I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
9
,
No
.
4
,
A
u
g
u
s
t
201
9
,
p
p
.
2
6
5
9
~
2
6
6
7
I
SS
N:
2
0
8
8
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v9
i
4
.
p
p
2
6
5
9
-
2667
2659
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
co
r
e
.
co
m/
jo
u
r
n
a
ls
/in
d
ex
.
p
h
p
/
I
JE
C
E
An
i
m
pro
v
ed
d
i
ff
erential
e
v
o
lution
a
lg
o
rith
m
f
o
r
d
a
ta
s
trea
m
c
lus
t
ering
B
ha
s
k
a
r
Ade
pu
1
,
J
a
y
a
dev
G
y
a
ni
2
,
G
.
Na
rsi
m
ha
3
1
De
p
a
rt
m
e
n
t
o
f
In
f
o
rm
a
ti
o
n
T
e
c
h
n
o
l
o
g
y
,
Ka
k
a
ti
y
a
In
stit
u
te
o
f
T
e
c
h
n
o
lo
g
y
&
S
c
ien
c
e
,
In
d
ia
2
De
p
a
rtme
n
t
o
f
CS
,
Co
l
leg
e
o
f
C
o
m
p
u
ter &
In
f
o
rm
a
ti
o
n
S
c
ien
c
e
s,
M
a
jm
a
a
h
Un
iv
e
rsit
y
,
S
a
u
d
i
A
ra
b
ia
3
De
p
a
rtme
n
t
o
f
CS
E,
JN
T
UH
Co
ll
e
g
e
o
f
En
g
in
e
e
rin
g
,
In
d
ia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
u
n
27
,
2
0
1
8
R
ev
i
s
ed
J
an
9
,
2
0
1
9
A
cc
ep
ted
J
an
11
,
2
0
1
9
A
F
e
w
a
l
g
o
rit
h
m
s
w
e
r
e
a
c
tu
a
li
z
e
d
b
y
th
e
a
n
a
ly
sts
f
o
r
p
e
r
f
o
rm
in
g
c
lu
ste
rin
g
o
f
d
a
ta
stre
a
m
s.
M
o
st
o
f
th
e
se
a
l
g
o
rit
h
m
s
re
q
u
ire
th
a
t
th
e
n
u
m
b
e
r
o
f
c
lu
ste
rs
(
K
)
h
a
s
to
b
e
f
ix
e
d
b
y
th
e
c
u
sto
m
e
r
b
a
se
d
o
n
i
n
p
u
t
d
a
ta
a
n
d
it
c
a
n
b
e
k
e
p
t
se
tt
led
a
ll
th
ro
u
g
h
th
e
c
lu
ste
ri
n
g
p
ro
c
e
ss
.
S
trea
m
c
lu
ste
rin
g
h
a
s
f
a
c
e
d
f
e
w
d
if
f
icu
lt
ies
in
p
ick
in
g
u
p
K
.
In
th
is
p
a
p
e
r,
w
e
p
ro
p
o
se
a
n
e
ff
icie
n
t
a
p
p
r
o
a
c
h
f
o
r
d
a
ta
stre
a
m
c
lu
ste
rin
g
b
y
e
m
b
ra
c
in
g
a
n
Im
p
ro
v
e
d
Diff
e
r
e
n
ti
a
l
Ev
o
lu
ti
o
n
(IDE)
a
lg
o
rit
h
m
.
T
h
e
IDE
a
lg
o
rit
h
m
is
o
n
e
o
f
th
e
q
u
ick
,
p
o
w
e
r
f
u
l
a
n
d
p
ro
d
u
c
ti
v
e
g
lo
b
a
l
o
p
t
im
iza
ti
o
n
a
p
p
r
o
a
c
h
f
o
r
p
r
o
g
ra
m
m
e
d
c
lu
ste
rin
g
.
In
o
u
r
p
ro
p
o
se
d
a
p
p
ro
a
c
h
,
w
e
a
d
d
it
io
n
a
ll
y
a
p
p
ly
a
n
e
n
tro
p
y
b
a
se
d
m
e
th
o
d
f
o
r
d
isti
n
g
u
is
h
in
g
th
e
c
o
n
c
e
p
t
d
rif
t
in
th
e
d
a
ta
stre
a
m
a
n
d
in
th
is
w
a
y
u
p
d
a
ti
n
g
th
e
c
lu
ste
rin
g
p
ro
c
e
d
u
re
o
n
li
n
e
.
W
e
d
e
m
o
n
stra
ted
th
a
t
o
u
r
p
r
o
p
o
se
d
m
e
th
o
d
is
c
o
n
tras
ted
w
it
h
G
e
n
e
ti
c
A
l
g
o
rit
h
m
a
n
d
id
e
n
ti
f
ied
a
s
p
ro
f
icie
n
t
o
p
ti
m
iza
ti
o
n
a
lg
o
rit
h
m
.
T
h
e
p
e
rf
o
r
m
a
n
c
e
o
f
o
u
r
p
ro
p
o
se
d
tec
h
n
i
q
u
e
is
a
ss
e
ss
e
d
a
n
d
c
r
e
a
tes
th
e
a
c
c
u
ra
c
y
o
f
9
2
.
2
9
%
,
th
e
p
re
c
isio
n
is
8
6
.
9
6
%
,
re
c
a
ll
is 9
0
.
3
0
%
a
n
d
F
-
m
e
a
su
re
e
sti
m
a
t
e
is 8
8
.
6
0
%
.
K
ey
w
o
r
d
s
:
C
o
n
ce
p
t
d
r
if
t
Data
s
tr
ea
m
c
l
u
s
ter
in
g
Dif
f
er
en
t
ial
e
v
o
l
u
tio
n
E
n
co
d
in
g
s
ch
e
m
e
E
n
tr
o
p
y
t
h
eo
r
y
Co
p
y
rig
h
t
©
2
0
1
9
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
B
h
ask
ar
A
d
ep
u
,
Kak
ati
y
a
I
n
s
tit
u
te
o
f
T
ec
h
n
o
lo
g
y
&
Scie
n
ce
,
Af
f
iliated
to
Kak
at
i
y
a
U
n
i
v
er
s
it
y
,
W
ar
an
g
al
,
506015
-
I
n
d
ia
.
E
m
ail: b
h
a
s
k
ar
_
ad
ep
u
@
y
a
h
o
o
.
co
m
1.
I
NT
RO
D
UCT
I
O
N
No
w
ad
a
y
s
,
d
atastre
a
m
s
b
ec
o
m
es
a
n
i
m
p
o
r
tan
t
s
o
u
r
ce
o
f
d
ata.
I
n
r
ec
en
t
y
ea
r
s
,
m
u
lt
ip
le
o
r
g
an
iza
tio
n
s
g
en
er
ate
h
u
g
e
a
m
o
u
n
t
s
o
f
d
ata.
Data
s
tr
ea
m
ap
p
licatio
n
d
o
m
ai
n
s
in
c
lu
d
es
i
n
f
o
r
m
atio
n
a
n
al
y
s
i
s
in
n
et
w
o
r
k
d
ata
f
lo
w
m
o
n
ito
r
i
n
g
,
I
n
ter
n
et
o
f
T
h
in
g
s
(
I
o
T
)
ap
p
licatio
n
s
r
eg
u
lar
l
y
s
e
n
d
i
n
g
s
e
n
s
o
r
s
d
ata
,
w
eb
p
ag
e
ac
ce
s
s
an
d
w
eb
clic
k
in
f
o
r
m
a
tio
n
,
w
ea
t
h
er
f
o
r
ec
asti
n
g
in
f
o
r
m
a
ti
o
n
a
n
d
th
e
ec
o
n
o
m
ic
i
n
f
o
r
m
atio
n
p
r
o
d
u
ce
d
b
y
f
i
n
an
ce
an
d
s
ec
u
r
ities
co
m
p
a
n
ies
an
d
s
o
o
n
[
1
-
5
]
.
C
o
n
v
e
n
tio
n
al
d
ata
m
i
n
i
n
g
m
et
h
o
d
s
m
o
s
tl
y
f
o
cu
s
ed
o
n
m
i
n
in
g
s
tatic
a
n
d
m
e
m
o
r
y
r
e
s
id
en
t
d
ata
r
ep
o
s
ito
r
ies.
Ho
wev
er
,
w
ith
th
e
e
m
er
g
en
ce
o
f
d
ata
s
tr
ea
m
s
an
d
tech
n
o
lo
g
ical
d
ev
elo
p
m
en
t
s
ch
an
g
ed
t
h
e
w
a
y
p
eo
p
le
s
to
r
e,
p
r
o
ce
s
s
an
d
co
m
m
u
n
ic
ate
th
e
d
ata
[
6
]
.
Data
s
tr
ea
m
s
ar
e
te
m
p
o
r
ar
il
y
o
r
d
er
e
d
,
f
ast
c
h
an
g
i
n
g
,
in
f
i
n
it
e
an
d
m
ass
iv
el
y
p
o
ten
tial.
I
t
m
a
y
b
e
n
o
t
p
o
s
s
ib
le
to
s
to
r
e
th
e
en
tire
d
ata
s
tr
ea
m
in
to
m
e
m
o
r
y
[
7
]
.
Stre
a
m
m
in
in
g
h
a
s
to
d
ea
l
w
it
h
r
ap
id
an
d
d
y
n
a
m
ic
d
ata
w
it
h
r
ea
l
ti
m
e
p
r
o
ce
s
s
in
g
an
d
ai
m
i
n
g
at
e
x
tr
ac
tin
g
u
s
e
f
u
l
an
d
in
t
er
esti
n
g
p
atter
n
s
.
T
h
e
b
ig
g
e
s
t
ch
alle
n
g
e
i
s
f
i
n
d
in
g
v
alu
ab
le
i
n
f
o
r
m
atio
n
in
a
s
i
n
g
l
e
s
ca
n
n
in
g
o
f
m
as
s
i
v
e
d
ata
s
tr
ea
m
s
[
7
,
8
]
.
Var
io
u
s
alg
o
r
it
h
m
s
a
n
d
p
r
o
c
ed
u
r
es
p
r
o
p
o
s
ed
f
o
r
m
i
n
i
n
g
d
ata
s
tr
ea
m
s
ca
n
b
e
g
r
o
u
p
ed
in
to
t
w
o
g
r
o
u
p
s
o
f
tech
n
iq
u
e
s
.
On
e
g
r
o
u
p
o
f
alg
o
r
it
h
m
s
ca
n
ac
h
ie
v
e
d
esire
d
clu
s
ter
i
n
g
r
es
u
lts
,
b
u
t
i
n
s
u
f
f
ic
ien
c
y
o
f
d
ata
s
to
r
ag
e
ca
p
ac
it
y
w
h
ich
lead
s
u
s
to
p
r
o
ce
s
s
d
ata
d
y
n
a
m
ic
all
y
i
n
e
x
tr
ac
ti
n
g
k
n
o
w
le
d
g
e.
An
o
t
h
er
g
r
o
u
p
o
f
alg
o
r
ith
m
s
r
ef
er
s
to
s
tr
ea
m
i
n
g
o
f
d
ata
a
n
d
ap
p
lies
m
i
n
i
n
g
tech
n
iq
u
e
s
[
9
]
.
T
h
ese
t
w
o
g
r
o
u
p
s
o
f
tech
n
iq
u
e
s
ex
p
r
ess
s
o
m
e
d
i
f
f
icu
l
ties
i
n
clu
s
ter
i
n
d
ata
s
tr
ea
m
s
.
So
m
e
o
f
th
e
d
if
f
ic
u
ltie
s
in
cl
u
d
es:
v
is
itin
g
o
f
d
ata
o
n
ce
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
20
88
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
4
,
A
u
g
u
s
t 2
0
1
9
:
2
6
5
9
-
2667
2660
d
u
r
in
g
t
h
e
p
r
o
ce
s
s
i
n
g
o
f
d
ata
s
tr
ea
m
,
p
er
f
o
r
m
an
ce
o
f
p
r
o
ce
s
s
in
g
s
tr
ea
m
is
cr
u
cial
a
n
d
d
ete
ctin
g
a
c
h
an
g
e
o
r
a
co
n
ce
p
t
d
r
if
t
d
u
r
in
g
th
e
t
i
m
e
w
h
et
h
er
g
r
ad
u
al
o
r
ab
r
u
p
t
in
t
h
e
ev
o
l
u
tio
n
ar
y
d
ata
s
tr
ea
m
is
d
i
f
f
icu
l
t
[
1
0
]
.
I
n
t
h
e
ar
ea
o
f
d
ata
s
tr
ea
m
cl
u
s
ter
in
g
,
m
u
ltip
le
tec
h
n
iq
u
es
h
av
e
b
ee
n
p
r
o
p
o
s
ed
.
