s
e
rv
i
c
e
s
a
n
d
a
p
p
l
i
c
a
t
i
o
n
s
.
2
0
1
0
;
1
:
7
-
18
.
[5
]
Curi
n
o
C
,
J
o
n
e
s
EP
,
Po
p
a
RA
,
M
a
l
v
i
y
a
N
,
W
u
E
,
M
a
d
d
e
n
S
,
Ba
l
a
k
r
i
s
h
n
a
n
H
,
Z
e
l
d
o
v
i
c
h
N
.
Rel
a
ti
o
n
a
l
c
l
o
u
d
:
A
d
a
ta
b
a
s
e
-
as
-
a
-
s
e
r
v
i
c
e
fo
r
th
e
c
l
o
u
d
.
5
th
B
i
e
n
n
i
a
l
Con
fe
re
n
c
e
o
n
In
n
o
v
a
ti
v
e
Da
ta
Sy
s
te
m
s
Res
e
a
r
c
h
.
2
0
1
1
.
[6
]
Al
Sh
e
h
r
i
W
.
Cl
o
u
d
D
a
ta
b
a
s
e
Dat
a
b
a
s
e
a
s
a
S
e
rv
i
c
e
.
In
te
r
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Dat
a
b
a
s
e
M
a
n
a
g
e
m
e
n
t
Sy
s
te
m
s
.
2
0
1
3
;
5
:
1
.
[7
]
Ag
ra
w
a
l
D,
El
Ab
b
a
d
i
A,
Em
e
k
c
i
F,
M
e
tw
a
l
l
y
A
.
Dat
a
b
a
s
e
m
a
n
a
g
e
m
e
n
t
a
s
a
s
e
r
v
i
c
e
:
Cha
l
l
e
n
g
e
s
a
n
d
o
p
p
o
rt
u
n
i
t
i
e
s
.
Dat
a
E
n
g
i
n
e
e
ri
n
g
,
2
0
0
9
.
ICDE'
0
9
.
IEE
E
2
5
th
I
n
te
rn
a
ti
o
n
a
l
C
o
n
fe
r
e
n
c
e
.
2009
:
1709
-
1
7
1
6
.
[8
]
Sa
m
m
o
u
r
M
,
Hus
s
i
n
B
,
O
th
m
a
n
M
FI
,
Doh
e
i
r
M
,
Al
Sh
a
i
k
h
d
e
e
b
B
,
T
a
l
i
b
MS
.
DN
S
T
u
n
n
e
l
i
n
g
:
a
R
e
v
i
e
w
o
n
F
e
a
tu
re
s
.
I
n
t.
J
.
En
g
.
Te
c
h
n
o
l
.
2018
;
7
:
1
-
5
.
[9
]
Be
th
e
n
c
o
u
r
t
J
,
Sa
h
a
i
A
,
W
a
te
rs
B
.
Cip
h
e
rte
x
t
-
p
o
l
i
c
y
a
tt
r
i
b
u
t
e
-
b
a
s
e
d
e
n
c
r
y
p
ti
o
n
.
S
e
c
u
r
i
ty
a
n
d
Pri
v
a
c
y
,
2
0
0
7
.
SP'
0
7
.
IEEE
Sy
m
p
o
s
i
u
m
.
2
0
0
7
:
321
-
3
3
4
.
[1
0
]
Ren
Y
,
X
u
J
,
W
a
n
g
J
,
Ki
m
JU
.
Des
i
g
n
a
te
d
-
v
e
ri
f
i
e
r
p
ro
v
a
b
l
e
d
a
ta
p
o
s
s
e
s
s
i
o
n
i
n
p
u
b
l
i
c
c
l
o
u
d
s
t
o
ra
g
e
.
In
te
rn
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
Se
c
u
ri
t
y
a
n
d
I
ts
Ap
p
l
i
c
a
ti
o
n
s
.
20
13
;
7
:
11
-
20
.
[1
1
]
L
i
R,
L
i
u
AX
.
Ad
a
p
ti
v
e
l
y
s
e
c
u
re
c
o
n
j
u
n
c
ti
v
e
q
u
e
r
y
p
r
o
c
e
s
s
i
n
g
o
v
e
r
e
n
c
ry
p
te
d
d
a
t
a
fo
r
c
l
o
u
d
c
o
m
p
u
ti
n
g
.
2
0
1
7
I
EEE 3
3
r
d
I
n
t
e
rn
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
Da
ta
En
g
i
n
e
e
ri
n
g
(I
CD
E)
.
2
0
1
7
:
6
9
7
-
708
.
[1
2
]
Sa
h
i
n
C,
Al
l
a
rd
T
,
Ak
b
a
r
i
n
i
a
R,
El
Ab
b
a
d
i
A,
Pa
c
i
tt
i
E
.
A
Dif
f
e
re
n
ti
a
l
l
y
Pri
v
a
te
I
n
d
e
x
fo
r
Ra
n
g
e
Q
u
e
ry
Pro
c
e
s
s
i
n
g
i
n
Clo
u
d
s
.