T
h
ey
ar
e
K
-
m
ea
n
s
cl
u
s
ter
in
g
alg
o
r
ith
m
[
1
1
-
1
3
]
,
Fas
t
E
v
o
l
u
tio
n
ar
y
A
l
g
o
r
ith
m
f
o
r
C
lu
s
t
er
in
g
(
FE
A
C
)
[
1
4
]
,
A
S
u
p
p
o
r
t
Vec
to
r
C
l
u
s
ter
i
n
g
(
SVC
)
b
ased
alg
o
r
ith
m
[
1
5
,
1
6
]
,
Mu
lticlas
s
No
v
elt
y
Dete
ctio
n
(
MN
D)
alg
o
r
ith
m
[
1
7
,
1
8
]
,
Fu
zz
y
C
-
Me
a
n
alg
o
r
ith
m
[
1
,
19
]
,
B
I
R
C
H
alg
o
r
ith
m
[
20
]
,
KNN
(
K
-
Nea
r
est
Neig
h
b
o
r
)
alg
o
r
ith
m
[
8
]
.
Fo
r
clu
s
ter
i
n
g
d
ata
in
m
a
n
y
r
ea
l
w
o
r
ld
p
r
o
b
lem
s
,
K
-
m
ea
n
s
a
lg
o
r
ith
m
i
s
p
o
p
u
lar
l
y
u
s
ed
b
ec
au
s
e
o
f
it
s
s
i
m
p
lic
it
y
a
n
d
s
ca
lab
ilit
y
o
f
t
h
at
alg
o
r
ith
m
f
o
r
m
o
s
t
o
f
th
e
r
ea
l
w
o
r
ld
ap
p
lica
tio
n
s
.
K
-
m
ea
n
s
al
g
o
r
ith
m
h
a
s
a
li
m
ita
tio
n
th
a
t
w
e
h
av
e
to
s
p
ec
if
y
t
h
e
n
u
m
b
er
o
f
clu
s
ter
s
i.
e.
k
-
v
a
lu
e
as
an
i
n
p
u
t
p
ar
a
m
e
ter
to
th
e
alg
o
r
ith
m
[
1
1
-
1
3
]
.
Fo
r
au
to
m
a
ticall
y
e
s
ti
m
ati
n
g
th
e
n
u
m
b
er
o
f
clu
s
t
er
s
(
i.e
.
k
-
v
alu
e)
f
r
o
m
t
h
e
in
p
u
t
d
ata,
th
er
e
is
an
alg
o
r
it
h
m
k
n
o
w
n
as
Fas
t
E
v
o
lu
tio
n
ar
y
A
l
g
o
r
ith
m
f
o
r
C
lu
s
ter
in
g
(
FE
AC
)
[
1
4
]
h
as
s
h
o
w
n
to
b
e
e
f
f
icie
n
t.
T
h
e
lim
i
tatio
n
o
f
t
h
is
al
g
o
r
ith
m
w
as
th
a
t
it
is
n
o
t
ap
p
lied
f
o
r
d
ata
s
tr
ea
m
an
al
y
s
i
s
.
Su
p
p
o
r
t
Vec
to
r
C
lu
s
ter
i
n
g
(
SVC
)
al
g
o
r
ith
m
i
s
a
n
e
f
f
ici
en
t
a
n
d
ef
f
ec
ti
v
e
d
ata
s
tr
ea
m
clu
s
ter
in
g
al
g
o
r
ith
m
.
T
h
e
d
i
s
ad
v
an
tag
e
o
f
t
h
i
s
alg
o
r
ith
m
i
s
th
a
t
it
ca
n
n
o
t
f
i
n
d
ar
b
itar
y
s
h
ap
e
cl
u
s
ter
s
b
ec
au
s
e
m
o
s
t
o
f
th
e
s
e
al
g
o
r
ith
m
s
ar
e
b
a
s
ed
o
n
K
-
m
ea
n
s
alg
o
r
ith
m
s
[
1
5
,
1
6
]
.
Mu
lticl
ass
No
v
elt
y
Dete
ctio
n
alg
o
r
ith
m
ca
n
b
e
u
s
ed
in
v
ar
io
u
s
ap
p
licati
o
n
s
lik
e
in
tr
u
s
io
n
,
f
a
u
lt
a
n
d
f
r
au
d
d
etec
tio
n
s
,
s
p
a
m
f
ilter
s
an
d
in
tex
t
m
i
n
i
n
g
[
1
7
,
1
8
]
.
B
I
R
C
H
is
a
h
ier
ar
ch
ical
clu
s
ter
i
n
g
alg
o
r
it
h
m
w
h
ic
h
is
b
ased
o
n
ca
lcu
lati
n
g
t
h
e
d
is
ta
n
ce
b
et
w
ee
n
t
h
e
n
e
w
d
ata
p
o
in
t
an
d
th
e
r
e
m
ai
n
in
g
k
n
o
w
n
d
atap
o
in
t
s
.
Af
ter
t
h
at,
co
m
p
ar
e
t
h
ese
d
is
ta
n
ce
s
w
it
h
a
th
r
es
h
o
ld
to
d
eter
m
in
e
th
e
c
ateg
o
r
y
o
f
n
e
w
d
ata
p
o
in
t.
T
h
e
li
m
itatio
n
o
f
th
is
alg
o
r
ith
m
is
t
h
at
it
d
o
es
n
o
t
f
u
n
ctio
n
ef
f
ec
ti
v
el
y
f
o
r
th
e
d
ata
w
it
h
ar
b
itra
r
y
s
h
ap
e
[
20
]
.
KNN
alg
o
r
it
h
m
i
s
a
s
k
e
w
ed
ap
p
r
o
ac
h
an
d
is
w
id
el
y
u
s
ed
m
eth
o
d
f
o
r
s
o
lv
i
n
g
clas
s
i
f
icatio
n
an
d
p
atter
n
r
ec
o
g
n
itio
n
p
r
o
b
le
m
s
in
m
ac
h
i
n
e
lear
n
i
n
g
.
KN
N
alg
o
r
ith
m
is
u
s
ed
to
av
o
id
th
e
h
ig
h
co
m
p
u
tatio
n
al
co
m
p
le
x
it
y
ev
e
n
t
h
o
u
g
h
w
e
co
u
ld
n
o
t
g
et
a
s
a
tis
f
ac
to
r
y
p
er
f
o
r
m
a
n
ce
i
n
m
a
n
y
ap
p
licatio
n
s
[
19,
21
-
2
4
]
.
I
n
th
is
p
ap
er
,
w
e
p
r
o
p
o
s
e
an
I
m
p
r
o
v
ed
Dif
f
er
en
tial
E
v
o
lu
t
i
o
n
alg
o
r
it
h
m
(
I
DE
)
f
o
r
t
h
e
d
ata
s
tr
ea
m
clu
s
ter
i
n
g
.
A
t
f
ir
s
t,
f
r
o
m
th
e
in
p
u
t
d
ata
s
tr
ea
m
s
,
t
h
e
n
ea
r
est
clu
s
ter
ce
n
ter
i
s
ass
e
s
s
ed
f
o
r
ev
er
y
i
n
co
m
in
g
o
b
j
ec
t
an
d
th
e
clu
s
ter
s
ar
e
u
p
d
ated
.
A
r
o
u
n
d
th
e
n
a
n
y
co
n
ce
p
t
d
r
if
t
o
c
cu
r
s
,
th
e
ap
p
r
o
ac
h
in
g
o
b
j
ec
ts
ar
e
p
u
t
i
n
th
e
b
u
f
f
er
till
a
s
ettled
ti
m
e
p
er
io
d
.
Fr
o
m
t
h
at
p
o
in
t
o
n
w
ar
d
s
,
u
s
e
th
e
I
DE
-
b
ased
o
p
ti
m
iza
tio
n
f
o
r
f
in
d
i
n
g
th
e
o
p
tim
a
l
K
v
alu
e.
O
n
t
h
e
o
f
f
ch
an
ce
th
at
a
n
y
co
n
ce
p
t d
r
if
t o
cc
u
r
s
,
th
e
u
n
d
er
l
y
i
n
g
ad
v
a
n
ce
s
a
r
e
r
eh
ash
ed
.
T
h
e
r
est
o
f
th
is
p
ap
er
is
p
o
r
tr
a
y
ed
in
t
h
e
s
e
g
m
e
n
t
u
n
d
er
n
ea
t
h
.
T
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
is
d
elin
ea
ted
i
n
s
ec
tio
n
2
.
T
h
e
o
v
er
w
ie
w
o
f
th
e
I
DE
alg
o
r
it
h
m
is
e
x
p
la
in
ed
in
s
ec
tio
n
2
.
2
.
Ou
r
p
r
o
p
o
s
ed
I
DE
Stre
am
alg
o
r
ith
m
is
d
eli
n
ea
ted
in
s
ec
ti
o
n
2
.
7
.
R
esu
lts
a
n
d
th
e
co
n
c
lu
s
io
n
w
er
e
d
ep
icted
in
s
ec
tio
n
s
3
an
d
4
.
2.
P
RO
P
O
SE
D
M
E
T
H
O
D
C
o
n
s
id
er
th
e
d
ata
s
tr
ea
m
w
h
ic
h
co
n
s
is
t
s
o
f
N
n
u
m
b
er
o
f
o
b
jects
an
d
ea
c
h
o
b
j
ec
t
is
a
n
l
-
d
i
m
en
s
io
n
a
l
f
ea
t
u
r
e
v
ec
to
r
x
i
=[
x
i
j
]
w
h
er
e
j
=1
to
l
an
d
1
<=
i
<=
N.
I
n
itiall
y
th
e
n
u
m
b
er
o
f
cl
u
s
ter
s
m
a
x
,
2
k
k
an
d
th
e
ce
n
tr
o
id
s
ar
e
r
a
n
d
o
m
l
y
s
e
lecte
d
f
r
o
m
t
h
e
g
iv
e
n
d
ata
s
tr
e
a
m
.
T
h
e
n
o
r
m
al
d
i
s
tan
ce
b
et
w
ee
n
clo
s
e
s
t
c
lu
s
ter
ce
n
ter
an
d
t
h
e
u
n
d
er
l
y
i
n
g
o
b
jects
ar
e
ev
al
u
ated
an
d
u
p
d
ated
.
T
h
e
ev
al
u
ated
clu
s
ter
ce
n
te
r
s
ar
e
u
p
d
ated
an
d
th
is
p
r
o
ce
d
u
r
e
is
r
eh
as
h
ed
u
n
t
il
th
e
p
o
in
t
w
h
en
s
o
m
e
ce
as
in
g
r
u
le
is
m
et.
A
t
t
h
e
m
ea
n
ti
m
e,
if
a
n
y
co
n
ce
p
t
d
r
if
t o
cc
u
r
s
,
it u
t
ilizes t
h
e
en
tr
o
p
y
t
h
eo
r
y
w
h
ich
i
s
p
r
esen
ted
in
th
e
s
ec
tio
n
2
.
1
.
Ou
r
p
r
o
p
o
s
ed
I
DE
alg
o
r
ith
m
d
o
esn
't r
eq
u
ir
e
th
e
s
tr
ea
m
in
g
m
o
d
u
le
to
s
to
r
e
t
h
e
d
ata
f
o
r
o
u
t
li
n
e
w
h
ic
h
u
tili
ze
s
t
h
e
cl
u
s
ter
i
n
g
m
o
d
u
l
e
o
r
m
o
d
u
le
f
o
r
esti
m
atio
n
o
f
k
-
v
al
u
e
f
o
r
ap
p
o
r
tio
n
in
g
t
h
e
d
ata.
Her
e,
t
h
e
u
p
d
ated
ap
p
o
r
tio
n
in
g
d
ata
ar
e
k
ep
t
u
p
b
y
u
t
ilizi
n
g
a
s
in
g
le
co
m
p
o
n
en
t;
it
is
d
o
n
e
i
n
th
e
o
n
li
n
e
m
o
d
e.
Dis
ti
n
cti
v
e
q
u
a
n
titi
e
s
o
f
tr
ail
ar
r
an
g
e
m
en
t
s
ar
e
u
t
ilized
f
o
r
o
u
r
ap
p
r
o
ac
h
,
w
h
ic
h
h
as
d
i
v
e
r
s
e
clu
s
ter
ce
n
ter
s
an
d
its
co
o
r
d
in
ates.
T
h
e
b
est
an
s
w
er
f
o
r
th
e
u
p
d
ated
clu
s
ter
in
g
d
ata
is
c
h
o
s
en
a
n
d
th
e
m
o
s
t
n
o
ticea
b
l
y
a
w
f
u
l
ar
r
an
g
e
m
en
t
s
ar
e
d
is
p
o
s
ed
o
f
in
th
e
li
g
h
t
o
f
th
e
I
DE
alg
o
r
it
h
m
.
I
n
th
e
d
ata
s
tr
ea
m
cl
u
s
te
r
in
g
p
r
o
ce
s
s
,
j
u
s
t
a
s
in
g
le
o
b
j
ec
t is to
u
ch
i
n
g
b
ase
at
o
n
ce
.
A
t
f
ir
s
t,
t
h
e
q
u
a
n
tit
y
o
f
clu
s
ter
s
i
s
ass
e
s
s
ed
b
y
I
DE
S
tr
ea
m
an
d
t
h
e
ea
ch
ev
alu
a
ted
clu
s
ter
m
ea
s
u
r
e
it
o
u
g
h
t
to
b
e
eq
u
iv
ale
n
t
to
th
e
u
n
d
er
l
y
in
g
s
ize
o
b
j
ec
ts
f
r
o
m
th
e
s
tr
ea
m
.
A
t
th
a
t
p
o
in
t,
th
e
ass
e
s
s
ed
cl
u
s
ter
is
k
ep
t u
p
in
an
o
n
l
in
e
m
a
n
n
er
,
s
u
b
s
eq
u
en
t to
b
u
ild
i
n
g
u
p
th
e
u
n
d
er
ly
i
n
g
clu
s
ter
.
T
h
e
p
r
o
ce
d
u
r
e
o
f
I
DE
Stre
a
m
al
g
o
r
ith
m
h
as
ap
p
ea
r
ed
in
Fig
u
r
e
1
.
A
t
t
h
e
u
n
d
er
l
y
in
g
p
h
ase
o
f
t
h
e
d
ata
s
tr
ea
m
p
r
o
ce
s
s
in
g
,
th
e
n
u
m
b
er
o
f
cl
u
s
ter
s
i
s
r
an
d
o
m
l
y
s
elec
ted
f
r
o
m
t
h
e
i
n
f
o
r
m
ati
o
n
o
f
d
ata
s
tr
ea
m
.
T
h
e
n
o
r
m
a
l
d
is
ta
n
ce
b
et
w
ee
n
clo
s
est
cl
u
s
ter
ce
n
ter
an
d
t
h
e
u
n
d
er
l
y
in
g
o
b
j
ec
ts
ar
e
ev
al
u
ated
an
d
u
p
d
ated
.
W
ith
s
p
ec
i
f
ic
g
o
al
to
r
ec
o
g
n
i
ze
ch
an
g
e
s
i
n
t
h
e
d
ata
p
ar
titi
o
n
,
th
e
cl
u
s
ter
s
ar
e
ad
m
in
i
s
te
r
ed
b
y
t
h
e
e
n
tr
o
p
y
th
eo
r
y
.
At
th
e
p
o
in
t
w
h
e
n
th
e
en
tr
o
p
y
test
tr
ig
g
er
s
an
alar
m
,
th
at
is
th
e
p
o
in
t
at
w
h
ic
h
th
e
r
ea
l
clu
s
ter
s
b
ein
g
u
p
d
ated
d
o
n
'
t
m
ir
r
o
r
th
e
ad
j
u
s
t
m
e
n
ts
in
t
h
e
d
ata
s
tr
ea
m
s
a
n
d
th
ese
cl
u
s
ter
ed
o
b
j
ec
ts
ar
e
p
u
t
i
n
th
e
b
u
f
f
er
f
o
r
s
o
m
e
ti
m
e
u
n
til
a
n
y
co
n
ce
p
t
d
r
if
t
o
r
ch
an
g
e
o
cc
u
r
s
.
A
t
th
at
p
o
in
t,
th
e
I
DE
s
ca
n
is
b
eg
u
n
f
o
r
o
p
ti
m
izatio
n
a
n
d
it is
d
ea
lt in
t
h
e
en
co
d
in
g
s
c
h
e
m
e.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
n
imp
r
o
ve
d
d
iffer
en
tia
l e
vo
lu
tio
n
a
lg
o
r
ith
m
fo
r
d
a
t
a
s
tr
ea
m
clu
s
teri
n
g
(
B
h
a
s
ka
r
A
d
ep
u
)
2661
2
.
1
.
E
ntr
o
py
t
heo
ry
f
o
r
det
ec
t
ing
co
ncept
drif
t
A
t
th
e
ti
m
e
o
f
p
ar
titi
o
n
in
g
t
h
e
d
ata,
th
e
d
ata
d
is
tr
ib
u
tio
n
m
a
y
ch
a
n
g
e
o
v
er
ti
m
e
i
n
u
n
f
o
r
es
ee
n
w
a
y
s
,
th
is
p
r
o
b
le
m
is
k
n
o
w
n
as
co
n
ce
p
t
d
r
if
t.
T
h
ec
o
n
ce
p
t
d
r
if
t
is
g
o
i
n
g
o
n
,
w
h
ile
co
n
ce
p
t
ev
o
l
u
tio
n
ap
p
ea
r
an
ce
to
v
an
i
s
h
in
g
o
f
cl
u
s
ter
s
.
T
o
r
ec
o
g
n
ize
t
h
e
co
n
ce
p
t
d
r
if
t,
t
h
e
ac
co
m
p
a
n
y
in
g
s
p
ec
u
latio
n
s
ar
e
co
n
s
id
er
ed
f
o
r
th
e
en
tr
o
p
y
t
h
eo
r
y
.
Her
e,
j
u
s
t
a
s
i
n
g
le
id
ea
is
h
o
led
b
y
t
h
e
d
ata
s
tr
ea
m
.
B
ec
au
s
e
o
f
th
i
s
r
ea
s
o
n
,
th
e
co
n
ce
p
t
d
r
if
t
is
n
'
t
h
ap
p
en
in
g
i
n
lig
h
t
o
f
th
e
f
ac
t
th
at
t
h
e
d
ata
s
tr
ea
m
is
s
te
ad
y
.
O
n
th
e
o
f
f
ch
a
n
ce
th
at
t
h
e
d
ata
s
tr
ea
m
d
o
es
n
o
t
h
a
v
e
a
n
y
co
n
ce
p
t
d
r
if
t,
t
h
e
ti
m
e
s
er
ies
o
f
th
e
e
n
t
r
o
p
y
s
tr
ateg
y
is
ab
o
u
t
ze
r
o
an
d
s
ta
y
s
s
tatio
n
ar
y
.
I
n
th
e
ev
en
t
th
at
th
e
d
ata
s
tr
ea
m
h
a
s
an
y
co
n
ce
p
t d
r
if
t,
t
h
e
en
tr
o
p
y
is
ca
n
'
t
ex
p
an
d
o
n
th
e
last
id
ea
,
it j
u
s
t c
o
o
r
d
in
ates
w
it
h
t
h
e
u
n
d
er
l
y
in
g
o
n
e.
T
h
at
i
m
p
lies
,
th
e
p
r
ec
is
io
n
j
u
d
g
m
en
t
o
f
th
e
b
ac
k
to
b
ac
k
d
ata
i
s
ca
n
'
t
co
n
tr
o
l
th
e
f
r
a
m
e
w
o
r
k
,
if
th
e
u
p
d
ati
n
g
i
s
n
o
t
f
in
i
s
h
ed
.
B
ec
au
s
e
o
f
th
is
r
ea
s
o
n
,
th
e
d
is
co
v
er
y
o
f
co
n
ce
p
t
d
r
if
t
is
cr
itical.
T
h
u
s
,
th
e
e
n
tr
o
p
y
esti
m
at
io
n
is
estab
lis
h
ed
i
n
th
e
li
g
h
t
o
f
t
h
e
m
e
m
b
er
s
h
ip
estee
m
s
a
n
d
it
is
cr
itical
f
o
r
id
en
ti
f
y
i
n
g
t
h
e
co
n
ce
p
t d
r
if
t.
I
n
o
u
r
m
et
h
o
d
s
h
a
n
n
o
n
’
s
e
n
tr
o
p
y
s
tr
ate
g
y
is
u
tili
ze
d
f
o
r
th
e
en
tr
o
p
y
e
s
ti
m
atio
n
.
Her
e,
th
e
d
is
cr
ete
r
an
d
o
m
v
ar
iab
les
ar
e
co
n
s
id
er
ed
as
X
an
d
th
e
co
n
ce
iv
ab
le
estee
m
s
ar
e
co
n
s
id
er
ed
as:
l
d
i
X
l
i
X
l
i
X
l
i
X
,
,
2
,
1
,
an
d
t
h
e
p
r
o
b
ab
ilit
y
m
a
s
s
f
u
n
c
ti
o
n
is
P
(
X
)
.
T
h
e
en
tr
o
p
y
E
(
X
)
is
ca
lc
u
lated
f
o
r
ev
er
y
r
an
d
o
m
v
ar
iab
le
X
u
s
in
g
th
e
f
o
llo
w
i
n
g
eq
u
atio
n
.
n
i
X
i
P
X
i
P
X
E
1
))
(
l
o
g
(
)
(
)
(
(
1
)
T
h
e
n
o
n
-
co
n
s
is
te
n
c
y
o
f
t
h
e
g
i
v
en
clu
s
ter
i
s
as
s
es
s
ed
in
v
ie
w
o
f
t
h
e
e
n
tr
o
p
y
m
ea
s
u
r
e.
Fo
r
th
e
al
l
-
o
u
t
d
ata
clu
s
ter
in
g
,
t
h
e
e
n
tr
o
p
y
t
h
eo
r
y
is
a
p
r
o
f
icien
t
tec
h
n
iq
u
e.
On
th
e
o
f
f
c
h
an
ce
th
a
t
t
h
e
e
n
tr
o
p
y
estee
m
s
ar
e
h
ig
h
,
th
e
v
u
ln
er
ab
ilit
y
o
f
t
h
e
I
DE
is
b
ig
g
er
.
T
h
e
ti
m
e
s
er
ies
o
f
th
e
clas
s
if
icatio
n
e
n
tr
o
p
y
p
o
in
ts
is
al
m
o
s
t
ze
r
o
if
t
h
e
d
ata
s
tr
ea
m
d
o
es n
o
t c
o
n
tain
an
y
co
n
ce
p
t d
r
if
t,
g
en
er
all
y
,
th
e
es
tee
m
w
i
ll t
u
r
n
o
u
t to
b
e
ex
ten
s
i
v
e.
2
.
2
.
O
v
er
v
ie
w
o
f
I
DE
Str
ea
m
a
lg
o
rit
h
m
Fig
u
r
e
1
s
h
o
w
n
t
h
e
p
r
o
ce
d
u
r
e
o
f
I
DE
S
tr
ea
m
al
g
o
r
ith
m
.
Fig
u
r
e
1
.
P
r
o
ce
s
s
o
f
I
DE
S
tr
ea
m
alg
o
r
it
h
m
2
.
3
.
B
uff
er
f
o
r
I
DE
S
t
re
a
m
I
n
t
h
e
w
a
k
e
o
f
r
ec
o
g
n
izi
n
g
t
h
e
co
n
ce
p
t
d
r
if
t,
t
h
e
g
en
u
i
n
e
c
lu
s
ter
s
ar
e
u
p
d
ated
,
an
d
t
h
ese
o
u
td
ated
clu
s
ter
s
ar
e
p
u
t
a
w
a
y
i
n
b
u
f
f
e
r
f
o
r
s
o
m
e
ti
m
e
to
r
u
n
th
e
I
D
E
alg
o
r
ith
m
.
Her
e,
th
e
b
ase
s
i
ze
o
f
th
e
b
u
f
f
er
is
1
0
%×in
itial
s
ize.
Fo
r
t
h
e
co
m
p
u
tatio
n
al
as
s
ets,
j
u
s
t
th
e
b
as
e
s
ize
o
f
th
e
b
u
f
f
er
is
ac
ce
s
s
i
b
le.
T
o
ch
ar
ac
ter
ize
th
e
w
ar
n
in
g
a
n
d
th
e
alar
m
s
tates,
t
h
er
e
ar
e
t
w
o
t
h
r
es
h
o
ld
v
alu
e
s
,
s
p
ec
i
f
icall
y
w
an
d
a
ar
e
u
tili
ze
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
20
88
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
4
,
A
u
g
u
s
t 2
0
1
9
:
2
6
5
9
-
2667
2662
(
i.e
.
w
a
)
.
R
ig
h
t
w
h
e
n
t
h
e
alar
m
t
h
r
es
h
o
ld
v
al
u
e
a
is
n
'
t
a
s
m
u
ch
as
t
h
e
e
s
ti
m
atio
n
o
f
t
h
e
c
o
n
tr
ast
b
et
w
ee
n
t
h
e
t
w
o
v
ar
iab
les,
t
h
e
s
tatio
n
ar
y
s
tate
i
s
tr
i
g
g
er
ed
.
Star
tin
g
n
o
w
a
n
d
i
n
to
t
h
e
f
o
r
eseea
b
le
f
u
tu
r
e,
to
ad
d
r
ess
th
e
d
ata
p
ar
titi
o
n
a
n
d
r
ec
o
g
n
ize
t
h
e
o
p
ti
m
ized
cl
u
s
ter
s
,
th
e
e
n
co
d
in
g
s
ch
e
m
e
i
s
ex
ec
u
ted
to
r
u
n
t
h
e
I
DE
alg
o
r
ith
m
.