2
0
1
8
I
EEE
3
4
th
In
te
rn
a
ti
o
n
a
l
Co
n
fe
r
e
n
c
e
o
n
Dat
a
En
g
i
n
e
e
r
i
n
g
(IC
DE)
.
2018
:
857
-
8
6
8
.
d
o
i
:1
0
.1
1
0
9
/I
CD
E.2
0
1
8
.0
0
0
8
2
.
[1
3
]
W
a
n
g
C,
Ca
o
N,
Ren
K,
L
o
u
W
.
E
n
a
b
l
i
n
g
s
e
c
u
re
a
n
d
e
ff
i
c
i
e
n
t
ra
n
k
e
d
k
e
y
w
o
rd
s
e
a
r
c
h
o
v
e
r o
u
ts
o
u
rc
e
d
c
l
o
u
d
d
a
ta
.
IEEE
Tra
n
s
a
c
ti
o
n
s
o
n
p
a
r
a
l
l
e
l
a
n
d
d
i
s
tri
b
u
te
d
s
y
s
te
m
s
.
2
0
1
2
;
23
(8
)
:
1
4
6
7
-
1
4
7
9
.
[1
4
]
G
ra
e
p
e
l
T
,
L
a
u
t
e
r
K,
N
a
e
h
ri
g
M
.
M
L
c
o
n
fi
d
e
n
ti
a
l
:
M
a
c
h
i
n
e
l
e
a
rn
i
n
g
o
n
e
n
c
ry
p
t
e
d
d
a
ta
.
I
n
te
rn
a
t
i
o
n
a
l
Con
fe
re
n
c
e
o
n
I
n
fo
r
m
a
ti
o
n
Se
c
u
ri
ty
a
n
d
Cry
p
to
l
o
g
y
.
2
0
1
2
:
1
-
21
.
[1
5
]
Za
rd
a
ri
M
A,
J
u
n
g
L
T
,
Za
k
a
ri
a
M
N
.
Hy
b
ri
d
M
u
l
ti
-
c
l
o
u
d
Da
ta
Se
c
u
ri
t
y
(HM
CD
S)
M
o
d
e
l
a
n
d
Dat
a
Cla
s
s
i
fi
c
a
t
i
o
n
.
Ad
v
a
n
c
e
d
C
o
m
p
u
te
r
S
c
i
e
n
c
e
Ap
p
l
i
c
a
t
i
o
n
s
a
n
d
T
e
c
h
n
o
l
o
g
i
e
s
(AC
SAT
),
2
0
1
3
In
te
rn
a
ti
o
n
a
l
Co
n
fe
re
n
c
e
.
2
013
:
166
-
171
.
[1
6
]
Al
b
a
d
r
i
H,
Su
l
a
i
m
a
n
R
.
A
Cla
s
s
i
fi
c
a
t
i
o
n
M
e
th
o
d
Fo
r
I
d
e
n
ti
fy
i
n
g
Con
fi
d
e
n
t
i
a
l
Da
ta
T
o
En
h
a
n
c
e
Eff
i
c
i
e
n
c
y
O
f
Q
u
e
ry
Pro
c
e
s
s
i
n
g
O
v
e
r
Clo
u
d
.
J
o
u
rn
a
l
o
f
Th
e
o
re
ti
c
a
l
&
Ap
p
l
i
e
d
In
fo
rm
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
.
2016
;
93
(2
)
:
4
1
2
-
420
.
[1
7
]
Za
rd
a
ri
M
A,
J
u
n
g
LT
.
Cla
s
s
i
fi
c
a
ti
o
n
o
f
F
i
l
e
Da
ta
Ba
s
e
d
o
n
Con
fi
d
e
n
ti
a
l
i
ty
i
n
C
l
o
u
d
C
o
m
p
u
ti
n
g
u
s
i
n
g
K
-
NN
Cl
a
s
s
i
fi
e
r
.
In
t
e
rn
a
t
i
o
n
a
l
J
o
u
rn
a
l
o
f
Bu
s
i
n
e
s
s
An
a
l
y
ti
c
s
(
IJ
BAN)
.
2
0
1
6
;
3
:
61
-
78
.
[1
8
]
Ren
u
S,
Ve
n
i
SK
.
An
En
h
a
n
c
e
d
CIA
tr
e
e
Us
i
n
g
S
tri
n
g
M
a
tc
h
i
n
g
Al
g
o
ri
th
m
.
In
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
Ap
p
l
i
e
d
En
g
i
n
e
e
ri
n
g
Re
s
e
a
rc
h
.
2017
;
12
:
6
1
2
3
-
6
1
2
6
.
[1
9
]
Ko
n
c
h
a
d
y
M.
T
e
x
t
m
i
n
i
n
g
a
p
p
l
i
c
a
t
i
o
n
p
ro
g
r
a
m
m
i
n
g
:
Cha
r
l
e
s
Riv
e
r M
e
d
i
a
,
In
c
.