Fo
r
th
e
e
f
f
ic
ie
n
t
a
n
d
r
o
b
u
s
t
o
p
ti
m
izat
io
n
p
r
o
ce
s
s
,
t
h
e
I
DE
al
g
o
r
ith
m
is
a
n
e
f
f
icien
t
a
n
d
w
ell
k
n
o
w
n
a
lg
o
r
it
h
m
w
h
ic
h
is
a
p
o
p
u
latio
n
b
ased
alg
o
r
ith
m
.
T
h
e
f
lo
atin
g
p
o
in
t
r
ep
r
esen
t
atio
n
s
ar
e
u
s
ed
in
th
is
al
g
o
r
ith
m
.
2
.
4
.
E
nco
din
g
s
c
he
m
e
Fo
r
th
e
c
lu
s
ter
in
g
p
r
o
b
le
m
s
,
th
e
ca
n
d
id
ate
s
o
l
u
tio
n
s
(
i
n
d
i
v
id
u
al
s
)
ar
e
d
escr
ib
ed
b
y
th
e
en
co
d
in
g
s
ch
e
m
e
o
f
[
1
4
]
is
a
d
o
p
ted
b
y
th
e
I
DE
-
s
tr
ea
m
al
g
o
r
ith
m
.
A
d
ata
s
tr
ea
m
X
is
a
s
ig
n
if
ican
t
ar
r
an
g
e
m
en
t
o
f
illu
s
tr
atio
n
s
an
d
it is
g
i
v
en
a
s
f
o
llo
w
:
l
j
j
i
x
X
e
i
t
l
d
i
x
l
i
x
l
i
x
l
i
x
t
l
i
X
1
.,
.
)}
(
,
,
,
2
,
1
,
{
)
(
(
2
)
w
h
er
e
i
I
n
co
m
in
g
o
b
j
ec
t
l
Nu
m
b
er
o
f
attr
ib
u
tes i
n
ea
ch
o
b
j
ec
t.
T
h
e
ab
o
v
e
eq
u
atio
n
is
p
o
ten
ti
all
y
u
n
b
o
u
n
d
e
d
).
(
N
E
ac
h
ca
s
e
is
d
ep
icted
b
y
a
n
-
d
i
m
e
n
s
io
n
al
attr
ib
u
tes
v
ec
to
r
n
l
l
i
x
i
X
1
.
A
t
t
h
at
p
o
i
n
t,
th
e
i
n
p
u
t
d
ata
p
o
in
t
s
(
D)
ar
e
p
ar
titi
o
n
ed
i
n
to
k
n
u
m
b
er
o
f
n
o
n
-
o
v
er
lap
p
in
g
cl
u
s
ter
s
}
,
,
2
,
1
{
k
C
C
C
C
s
u
c
h
t
h
at
s
atis
f
y
in
g
t
h
e
eq
u
atio
n
D
i
C
k
i
j
i
k
j
i
j
C
i
C
i
C
1
;
,
,
2
,
1
,
,
;
(
3
)
T
h
e
ab
o
v
e
ex
p
r
ess
io
n
clar
i
f
ie
d
,
th
at
t
h
e
o
b
j
ec
ts
i
n
t
h
e
s
i
m
i
lar
clu
s
ter
s
ar
e
l
ik
e
ea
ch
o
th
e
r
,
an
d
th
e
o
b
j
ec
ts
in
th
e
d
i
v
er
s
e
clu
s
ter
s
ar
e
d
is
p
ar
ate.
I
n
th
e
d
ata
s
et,
th
e
clo
s
en
e
s
s
a
n
d
d
is
p
ar
it
y
b
et
w
ee
n
t
h
e
o
b
j
ec
ts
ar
e
f
o
u
n
d
b
y
ev
al
u
ati
n
g
t
h
e
E
u
clid
ea
n
d
i
s
tan
ce
b
et
w
ee
n
t
h
e
p
o
in
ts
n
l
l
i
x
i
X
1
.
T
h
e
ab
o
v
e
p
ar
titi
o
n
ed
d
ata
C
is
ch
an
g
ed
in
to
th
e
in
te
g
er
s
tr
in
g
o
f
N
p
o
s
itio
n
s
b
y
u
s
i
n
g
en
co
d
ed
p
r
o
ce
s
s
.
Her
e,
th
e
p
o
s
itio
n
o
f
ev
er
y
s
tr
in
g
a
n
d
th
e
n
u
m
er
ical
o
r
d
er
s
f
o
r
th
e
o
b
j
ec
ts
o
f
t
h
e
d
at
asets
ar
e
al
m
o
s
t
s
i
m
ilar
.
T
h
e
ex
a
m
p
le
e
n
co
d
in
g
s
ch
e
m
e
o
f
t
h
e
d
ataset
is
s
h
o
wn
in
Fig
u
r
e
2
.
1
1
1
1
3
1
2
3
3
2
3
O
1
O
2
O
3
O
4
O
5
O
6
O
7
O
8
O
9
O
10
K
Fig
u
r
e
2
.
E
n
co
d
in
g
s
c
h
e
m
e
I
n
th
e
ab
o
v
e
ex
a
m
p
l
e,
th
e
d
ataset
co
n
s
i
s
ts
o
f
te
n
o
b
j
ec
ts
}
10
,
,
1
{
,
.
.
i
i
O
e
i
an
d
th
e
en
co
d
in
g
clu
s
ter
s
ar
e
th
r
ee
)
3
(
k
.
Her
e,
th
e
i
th
o
b
j
ec
t
o
f
th
e
d
ataset
is
s
to
r
ed
in
th
e
i
th
p
o
s
itio
n
an
d
t
h
e
k
-
v
al
u
e
is
s
to
r
ed
in
th
e
las
t p
o
s
itio
n
.
Her
e,
th
e
f
ea
t
u
r
e
v
ec
to
r
s
ar
e
u
s
ed
to
d
escr
ib
e
t
h
e
e
v
er
y
c
l
u
s
ter
a
n
d
t
h
e
n
u
m
b
er
o
f
o
b
j
ec
ts
N,
d
ata
o
b
j
ec
t’
s
l
in
ea
r
s
u
m
ca
lled
S
1
an
d
t
h
eir
s
q
u
ar
ed
s
u
m
ca
lled
S
2
an
d
ti
m
e
‘
t
’
o
f
th
e
m
o
s
t
r
e
ce
n
t
o
b
j
ec
ts
t
h
at
t
h
e
clu
s
ter
r
ec
eiv
ed
ar
e
t
h
e
f
o
u
r
g
en
u
i
n
e
q
u
an
t
ities
o
f
m
ea
s
u
r
e
m
en
t.
T
h
e
ce
n
tr
o
id
o
f
t
h
e
c
lu
s
ter
is
ev
a
lu
ated
b
y
u
s
i
n
g
th
e
in
i
tial
t
h
r
ee
co
m
p
o
n
en
ts
a
n
d
th
e
i
m
p
o
r
tan
ce
o
f
t
h
e
clu
s
ter
i
s
w
ei
g
h
ted
b
y
t
h
e
r
est
o
f
th
e
co
m
p
o
n
en
t.
A
t ti
m
e
‘
T
’
,
th
e
cl
u
s
ter
w
ei
g
h
t
‘
W
’
is
e
v
al
u
ated
as f
o
llo
w
s
:
v
t
T
e
W
(
4
)
w
h
er
e
W
is
r
ep
r
esen
ted
as
t
h
e
weig
h
t
o
f
t
h
e
cl
u
s
ter
,
v
is
r
ep
r
ese
n
ted
as
th
e
u
s
er
d
ef
i
n
ed
p
ar
a
m
eter
u
s
ed
to
co
n
tr
o
l
th
e
f
ad
in
g
f
ac
to
r
.
W
h
en
th
e
w
ei
g
h
tin
g
v
a
lu
e
o
f
t
h
e
clu
s
ter
is
l
e
s
s
th
a
n
0
.
1
(
i.e
.
W
<
0
.
1
)
,
th
at
clu
s
ter
i
s
r
e
m
o
v
ed
f
r
o
m
t
h
e
d
ata
p
ar
titi
o
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
n
imp
r
o
ve
d
d
iffer
en
tia
l e
vo
lu
tio
n
a
lg
o
r
ith
m
fo
r
d
a
t
a
s
tr
ea
m
clu
s
teri
n
g
(
B
h
a
s
ka
r
A
d
ep
u
)
2663
2
.
5
.
Va
lid
it
y
ind
ices o
f
t
he
clus
t
e
rs
T
h
e
v
alid
it
y
in
d
e
x
is
u
tili
ze
d
to
ev
alu
ate
th
e
n
u
m
b
er
o
f
clu
s
ter
s
.
T
h
e
ev
o
lu
tio
n
ar
y
alg
o
r
ith
m
o
p
tim
izes
th
e
k
-
v
al
u
e
b
ased
o
n
t
h
e
f
it
n
es
s
f
u
n
ct
io
n
a
n
d
it
is
esti
m
ated
b
y
u
s
i
n
g
t
h
e
v
alid
i
t
y
i
n
d
ex
.
T
h
e
w
ell
k
n
o
w
n
v
alid
it
y
i
n
d
ex
u
s
ed
in
o
u
r
m
eth
o
d
is
a
Si
m
p
li
f
ied
Sil
h
o
u
ette
(
SS
)
i
n
d
e
x
w
h
ic
h
is
u
s
ed
in
[
5
,
1
4
]
.
T
h
e
co
m
p
ac
t
n
es
s
an
d
t
h
e
p
ar
titi
o
n
o
f
th
e
clu
s
ter
s
ar
e
a
s
s
es
s
ed
b
y
u
t
ilizi
n
g
th
e
Sil
h
o
u
ett
e
w
id
t
h
.
T
h
e
g
i
v
e
n
ar
r
an
g
e
m
en
t o
f
th
e
d
ata
p
o
in
t
is
g
i
v
e
n
as tak
e
s
af
ter
,
l
j
i
i
i
i
l
i
x
x
x
x
X
,
3
2
1
,
,
,
,
(
5
)
w
h
er
e
in
t
h
e
ab
o
v
e
eq
u
atio
n
,
N
i
3
,
2
,
1
th
e
n
u
m
b
er
o
f
o
b
j
ec
ts
in
th
e
p
ar
titi
o
n
an
d
K
j
3
,
2
,
1
th
e
n
u
m
b
er
o
f
cl
u
s
ter
’
s
r
a
n
g
e
v
ar
iatio
n
s
.
T
h
e
ab
o
v
e
co
n
s
id
er
ed
d
ata
p
o
in
ts
l
i
X
co
m
p
r
is
es
th
e
clu
s
ter
a
C
.
I
n
a
C
,
th
e
alter
n
ate
o
b
j
ec
tiv
es
ar
e
d
i
s
tin
g
u
is
h
ed
i
n
v
ie
w
o
f
t
h
e
l
i
X
d
if
f
er
en
ce
ca
p
ac
it
y
a
n
d
it
is
m
e
an
t
b
y
)
(
l
i
X
a
.
I
n
clu
s
ter
a
C
,
th
e
n
o
r
m
al
d
i
v
er
g
en
ce
ca
p
ac
it
y
o
f
e
v
er
y
s
i
n
g
le
o
th
er
o
b
j
ec
t
is
in
d
icate
d
as
)
,
(
b
C
l
i
X
d
.
T
h
e
m
i
n
i
m
u
m
d
i
s
s
i
m
ilar
it
y
f
u
n
ctio
n
o
f
t
h
e
d
ata
p
o
in
ts
i
s
s
elec
ted
,
if
t
h
e
clu
s
ter
a
C
b
C
,
w
h
ich
i
s
g
iv
e
n
b
en
ea
t
h
,
)
1
(
,
)
(
j
c
e
n
t
r
o
i
d
l
i
X
d
i
s
l
i
X
a
(
6
)
)
(
,
m
i
n
)
(
i
j
c
en
t
r
o
i
d
l
i
X
dis
l
i
X
b
(
7
)
\
Af
ter
co
m
p
u
tin
g
t
h
e
d
is
s
i
m
i
la
r
it
y
f
u
n
ctio
n
,
t
h
e
s
il
h
o
u
e
tte
est
i
m
atio
n
)
(
l
i
X
S
is
g
iv
e
n
as
f
o
llo
w
s
,
)
(
),
(
m
a
x
)
(
)
(
)
(
l
i
X
b
l
i
X
a
l
i
X
a
l
i
X
b
l
i
X
S
(
8
)
T
h
e
s
ilh
o
u
ette
v
al
u
e
s
ar
e
o
n
ly
r
a
n
g
e
s
b
et
w
ee
n
th
e
i
n
ter
v
a
l
[
0
,
1
]
,
th
e
clo
s
est
clu
s
ter
v
alu
es
ar
e
ac
ce
s
s
ed
ac
co
r
d
in
g
to
th
e
eq
u
atio
n
(
1
1
)
.
A
t
th
at
p
o
in
t,
w
h
e
n
th
e
s
il
h
o
u
ette
v
al
u
e
is
n
ea
r
er
to
1
,
it
r
e
p
r
esen
ts
l
i
X
is
c
lu
s
ter
ed
p
r
ec
is
el
y
,
g
e
n
er
al
l
y
t
h
e
d
ata
p
o
in
t
s
ar
e
w
r
o
n
g
l
y
clu
s
ter
ed
.
T
h
e
o
v
er
all
s
ilh
o
u
et
te
in
d
e
x
o
f
t
h
e
p
o
r
tio
n
in
g
cl
u
s
ter
k
C
C
C
C
,
,
2
,
1
,
is
g
i
v
en
a
s
:
N
i
l
i
X
S
N
SS
1
)
(
1
(
9
)
I
n
th
e
ab
o
v
e
eq
u
atio
n
t
h
e
m
a
x
i
m
u
m
v
al
u
e
o
f
SS
)
m
a
x
(
SS
is
k
n
o
wn
as
th
e
f
it
n
e
s
s
f
u
n
ctio
n
o
f
th
e
o
b
j
ec
t
in
t
h
e
cl
u
s
ter
,
w
h
ic
h
i
s
u
til
ized
to
d
eter
m
i
n
e
t
h
e
q
u
a
lit
y
o
f
t
h
e
p
ar
titi
o
n
in
g
d
ata.
Her
e,
w
h
ic
h
cl
u
s
ter
h
as
th
e
m
ax
i
m
u
m
SS
v
al
u
e,
th
at
is
co
n
s
id
er
ed
as
t
h
e
b
est
clu
s
ter
in
g
an
d
th
e
n
th
e
ev
o
l
u
tio
n
ar
y
s
ea
r
ch
i
s
s
tar
ted
.
2
.
6
.
I
nn
o
v
a
t
iv
e
ev
o
lutio
na
r
y
s
ea
rc
h
I
n
o
u
r
m
eth
o
d
,
th
e
DE
alg
o
r
i
th
m
i
s
m
o
d
i
f
ied
b
ased
o
n
th
e
ad
j
u
s
t
m
en
t
o
f
p
o
p
u
latio
n
X
,
m
u
tatio
n
an
d
cr
o
s
s
o
v
er
CR
esti
m
atio
n
s
,
b
e
ca
u
s
e
th
e
b
est
o
p
ti
m
izatio
n
r
e
s
u
lt
s
ar
e
o
b
tai
n
ed
t
h
r
o
u
g
h
th
e
cr
o
s
s
o
v
er
a
n
d
th
e
m
u
tat
io
n
r
ate.
I
n
o
u
r
I
DE
i
m
p
le
m
e
n
tatio
n
,
in
it
iall
y
t
h
e
ca
n
d
id
ate
s
o
lu
tio
n
s
)}
(
,
,
,
,
{
,
2
1
t
x
x
x
x
l
d
i
l
i
l
i
l
i
ar
e
g
en
er
ated
f
r
o
m
t
h
e
i
n
itial
p
o
p
u
latio
n
X
an
d
it
i
s
s
h
o
w
n
i
n
s
ec
tio
n
2
.
4
as
th
e
en
co
d
i
n
g
s
c
h
e
m
e.
T
h
en
th
e
s
ilh
o
u
ette
i
n
d
ex
m
a
x
SS
is
ev
a
lu
ated
f
o
r
ea
ch
in
d
iv
id
u
al
to
k
n
o
w
t
h
e
f
it
n
es
s
an
d
t
h
is
e
s
ti
m
atio
n
i
s
s
h
o
w
n
i
n
s
ec
tio
n
2
.
5
.
A
f
ter
th
a
t,
t
h
e
o
t
h
er
o
b
j
ec
ts
)
(
),
(
t
x
t
x
l
j
l
i
an
d
)
(
t
x
l
p
ar
e
r
an
d
o
m
l
y
g
en
er
ated
f
r
o
m
t
h
e
i
n
it
ial
g
en
er
ated
ca
n
d
id
ate
s
o
lu
tio
n
.
T
h
en
th
e
d
i
f
f
er
e
n
ce
b
et
w
ee
n
)
(
t
x
l
i
an
d
)
(
t
x
l
j
ar
e
esti
m
ated
an
d
t
h
e
esti
m
ated
d
if
f
er
e
n
ce
v
al
u
es
ar
e
s
ca
led
b
y
s
ca
lar
S
,
it
i
s
r
ep
r
ese
n
ted
as
]
1
,
0
[
S
.
Her
e,
th
e
s
ca
led
v
al
u
e
o
f
th
e
t
w
o
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
20
88
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
4
,
A
u
g
u
s
t 2
0
1
9
:
2
6
5
9
-
2667
2664
r
an
d
o
m
v
ec
to
r
’
s
w
ei
g
h
ted
d
i
f
f
er
en
ce
v
al
u
es
))
(
,
)
(
,
(
t
l
u
j
X
t
u
l
i
X
O
is
ad
d
ed
to
th
e
t
h
ir
d
v
ec
to
r
)
(
t
x
l
p
to
g
en
er
ate
t
h
e
n
e
w
v
ec
to
r
)
1
(
t
O
l
i
,
it is
k
n
o
w
n
as t
h
e
m
u
tatio
n
a
n
d
it is
g
iv
e
n
as
f
o
llo
w
s
:
)
(
,
,
))
(
,
)
(
,
(
)
(
,
,
]
1
,
0
[
)
1
(
,
t
u
k
X
o
t
h
e
r
w
i
s
e
t
l
u
j
X
t
u
l
i
X
O
t
u
p
X
CR
u
r
a
n
d
if
t
u
k
O
(
1
0
)
T
h
en
th
e
o
t
h
er
p
r
ed
eter
m
in
ed
v
ec
to
r
is
m
i
x
ed
w
it
h
t
h
e
m
u
t
ated
v
ec
to
r
,
it
is
k
n
o
w
n
a
s
t
h
e
cr
o
s
s
o
v
er
r
ate
(
C
R
)
.
I
t
is
th
e
s
ca
lar
p
ar
a
m
eter
o
f
t
h
e
al
g
o
r
ith
m
an
d
i
n
b
et
w
ee
n
0
a
n
d
1
,
i.e
.
,
]
1
,
0
[
CR
.
A
t
last
,
ass
e
s
s
th
e
n
e
w
ca
n
d
id
ate
s
o
lu
t
io
n
with
t
h
e
S
S
m
ax
a
n
d
th
e
n
t
h
e
i
n
iti
al
ca
n
d
id
ate
s
o
l
u
tio
n
)
(
t
l
i
X
is
s
u
b
s
t
itu
ted
i
n
t
h
e
n
e
w
ca
n
d
id
ate
s
o
l
u
tio
n
)
1
(
t
l
i
X
.
Her
e,
if
t
h
e
n
e
w
l
y
g
e
n
er
ated
c
an
d
id
ate
s
o
l
u
tio
n
y
ield
s
t
h
e
m
a
x
i
m
u
m
o
b
j
ec
tiv
e
f
u
n
ct
io
n
,
th
e
s
e
s
o
lu
t
io
n
s
ar
e
co
n
s
id
er
ed
as
th
e
b
est
s
o
lu
tio
n
o
t
h
er
w
is
e
t
h
e
s
o
lu
t
io
n
is
r
etain
ed
i
n
th
e
p
o
p
u
latio
n
; it
is
d
ep
icted
in
t
h
e
eq
u
atio
n
b
elo
w
,
)
(
)
)
,
(
(
))
1
(
(
)
1
(
)
)
,
(
(
))
1
(
(
)
1
(
t
i
X
t
l
i
X
S
t
i
O
if
t
i
O
t
l
i
X
S
t
i
O
S
if
t
l
i
X
(
1
1
)
I
n
th
e
ab
o
v
e
eq
u
at
io
n
,
S(.
)
is
th
e
o
b
j
ec
tiv
e
eq
u
atio
n
.
T
o
en
h
a
n
ce
o
u
r
I
DE
ca
lcu
latio
n
,
w
e
h
av
e
en
h
a
n
ce
d
th
e
p
r
o
p
er
ties
o
f
I
D
E
in
d
if
f
er
e
n
t
w
a
y
s
.
T
o
s
ca
le
th
e
w
eig
h
ted
co
n
tr
ast
v
ec
to
r
,
)
(
,
)
(
,
t
l
u
j
X
t
u
l
i
X
,
th
e
s
ca
li
n
g
f
ac
to
r
S is
u
s
in
g
an
d
it e
x
ten
d
s
i
n
t
h
e
r
an
g
e
o
f
0
.
5
an
d
1
.
2
.
7
.
P
s
eudo
c
o
de
o
f
I
DE
S
t
re
a
m
a
lg
o
rit
h
m
Step
1
:
Gen
er
ate
th
e
ca
n
d
id
at
e
s
o
lu
t
io
n
s
f
r
o
m
t
h
e
in
i
tial p
o
p
u
latio
n
(
X)
i.e
.
th
e
in
d
i
v
id
u
al
p
o
in
t
s
l
n
X
l
X
l
X
,
2
,
1
ar
e
r
an
d
o
m
l
y
g
e
n
er
ated
f
r
o
m
th
e
d
ata
s
ets.
Step
2
:
C
alcu
a
lteSS
max
f
o
r
ea
ch
i
n
d
i
v
id
u
al.
[
Use e
q
.
9
]
Step
3
:
R
an
d
o
m
l
y
c
h
o
o
s
e
th
r
ee
o
b
j
ec
t
s
)
(
),
(
t
l
j
X
t
l
i
X
,
)
(
t
l
p
X
f
r
o
m
t
h
e
i
n
itiall
y
g
en
er
ated
c
an
d
id
ate
s
o
lu
tio
n
s
.
Step
4
:
E
s
ti
m
a
te
th
e
d
i
f
f
er
en
ce
b
et
w
e
en
an
y
t
w
o
o
b
j
ec
ts
an
d
s
ca
le
i
t
in
t
h
e
r
an
g
e
[
0
,
1
]
an
d
ad
d
t
h
is
v
al
u
e
to
th
ir
d
o
b
j
ec
t to
g
en
er
ate
a
n
e
w
o
b
j
ec
t.
[
Use e
q
.
1
0
]
Step
5
:
Pe
r
f
o
r
m
cr
o
s
s
o
v
er
b
y
m
i
x
i
n
g
th
is
m
u
tated
o
b
j
ec
t
r
esu
lted
f
r
o
m
s
tep
4
w
it
h
p
r
ed
ef
i
n
ed
o
b
ject
s
u
c
h
th
at
cr
o
s
s
o
v
er
r
ate
is
i
n
t
h
e
r
an
g
e
[
0
,
1
]
.
Step
6
:
Ass
es
s
t
h
e
n
e
w
l
y
g
en
er
ated
ca
n
d
id
ate
s
o
lu
tio
n
an
d
o
u
tp
u
t th
e
b
est r
esu
lt.
[
Use e
q
.
1
1
]
R
ep
ea
t step
2
to
s
tep
6
u
n
til s
t
o
p
p
in
g
cr
iter
ia
is
m
e
t.i.
e.
K<
=SS
m
ax
3.
RE
SU
L
T
S
A
ND
AN
AL
Y
SI
S
T
h
e
p
er
f
o
r
m
a
n
ce
o
f
t
h
e
I
DE
a
lg
o
r
ith
m
is
e
x
p
er
i
m
e
n
tal
l
y
ass
ess
ed
b
y
co
n
tr
ast
in
g
a
n
d
t
h
e
u
s
e
o
f
lat
e
cr
ea
ted
o
p
tim
iza
tio
n
al
g
o
r
ith
m
,
to
b
e
s
p
ec
i
f
ic
g
en
e
tic
a
lg
o
r
ith
m
[
2
5
]
.
T
h
e
p
r
im
ar
y
o
b
j
e
ctiv
e
o
f
th
i
s
ca
lcu
latio
n
is
,
to
o
p
tim
ize
t
h
e
ca
n
d
id
ate
s
o
lu
tio
n
to
p
r
o
d
u
ce
th
e
b
est
o
u
tco
m
e.
T
h
e
ca
n
d
id
ate
s
o
lu
tio
n
s
ar
e
p
ick
ed
r
elies
u
p
o
n
t
h
e
f
it
n
e
s
s
f
u
n
ctio
n
,
th
e
n
atu
r
e
o
f
t
h
e
ca
n
d
id
ate
s
o
l
u
tio
n
s
is
a
s
s
ess
ed
as
f
o
r
th
e
o
p
tim
izatio
n
is
s
u
e.
T
h
e
p
r
in
cip
l
e
f
av
o
r
ab
le
cir
cu
m
s
tan
ce
s
o
f
th
is
al
g
o
r
ith
m
ar
e,
it
ca
n
d
ea
l
w
it
h
a
f
e
w
co
m
p
eti
to
r
ar
r
an
g
e
m
e
n
t
s
all
t
h
e
w
h
i
le.