2
0
0
6
.
[2
0
]
Su
n
Q
,
L
i
u
H
,
M
a
L
,
Z
h
a
n
g
T
.
A
n
o
v
e
l
h
i
e
ra
rc
h
i
c
a
l
B
a
g
-
of
-
W
o
rd
s
m
o
d
e
l
f
o
r
c
o
m
p
a
c
t
a
c
ti
o
n
re
p
re
s
e
n
ta
ti
o
n
.
Neu
r
o
c
o
m
p
u
ti
n
g
.
2
0
1
6
;
174
:
722
-
732.
[2
1
]
El
s
h
o
u
rb
a
g
y
M
,
Hem
a
y
e
d
E
,
Fa
y
e
k
M
.
En
h
a
n
c
e
d
b
a
g
o
f
w
o
rd
s
u
s
i
n
g
m
u
l
ti
l
e
v
e
l
k
-
m
e
a
n
s
fo
r
h
u
m
a
n
a
c
ti
v
i
ty
re
c
o
g
n
i
ti
o
n
.
E
g
y
p
ti
a
n
I
n
fo
rm
a
ti
c
s
J
o
u
r
n
a
l
.
2
0
1
6
;
17
:
227
-
2
3
7
.
[2
2
]
Y
i
n
C
,
X
i
a
n
g
J
,
Zh
a
n
g
H
,
W
a
n
g
J
,
Y
i
n
Z
,
Ki
m
JU
.
A n
e
w s
v
m
m
e
th
o
d
f
o
r s
h
o
rt
te
x
t
c
l
a
s
s
i
f
i
c
a
ti
o
n
b
a
s
e
d
o
n
s
e
m
i
-
s
u
p
e
rv
i
s
e
d
l
e
a
rn
i
n
g
.
Ad
v
a
n
c
e
d
In
f
o
rm
a
t
i
o
n
T
e
c
h
n
o
l
o
g
y
a
n
d
Se
n
s
o
r
A
p
p
l
i
c
a
t
i
o
n
(
AIT
S),
2
0
1
5
4
th
I
n
t
e
rn
a
ti
o
n
a
l
Co
n
fe
re
n
c
e
.
2015
:
100
-
103
.
[2
3
]
Pa
n
X
,
Y
a
n
g
Z
,
X
u
Y
,
W
a
n
g
L
.
Sa
f
e
s
c
re
e
n
i
n
g
ru
l
e
s
fo
r
a
c
c
e
l
e
ra
ti
n
g
tw
i
n
s
u
p
p
o
r
t
v
e
c
to
r
m
a
c
h
i
n
e
c
l
a
s
s
i
fi
c
a
ti
o
n
.
IEEE
tra
n
s
a
c
ti
o
n
s
o
n
n
e
u
ra
l
n
e
two
r
k
s
a
n
d
l
e
a
rn
i
n
g
s
y
s
te
m
s
.
2
0
1
8
;
29
:
1
8
7
6
-
1887
.
[2
4
]
Is
h
i
d
a
H
,
O
i
s
h
i
Y
,
M
o
ri
ta
K
,
M
o
ri
w
a
k
i
K
,
Na
k
a
j
i
m
a
T
Y
.
D
e
v
e
l
o
p
m
e
n
t
o
f
a
s
u
p
p
o
rt
v
e
c
t
o
r
m
a
c
h
i
n
e
b
a
s
e
d
c
l
o
u
d
d
e
t
e
c
t
i
o
n
m
e
t
h
o
d
fo
r
M
O
DIS
w
i
th
th
e
a
d
j
u
s
t
a
b
i
l
i
ty
to
v
a
ri
o
u
s
c
o
n
d
i
ti
o
n
s
.
Rem
o
te
s
e
n
s
i
n
g
o
f
e
n
v
i
ro
n
m
e
n
t
.
2
0
1
8
;
205
:
3
9
0
-
407
.
[2
5
]
Hoa
n
g
T
.
An
e
ff
i
c
i
e
n
t
FPG
A
i
m
p
l
e
m
e
n
t
a
ti
o
n
o
f
th
e
Ad
v
a
n
c
e
d
En
c
r
y
p
t
i
o
n
Sta
n
d
a
rd
a
l
g
o
ri
th
m
.
Com
p
u
t
i
n
g
a
n
d
Co
m
m
u
n
i
c
a
ti
o
n
T
e
c
h
n
o
l
o
g
i
e
s
,
Re
s
e
a
r
c
h
,
In
n
o
v
a
ti
o
n
,
a
n
d
Vi
s
i
o
n
f
o
r t
h
e
F
u
t
u
re
(RIVF
),
2
0
1
2
I
EEE
RIVF
In
t
e
rn
a
t
i
o
n
a
l
Con
fe
re
n
c
e
.
2
0
1
2
: 1
-
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.