All
th
i
n
g
s
co
n
s
id
er
ed
,
in
n
u
m
er
o
u
s
h
a
n
d
y
ap
p
licatio
n
s
,
th
e
r
o
u
g
h
ass
u
r
an
ce
o
f
t
h
e
d
ata
s
et
is
u
n
th
i
n
k
ab
le.
I
n
o
u
r
tec
h
n
iq
u
e,
i
f
an
y
p
r
o
g
r
es
s
io
n
s
ar
e
h
ap
p
en
i
n
g
i
n
t
h
e
s
ea
s
o
n
o
f
DE
b
ased
clu
s
ter
in
g
,
w
h
ic
h
is
co
n
tr
asted
an
d
t
h
e
I
DE
al
g
o
r
it
h
m
f
o
r
u
tili
zi
n
g
a
s
i
m
ilar
o
b
j
e
ctiv
e
p
o
r
tr
a
y
al
a
n
d
th
e
s
il
h
o
u
e
tte
in
d
e
x
o
f
t
h
e
I
DE
.
3
.
1
.
Da
t
a
s
et
s
d
escript
io
n
W
e
u
s
ed
th
r
ee
d
atase
ts
a
n
d
th
e
y
ar
e
KDD
C
UP
’
9
9
[
4
]
,
[
5
]
,
f
o
r
est
co
v
er
t
y
p
e
[
5
]
,
[
2
6
]
,
elec
tr
ic
p
o
w
e
r
co
n
s
u
m
p
tio
n
d
ataset
[
2
6
]
.
W
e
co
m
p
ar
e
d
th
e
ac
c
u
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all
an
d
F
-
Me
a
s
u
r
e
s
o
n
th
ese
d
ataset
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
n
imp
r
o
ve
d
d
iffer
en
tia
l e
vo
lu
tio
n
a
lg
o
r
ith
m
fo
r
d
a
t
a
s
tr
ea
m
clu
s
teri
n
g
(
B
h
a
s
ka
r
A
d
ep
u
)
2665
(
i)
KDD
C
UP
’
9
9
Data
s
et:
KDD
C
UP
’
9
9
is
th
e
f
ir
s
t
r
ea
l
d
ata,
w
e
u
s
e
1
0
%
v
er
s
io
n
o
f
d
etec
tio
n
s
et,
an
d
th
i
s
v
er
s
io
n
is
m
o
r
e
ch
a
lle
n
g
i
n
g
a
n
d
m
o
r
e
co
n
ce
n
tr
ated
t
h
an
th
e
f
u
ll
v
er
s
io
n
.
3
4
co
n
ti
n
u
o
u
s
attr
ib
u
te
s
ar
e
u
s
ed
i
n
th
is
d
ata
s
et
w
h
ic
h
co
n
tai
n
s
2
3
d
if
f
er
en
t
cla
s
s
e
s
o
f
d
ata
an
d
4
9
4
,
0
2
0
co
n
n
ec
tio
n
r
ec
o
r
d
s
.
E
ac
h
o
b
j
ec
t
is
eit
h
er
a
n
o
r
m
al
co
n
n
ec
tio
n
o
r
an
in
tr
u
s
io
n
.
P
o
p
u
lar
m
in
i
m
u
m
-
m
a
x
i
m
u
m
p
r
o
ce
d
u
r
e
is
u
s
ed
f
o
r
n
o
r
m
alizi
n
g
th
e
d
ataset
attr
ib
u
te
v
al
u
es
w
i
t
h
in
[
0
,
1
]
b
ef
o
r
e
r
u
n
n
i
n
g
th
e
e
x
p
er
i
m
e
n
t.
(
ii)
Fo
r
est
C
o
v
er
T
y
p
e
Data
s
e
t:
Fo
r
est
co
v
er
t
y
p
e
i
s
t
h
e
s
ec
o
n
d
r
ea
l
d
ataset,
w
h
ic
h
co
n
tai
n
7
f
o
r
est
co
v
er
ty
p
e
o
f
5
8
1
,
0
1
2
g
eo
s
p
a
tial
d
escr
ip
tio
n
s
.
T
o
tally
5
4
a
ttrib
u
tes
ar
e
d
escr
ib
ed
in
o
b
j
e
ctso
u
t
o
f
w
h
ich
1
0
ar
e
q
u
an
titati
v
e
attr
ib
u
te
s
an
d
4
4
a
r
e
b
in
ar
y
attr
ib
u
te
s
.
T
h
ese
1
0
q
u
an
titati
v
e
attr
ib
u
tes
a
r
e
w
id
el
y
u
s
ed
in
m
an
y
ex
p
er
i
m
e
n
ts
as
r
ep
o
r
ted
in
th
e
liter
atu
r
e.
Si
m
ilar
to
t
h
e
KDD
C
UP
’
9
9
,
P
o
p
u
lar
m
i
n
i
m
u
m
-
m
ax
i
m
u
m
p
r
o
ce
d
u
r
e
is
u
s
ed
f
o
r
n
o
r
m
ali
zin
g
th
e
d
ataset
attr
ib
u
te
v
al
u
es
w
it
h
in
[
0
,
1
]
b
ef
o
r
e
r
u
n
n
i
n
g
t
h
e
ex
p
er
i
m
en
t.
B
y
ta
k
i
n
g
t
h
e
d
ata
in
p
u
t,
th
e
d
ata
s
et
is
co
n
v
er
ted
to
a
d
ata
s
tr
ea
m
.
(
iii)
E
lectr
ic
Po
w
er
C
o
n
s
u
m
p
t
io
n
Data
s
e
t
:
E
lectr
icit
y
d
ata
s
e
t
is
th
e
th
ir
d
r
ea
l
d
ataset,
w
h
ic
h
co
n
tai
n
s
2
,
0
7
5
,
2
5
9
n
u
m
b
er
o
f
i
n
s
ta
n
ce
s
an
d
9
d
if
f
er
e
n
t
attr
ib
u
t
es.
T
h
e
m
ai
n
c
h
ar
ac
ter
is
tic
o
f
t
h
is
d
ata
s
et
is
m
u
lti
v
ar
iate
an
d
ti
m
e
s
er
ies.
T
h
e
attr
ib
u
te
v
al
u
es
ar
e
n
o
r
m
alize
d
w
it
h
i
n
[
0
,
1
]
.
T
h
is
d
ata
s
etg
a
v
e
b
etter
r
es
u
lts
f
o
r
ea
ch
ac
tiv
a
tio
n
f
u
n
ctio
n
an
d
co
n
f
ig
u
r
atio
n
.
I
t p
er
f
o
r
m
ed
b
etter
th
an
t
h
e
o
th
er
d
atasets
.
3
.
2
.
E
v
a
lua
t
io
n
m
ea
s
ures
T
h
e
ac
cu
r
ate
k
-
v
al
u
es
ar
e
p
r
ed
icted
b
y
u
s
i
n
g
cl
u
s
ter
i
n
g
o
r
ien
ted
ap
p
r
o
ac
h
.
I
n
o
u
r
m
et
h
o
d
,
w
e
u
s
e
f
o
u
r
i
m
p
o
r
tan
t
ev
al
u
at
io
n
m
e
tr
ics
n
a
m
e
l
y
p
r
ec
is
io
n
,
ac
c
u
r
ac
y
,
r
ec
all
a
n
d
F
-
m
ea
s
u
r
e
f
o
r
o
u
r
ex
p
er
i
m
en
t
s
.
I
n
a
clu
s
ter
g
r
o
u
p
,
th
e
ac
c
u
r
ate
g
r
o
u
p
ed
d
ata
p
er
ce
n
tag
e
is
m
ea
s
u
r
ed
b
y
t
h
e
cl
u
s
ter
p
u
r
it
y
.
T
h
e
o
n
li
n
e
co
m
p
o
n
e
n
t
s
g
e
n
er
ate
th
e
m
ic
r
o
clu
s
ter
s
an
d
th
e
q
u
alit
y
o
f
th
i
s
clu
s
ter
is
ev
al
u
ate
d
b
ased
o
n
th
is
cl
u
s
ter
p
u
r
it
y
.
T
h
e
ac
c
u
r
ac
y
e
v
al
u
ati
o
n
is
u
ti
lized
to
d
etec
t
t
h
e
co
r
r
ec
tl
y
as
s
i
g
n
ed
cla
s
s
,
b
ased
o
n
t
h
is
ac
cu
r
ac
y
,
t
h
e
clu
s
ter
p
u
r
it
y
is
e
v
al
u
ated
.
Fo
r
s
tr
ea
m
clu
s
ter
in
g
,
t
h
e
p
r
ec
is
io
n
,
r
ec
all,
ac
cu
r
ac
y
an
d
F
-
m
ea
s
u
r
e
ar
e
u
s
ed
as
ev
alu
a
tio
n
m
ea
s
u
r
e
s
.
T
h
e
r
ati
o
b
et
w
ee
n
t
h
e
n
u
m
b
er
o
f
r
etr
iev
ed
r
elev
a
n
t
s
a
m
p
les
an
d
t
h
e
to
tal
n
u
m
b
er
o
f
g
r
o
u
p
ed
s
a
m
p
les
is
k
n
o
w
n
as
th
e
p
r
ec
is
io
n
,
w
h
ic
h
i
s
u
til
iz
ed
to
d
eter
m
in
e
t
h
e
ac
c
u
r
ate
clu
s
ter
i
n
g
r
es
u
lt
s
o
f
th
e
cl
u
s
ter
i
n
g
m
et
h
o
d
.
T
h
e
r
atio
b
et
w
ee
n
t
h
e
n
u
m
b
er
o
f
r
et
r
iev
ed
r
elev
an
t
s
a
m
p
les
a
n
d
th
e
to
tal
n
u
m
b
er
o
f
s
a
m
p
les
i
s
k
n
o
w
n
as
ac
c
u
r
ac
y
.
T
h
e
co
n
tr
ib
u
tio
n
o
f
th
e
p
r
ec
is
io
n
an
d
r
ec
all
is
k
n
o
wn
as
an
F
-
m
ea
s
u
r
e.
T
h
e
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all
an
d
F
-
m
ea
s
u
r
e
ar
e
g
i
v
en
a
s
f
o
llo
w
s
:
FN
FP
TP
TP
A
c
c
u
r
a
c
y
(
1
2
)
FP
TP
TP
(P
)
P
re
c
i
s
i
o
n
(
1
3
)
FN
TP
TP
(R)
Re
c
a
l
l
(
1
4
)
R
P
R
P.
2
m
ea
s
u
r
e
F
(
1
5
)
I
n
th
e
ab
o
v
e
eq
u
atio
n
s
,
T
P
is
th
e
tr
u
e
p
o
s
iti
v
e,
T
N
is
th
e
tr
u
e
n
e
g
ati
v
e,
FP
is
th
e
f
a
ls
e
p
o
s
itiv
e
a
n
d
FN is t
h
e
f
al
s
e
n
e
g
ati
v
e
s
a
m
p
l
es.
T
ab
le
1
s
h
o
w
s
th
e
r
es
u
lt
s
o
f
ev
al
u
atio
n
m
e
s
u
r
es c
o
m
p
u
te
d
o
n
th
r
ee
d
if
f
er
en
t
d
ata
s
ets.
T
h
is
s
h
o
w
s
th
at
o
u
r
p
r
o
p
o
s
ed
I
DE
Stre
a
m
alg
o
r
ith
m
p
er
f
o
r
m
s
b
etter
co
m
p
ar
ed
w
it
h
Ge
n
eti
c
A
l
g
o
r
ith
m
(
G
A
)
.
Fo
llo
w
in
g
g
r
ap
h
s
r
ep
r
esen
t
s
co
m
p
a
r
is
o
n
o
f
o
u
r
p
r
o
p
o
s
ed
I
DE
Stre
a
m
alg
o
r
it
h
m
w
it
h
g
en
et
ic
alg
o
r
it
h
m
i
n
f
o
u
r
ev
al
u
atio
n
m
ea
s
u
r
es o
n
t
h
r
ee
d
atas
ets
as s
h
o
w
n
i
n
Fi
g
u
r
e
3
.
T
ab
le
1
.
E
v
alu
atio
n
m
ea
s
u
r
es
o
n
th
r
ee
d
atasets
D
a
t
a
S
e
t
s
Ev
a
l
u
a
t
i
o
n
M
e
a
s
u
r
e
s
P
r
e
c
i
si
o
n
R
e
c
a
l
l
A
c
c
u
r
a
c
y
F
_
M
e
a
su
r
e
GA
P
r
o
p
o
se
d
I
D
E
GA
P
r
o
p
o
se
d
I
D
E
GA
P
r
o
p
o
se
d
I
D
E
GA
P
r
o
p
o
se
d
I
D
E
1
.
K
D
D
C
U
P
’
9
9
0
.
7
8
1
5
0
.
8
6
3
1
0
.
8
1
3
4
0
.
9
0
2
3
0
.
8
6
3
4
0
.
9
1
4
9
0
.
7
9
7
2
0
.
8
8
2
3
2
.
F
o
r
e
se
t
C
o
v
e
r
Ty
p
e
0
.
8
1
1
7
0
.
8
6
5
7
0
.
8
4
3
0
0
.
9
0
5
6
0
.
8
4
5
2
0
.
9
3
1
4
0
.
8
2
7
1
0
.
8
8
5
2
3
.
El
e
c
t
r
i
c
P
o
w
e
r
C
o
n
s
u
mp
t
i
o
n
0
.
8
1
7
4
0
.
8
8
0
2
0
.
8
4
7
5
0
.
9
0
1
2
0
.
8
8
1
9
0
.
9
2
2
5
0
.
8
3
2
1
0
.
8
9
0
6
A
v
e
r
a
g
e
v
a
l
u
e
0
.
8
0
3
5
0
.
8
6
9
7
0
.
8
3
4
6
0
.
9
0
3
0
0
.
8
6
3
5
0
.
9
2
2
9
0
.
8
1
8
8
0
.
8
8
6
0
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
20
88
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
4
,
A
u
g
u
s
t 2
0
1
9
:
2
6
5
9
-
2667
2666
P
r
o
p
o
s
ed
I
DE
Stre
am
A
l
g
o
r
ith
m
Gen
etic
A
l
g
o
r
ith
m
(
G
A
)
Fig
u
r
e
3
.
T
h
e
e
v
alu
atio
n
o
f
co
m
p
ar
i
s
o
n
b
et
w
ee
n
p
r
o
p
o
s
e
d
I
DE
Stre
a
m
al
g
o
r
ith
m
a
n
d
g
en
e
tic
alg
o
r
ith
m
4.
CO
NCLU
SI
O
N
I
n
t
h
is
p
ap
er
w
e
p
r
ese
n
ted
an
I
DE
S
tr
ea
m
a
lg
o
r
it
h
m
f
o
r
au
to
m
at
ic
cl
u
s
ter
i
n
g
o
f
s
tr
ea
m
d
ata.
P
r
im
ar
y
co
n
tr
ib
u
tio
n
s
o
f
t
h
i
s
p
ap
er
ar
e
to
d
etec
t th
e
o
p
ti
m
al
n
u
m
b
er
o
f
c
lu
s
ter
s
a
u
to
m
a
tic
all
y
f
o
r
all
d
ata
s
et
s
.
Ou
r
test
in
g
d
ata
s
e
ts
ar
e
ta
k
e
n
f
r
o
m
t
h
e
KDD
C
UP
’
9
9
,
f
o
r
est
co
v
er
t
y
p
e
s
,
an
d
elec
tr
ic
p
o
w
er
co
n
s
u
m
p
tio
n
d
atasets
.
O
u
r
p
r
o
p
o
s
ed
I
DE
alg
o
r
ith
m
is
m
o
r
e
e
f
f
ic
ien
t
t
h
an
t
h
e
e
x
is
t
in
g
w
e
ll
k
n
o
w
n
cl
u
s
ter
i
n
g
al
g
o
r
ith
m
s
s
u
c
h
a
s
Ge
n
etic
al
g
o
r
it
h
m
.
I
n
o
u
r
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
,
w
e
a
ls
o
ap
p
ly
a
n
e
n
tr
o
p
y
b
a
s
ed
tech
n
iq
u
e
f
o
r
d
etec
ti
n
g
th
e
co
n
ce
p
t
d
r
if
t
th
er
eb
y
u
p
d
atin
g
th
e
c
lu
s
ter
in
g
p
r
o
ce
s
s
.
Ou
r
ex
p
er
i
m
en
tal
r
es
u
lts
s
h
o
w
t
h
at
t
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
is
s
i
m
p
le,
p
r
ac
tical
an
d
i
m
p
ac
t
f
u
l.
T
h
e
p
er
f
o
r
m
a
n
ce
o
f
o
u
r
m
et
h
o
d
is
e
s
ti
m
ated
b
ased
o
n
t
h
e
ac
cu
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all
an
d
F
-
m
ea
s
u
r
e
v
al
u
es.
I
n
th
e
f
u
tu
r
e
w
o
r
k
,
w
e
co
u
ld
s
t
u
d
y
h
o
w
to
e
n
h
an
ce
t
h
e
s
y
s
te
m
to
i
m
p
r
o
v
e
t
h
e
p
r
ec
is
io
n
.
I
n
f
u
t
u
r
e
w
e
w
i
ll
i
m
p
l
e
m
en
t
t
h
is
I
DE
alg
o
r
it
h
m
f
o
r
d
if
f
er
e
n
t
t
y
p
es
o
f
attr
ib
u
tes i
n
th
e
cl
u
s
ter
in
g
d
o
m
ai
n
.
RE
F
E
R
E
NC
E
S
[1
]
B.
Zh
a
n
g
,
e
t
a
l
.
,
“
Da
ta
stre
a
m
c
lu
ste
rin
g
b
a
se
d
o
n
F
u
z
z
y
C
-
Me
a
n
a
lg
o
rit
h
m
a
n
d
e
n
tro
p
y
t
h
e
o
ry
,
”
S
ig
n
a
l
Pro
c
e
ss
in
g
,
p
p
.
1
1
1
-
1
1
6
,
2
0
1
5
.
[2
]
J.
Ba
rd
d
a
l,
e
t
a
l
.
,
“
S
NCS
trea
m
+
:
Ex
ten
d
in
g
a
h
ig
h
q
u
a
li
ty
tru
e
a
n
y
ti
m
e
d
a
ta
stre
a
m
c
lu
ste
rin
g
a
lg
o
rit
h
m
,
”
J
o
u
rn
a
l
of
In
fo
rm
a
t
io
n
S
y
ste
ms
,
v
o
l.
6
2
,
p
p
.
6
0
-
7
3
,
2
0
1
6
.
[3
]
M
.
Ho
ss
e
in
i,
e
t
a
l
.
,
“
A
n
e
n
se
m
b
le
o
f
c
lu
ste
r
-
b
a
se
d
c
las
sif
i
e
rs
f
o
r
se
m
i
-
su
p
e
rv
ise
d
c
las
si
f
ic
a
ti
o
n
o
f
n
o
n
-
sta
ti
o
n
a
ry
d
a
ta strea
m
s,
”
Kn
o
wled
g
e
a
n
d
In
f
o
rm
a
ti
o
n
S
y
ste
ms
,
v
o
l
/i
ss
u
e
:
46
(
3
)
,
p
p
.
5
6
7
-
5
9
7
,
2
0
1
5
.
[4
]
M.
Kh
a
li
li
a
n
,
e
t
a
l
.
,
“
Da
ta
stre
a
m
c
lu
ste
rin
g
b
y
d
iv
id
e
a
n
d
c
o
n
q
u
e
r
a
p
p
ro
a
c
h
b
a
se
d
o
n
v
e
c
to
r
m
o
d
e
l
,
”
J
o
u
rn
a
l
o
f
Bi
g
Da
ta
,
v
o
l
/i
ss
u
e
:
3
(
1
)
,
2
0
1
6
.
[5
]
J.
A.
S
il
v
a
,
e
t
a
l
.
,
“
A
n
e
v
o
lu
ti
o
n
a
ry
a
lg
o
rit
h
m
f
o
r
c
lu
ste
rin
g
d
a
ta
stre
a
m
s
w
it
h
a
v
a
riab
le
n
u
m
b
e
r
o
f
c
lu
ste
rs
,
”
Exp
e
rt S
y
ste
ms
wit
h
A
p
p
l
ica
ti
o
n
s,
v
o
l.
6
7
,
p
p
.
2
2
8
-
2
3
8
,
2
0
1
7
.
[6
]
E.
d
e
F
a
ria,
e
t
a
l
.
,
“
Ev
a
lu
a
ti
o
n
o
f
M
u
l
ti
c
las
s
No
v
e
lt
y
De
te
c
t
io
n
A
lg
o
rit
h
m
s
f
o
r
Da
ta
S
trea
m
s
,
’
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
K
n
o
wle
d
g
e
a
n
d
Da
t
a
E
n
g
i
n
e
e
rin
g
,
v
o
l
/i
ss
u
e
:
27
(
11
)
,
p
p
.
2
9
6
1
-
2
9
7
3
,
2
0
1
5
.
[7
]
C.
W
a
n
g
,
e
t
a
l
.
,
“
S
V
S
trea
m
:
A
S
u
p
p
o
rt
V
e
c
to
r
-
Ba
se
d
A
lg
o
rit
h
m
fo
r
Clu
ste
ri
n
g
Da
ta
S
trea
m
s,
”
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
K
n
o
wle
d
g
e
a
n
d
Da
t
a
E
n
g
i
n
e
e
rin
g
,
v
o
l
/i
ss
u
e
:
25
(
6
)
,
p
p
.
1
4
1
0
-
1
4
2
4
,
2
0
1
3
.
[8
]
D.
A
d
e
n
i
y
i,
e
t
a
l
.,
“
A
u
to
m
a
ted
w
e
b
u
sa
g
e
d
a
ta
m
in
in
g
a
n
d
re
c
o
m
m
e
n
d
a
ti
o
n
sy
ste
m
u
sin
g
K
-
N
e
a
re
st
Ne
ig
h
b
o
r
(KN
N) cla
ss
i
f
ica
ti
o
n
m
e
th
o
d
,
”
A
p
p
li
e
d
Co
m
p
u
ti
n
g
a
n
d
I
n
f
o
rm
a
ti
c
s
,
v
o
l
/i
ss
u
e
:
12
(
1
)
,
p
p
.
9
0
-
1
0
8
,
2
0
1
6
.
[9
]
J.
Zh
a
n
g
,
e
t
a
l
.,
“
Distri
b
u
ted
d
a
ta
stre
a
m
c
lu
ste
rin
g
a
lg
o
rit
h
m
b
a
se
d
o
n
a
ff
in
it
y
p
ro
p
a
g
a
ti
o
n
,
”
J
o
u
r
n
a
l
o
f
C
o
mp
u
ter
Ap
p
li
c
a
ti
o
n
s
,
v
o
l
/i
ss
u
e
:
33
(
9
)
,
p
p
.
2
4
7
7
-
2
4
8
1
,
2
0
1
3
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
n
imp
r
o
ve
d
d
iffer
en
tia
l e
vo
lu
tio
n
a
lg
o
r
ith
m
fo
r
d
a
t
a
s
tr
ea
m
clu
s
teri
n
g
(
B
h
a
s
ka
r
A
d
ep
u
)
2667
[1
0
]
K.
G
u
o
a
n
d
Q.
Zh
a
n
g
,
“
F
a
st
Clu
ste
rin
g
-
Ba
se
d
A
n
o
n
y
m
iza
ti
o
n
A
l
g
o
rit
h
m
f
o
r
Da
ta
S
trea
ms
,
”
J
o
u
rn
a
l
o
f
S
o
ft
wa
re
En
g
i
n
e
e
rin
g
,
v
o
l
/i
ss
u
e
:
24
(
8
)
,
p
p
.
1
8
5
2
-
1
8
6
7
,
2
0
1
4
.
[1
1
]
R.
Hy
d
e
,
e
t
a
l
.
,
“
F
u
ll
y
o
n
li
n
e
c
lu
ste
rin
g
o
f
e
v
o
lv
in
g
d
a
ta
stre
a
m
s
in
to
a
rb
it
ra
ril
y
sh
a
p
e
d
c
l
u
ste
rs,
”
J
o
u
rn
a
l
o
f
In
fo
rm
a
t
io
n
S
c
ien
c
e
s
,
v
o
l
.
3
8
2
-
3
8
3
,
p
p
.
9
6
-
1
1
4
,
2
0
1
7
.
[1
2
]
N.
Ch
a
tu
rv
e
d
i
,
e
t
a
l
.
,
“
A
n
Im
p
r
o
v
e
m
e
n
t
in
K
-
m
e
a
n
Clu
ste
rin
g
A
l
g
o
rit
h
m
Us
in
g
Be
tt
e
r
T
i
m
e
a
n
d
A
c
c
u
ra
cy
,
”
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
Pr
o
g
r
a
mm
in
g
L
a
n
g
u
a
g
e
s a
n
d
A
p
p
l
ica
ti
o
n
s
,
v
o
l
/i
ss
u
e
:
3
(
4
)
,
p
p
.
1
3
-
1
9
,
2
0
1
3
.
[1
3
]
M.
Na
ld
i
a
n
d
R.
Ca
m
p
e
ll
o
,
“
Ev
o
lu
ti
o
n
a
ry
k
-
m
e
a
n
s
f
o
r
d
istri
b
u
ted
d
a
ta
se
ts,
”
Ne
u
r
o
Co
m
p
u
ti
n
g
,
v
o
l.
1
2
7
,
p
p
.
3
0
-
4
2
,
2
0
1
4
.
[1
4
]
V.
S.
A
lv
e
s,
e
t
a
l
.
,
“
T
o
wa
rd
s
a
F
a
st
Ev
o
l
u
t
i
o
n
a
ry
A
l
g
o
rit
h
m
fo
r
Cl
u
ste
rin
g
,”
IEE
E
c
o
n
g
re
ss
o
n
e
v
o
lu
ti
o
n
a
ry
c
o
mp
u
t
a
ti
o
n
,
IEE
E
P
re
ss
,
p
p
.
1
7
7
6
-
1
7
8
3
,
2
0
0
6
.
[1
5
]
Y.
P
in
g
,
e
t
a
l
.
,
“
F
a
st
a
n
d
sc
a
l
a
b
le
su
p
p
o
rt
v
e
c
to
r
c
l
u
ste
rin
g
f
o
r
larg
e
-
sc
a
le
d
a
ta
a
n
a
l
y
sis,
”
K
n
o
wled
g
e
a
n
d
In
fo
rm
a
t
io
n
S
y
ste
ms
,
v
o
l
/is
s
u
e
:
43
(
2
)
,
p
p
.
2
8
1
-
3
1
0
,
2
0
1
4
.
[1
6
]
E.
d
e
F
a
ria,
e
t
a
l
.
,
“
M
IN
A
S
:
m
u
lt
icla
ss
lea
rn
in
g
a
l
g
o
rit
h
m
f
o
r
n
o
v
e
lt
y
d
e
te
c
ti
o
n
in
d
a
ta
stre
a
m
s,
”
Da
ta
M
in
in
g
a
n
d
Kn
o
wled
g
e
Disc
o
v
e
ry
,
v
o
l
/i
ss
u
e
:
30
(
3
)
,
p
p
.
6
4
0
-
6
8
0
,
2
0
1
5
.
[1
7
]
P
.
D.
a
n
d
A
.
Dix
it
,
“
M
u
lt
i
No
v
e
l
Clas
s
Cla
ss
i
f
ica
ti
o
n
o
f
F
e
a
tu
re
Ev
o
lv
in
g
Da
ta
S
trea
m
s
w
it
h
J4
8
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Co
m
p
u
ter
A
p
p
l
ica
ti
o
n
s
,
v
o
l
/i
ss
u
e
:
1
2
4
(
11
)
,
p
p
.
3
1
-
3
6
,
2
0
1
5
.
[1
8
]
Y.
Ha
n
,
“
Im
p
ro
v
e
d
BIRCH
Clu
ste
rin
g
A
lg
o
rit
h
m
a
n
d
Hu
m
a
n
Re
so
u
rc
e
M
a
n
a
g
e
m
e
n
t
Ef
f
ici
e
n
c
y
:
A
n
Or
g
a
n
iza
ti
o
n
a
l
L
e
a
rn
in
g
P
e
rsp
e
c
ti
v
e
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
S
e
c
u
rity
a
n
d
Its
A
p
p
li
c
a
ti
o
n
s
,
v
o
l
/i
ss
u
e
:
10
(
8
)
,
p
p
.
3
8
5
-
3
9
4
,
2
0
1
6
.
[1
9
]
Y.
L
iu
,
“
F
u
z
z
y
-
Clu
ste
rin
g
W
e
b
b
a
se
d
o
n
M
in
i
n
g
,
”
J
o
u
r
n
a
l
o
f
M
u
lt
ime
d
ia
,
v
o
l
/i
ss
u
e
:
9
(
1
)
,
2
0
1
4
.
[2
0
]
H.
L
e
e
,
e
t
a
l
.
,
“
A
M
a
p
Re
d
u
c
e
-
b
a
se
d
k
NN
Jo
in
Qu
e
ry
P
ro
c
e
ss
in
g
A
lg
o
r
it
h
m
f
o
r
A
n
a
l
y
z
in
g
Larg
e
-
sc
a
le
D
a
ta
,
”
J
o
u
rn
a
l
o
f
KIIS
E
,
v
o
l
/i
ss
u
e
:
42
(
4
)
,
p
p
.
5
0
4
-
5
1
1
,
2
0
1
5
.
[2
1
]
Z.
M
il
ler,
e
t
a
l
.
,
“
T
w
it
ter
sp
a
m
m
e
r
d
e
tec
ti
o
n
u
sin
g
d
a
ta
stre
a
m
c
lu
ste
rin
g
,
”
In
fo
rm
a
ti
o
n
S
c
i
e
n
c
e
s,
v
o
l.
2
6
0
,
p
p
.
6
4
-
7
3
,
2
0
1
4
.
[2
2
]
T
.
V
e
lm
u
ru
g
a
n
,
“
P
e
rf
o
rm
a
n
c
e
b
a
se
d
a
n
a
ly
sis
b
e
t
w
e
e
n
k
-
M
e
a
n
s
a
n
d
F
u
z
z
y
C
-
M
e
a
n
s
c
lu
ste
rin
g
a
lg
o
rit
h
m
s
f
o
r
c
o
n
n
e
c
ti
o
n
o
rien
ted
tele
c
o
m
m
u
n
ica
ti
o
n
d
a
ta
,
”
A
p
p
li
e
d
S
o
f
t
Co
m
p
u
t
in
g
,
v
o
l.
1
9
,
p
p
.
1
3
4
-
1
4
6
,
2
0
1
4
.
[2
3
]
R.
F
o
k
,
e
t
a
l
.
,
“
M
in
i
n
g
Ev
o
lv
in
g
Da
ta
S
trea
m
s
w
it
h
P
a
rti
c
le
F
il
ter
s,
”
Co
mp
u
t
a
ti
o
n
a
l
In
telli
g
e
n
c
e
,
v
o
l
/i
ss
u
e
:
33
(
2
)
,
p
p
.
1
4
7
-
1
8
0
,
2
0
1
5
.
[2
4
]
V
.
B
h
a
tn
a
g
a
r,
e
t
a
l
.
,
“
Clu
ste
ri
n
g
d
a
ta
stre
a
m
s
u
sin
g
g
rid
-
b
a
se
d
s
y
n
o
p
sis,
”
Kn
o
wled
g
e
a
n
d
In
fo
r
ma
ti
o
n
S
y
ste
ms
,
v
o
l
/i
ss
u
e
:
41
(
1
)
,
p
p
.
1
2
7
-
1
5
2
,
2
0
1
3
.
[2
5
]
X
.
Y
u
a
n
,
e
t
a
l
.
,
“
A
G
e
n
e
ti
c
A
lg
o
rit
h
m
-
Ba
se
d
,
Dy
n
a
m
i
c
Clu
ste
rin
g
M
e
t
h
o
d
t
o
w
a
rd
s
Im
p
ro
v
e
d
WS
N
L
o
n
g
e
v
it
y
,
”
J
o
u
rn
a
l
o
f
Ne
two
rk
a
n
d
S
y
ste
ms
M
a
n
a
g
e
me
n
t
,
v
o
l
/i
ss
u
e
:
25
(
1
)
,
p
p
.
2
1
-
4
6
,
2
0
1
6
.
[2
6
]
D.
M
a
rró
n
,
e
t
a
l
.
,
“
Da
ta
stre
a
m
c
las
si
f
ica
ti
o
n
u
sin
g
ra
n
d
o
m
f
e
a
t
u
re
f
u
n
c
ti
o
n
s
a
n
d
n
o
v
e
l
m
e
th
o
d
c
o
m
b
in
a
ti
o
n
s
,
”
J
o
u
rn
a
l
o
f
S
y
ste
ms
a
n
d
S
o
f
twa
re
,
v
o
l.
1
2
7
,
p
p
.
1
9
5
-
2
0
4
,
2
0
1
7
.
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
B
h
a
s
k
a
r
Adepu
is
a
n
A
ss
o
c
iate
P
r
o
f
e
ss
o
r
a
t
Ka
k
a
ti
y
a
In
stit
u
te
o
f
T
e
c
h
n
o
lo
g
y
&
S
c
ien
c
e
,
W
a
ra
n
g
a
l
a
n
d
A
ff
il
iate
d
to
Ka
k
a
ti
y
a
Un
iv
e
rsit
y
,
In
d
ia.
He
is
p
u
rsu
in
g
h
is
P
h
.
D.
f
ro
m
Ja
wa
h
a
rlal
Ne
h
ru
T
e
c
h
n
o
l
o
g
ica
l
Un
iv
e
rsity
(J
NTU
),
Hy
d
e
ra
b
a
d
.
He
re
c
e
i
v
e
d
M
.
T
e
c
h
.
(CS
E)
f
ro
m
JN
TU
-
H
y
d
e
ra
b
a
d
in
2
0
1
0
.
His
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
Da
ta
M
in
i
n
g
,
Im
a
g
e
P
ro
c
e
ss
in
g
a
n
d
A
rti
f
icia
l
In
telli
g
e
n
c
e
.
He
d
e
li
v
e
re
d
g
u
e
st l
e
c
tu
re
s in
th
e
f
ield
o
f
d
a
ta
m
in
in
g
a
n
d
a
rti
f
icia
l
in
t
e
ll
ig
e
n
c
e
a
t
v
a
rio
u
s p
latf
o
r
m
s.
He
is a M
e
m
b
e
r
o
f
IEE
E
a
n
d
a
m
e
m
b
e
r
o
f
IS
T
E.
J
a
y
a
d
e
v
G
y
a
n
i
is
a
n
A
s
sista
n
t
P
r
o
f
e
ss
o
r
a
t
th
e
Co
l
leg
e
o
f
Co
m
p
u
ter
a
n
d
In
f
o
rm
a
ti
o
n
S
c
ien
c
e
s,
M
a
jm
a
a
h
Un
iv
e
rsit
y
,
A
l
M
a
j
m
a
a
h
1
5
3
4
1
,
S
a
u
d
i
A
ra
b
ia.
He
re
c
e
iv
e
d
h
is
P
h
.
D.
f
ro
m
t
h
e
Un
iv
e
rsit
y
o
f
H
y
d
e
ra
b
a
d
in
2
0
0
9
.
He
h
a
s
p
u
b
li
s
h
e
d
m
o
re
th
a
n
8
0
re
f
e
re
e
d
jo
u
r
n
a
ls
a
n
d
c
o
n
f
e
re
n
c
e
a
rti
c
les
in
th
e
a
re
a
o
f
so
f
t
w
a
re
e
n
g
in
e
e
ri
n
g
,
d
a
ta
m
in
in
g
,
w
e
b
in
f
o
rm
a
t
io
n
sy
ste
m
s
a
n
d
d
ig
it
a
l
im
a
g
e
p
ro
c
e
ss
in
g
.
Dr.Gy
a
n
i
wa
s
p
ro
g
ra
m
Co
-
c
h
a
ir
a
n
d
V
ice
Ch
a
ir
f
o
r
f
e
w
in
tern
a
ti
o
n
a
l
c
o
n
f
e
re
n
c
e
s.
He
i
s
a
M
e
m
b
e
r
o
f
IEE
E
Co
m
p
u
ter S
o
c
iety
a
n
d
a
M
e
m
b
e
r
o
f
A
CM
.
Na
r
si
m
h
a
G
u
g
u
lo
t
u
is
a
P
r
o
f
e
ss
o
r
a
n
d
He
a
d
a
t
JN
T
UH
Co
ll
e
g
e
o
f
En
g
in
e
e
rin
g
,
S
u
lt
a
n
p
u
r
,
Ja
w
a
h
a
rlal
Ne
h
ru
T
e
c
h
n
o
lo
g
ica
l
Un
iv
e
rsity
,
H
y
d
e
r
a
b
a
d
,
In
d
ia.
He
re
c
e
iv
e
d
h
is
P
h
.
D.
f
ro
m
th
e
Os
m
a
n
ia
Un
iv
e
r
sit
y
,
H
y
d
e
r
a
b
a
d
in
2
0
0
9
.
He
h
a
s
p
u
b
li
sh
e
d
m
o
re
th
a
n
6
0
re
f
e
r
e
e
d
jo
u
rn
a
ls
a
n
d
c
o
n
f
e
re
n
c
e
a
rti
c
les
in
th
e
a
re
a
o
f
Da
ta
m
in
in
g
,
M
o
b
il
e
C
o
m
p
u
ti
n
g
,
Co
m
p
u
ter
Ne
tw
o
rk
s,
Clo
u
d
Co
m
p
u
ti
n
g
a
n
d
B
ig
d
a
ta
a
n
a
ly
ti
c
s.
Dr.Na
rsi
m
h
a
is
c
a
terin
g
v
a
ri
o
u
s
a
d
m
in
istrativ
e
a
n
d
a
c
a
d
e
m
ic
re
sp
o
n
si
b
il
it
ies
a
t
v
a
rio
u
s
c
a
p
a
c
it
ies
.
He
is
a
lso
re
v
ie
we
r
f
o
r
f
e
w
i
n
tern
a
ti
o
n
a
l
c
o
n
f
e
re
n
c
e
s.
He
is
a
M
e
m
b
e
r
o
f
IEE
E
Co
m
p
u
ter S
o
c
iety
a
n
d
a
M
e
m
b
e
r
o
f
IS
T
E.
Evaluation Warning : The document was created with Spire.PDF for Python.