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.
8
,
No
.
4
,
A
u
g
u
s
t
201
8
,
p
p
.
2
3
5
1
~
2
3
5
7
I
SS
N:
2088
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
8
i
4
.
p
p
2
3
5
1
-
2357
2351
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
Ex
tensiv
e Analy
sis
on G
en
eratio
n a
nd Cons
ensus
Me
cha
nis
m
s
o
f
Cluste
ring
Ens
e
m
bl
e
:
A Su
rv
ey
Y
.
L
ee
la
Sa
nd
hy
a
Ra
ni
1
,
V
.
Su
cha
rit
a
2
,
K
.
V
.
V
.
Sa
t
y
a
na
ra
y
a
na
3
1
De
p
a
rtme
n
t
o
f
C
SE
,
S
c
h
o
lar,
K
o
n
e
ru
L
a
k
sh
m
a
iah
Ed
u
c
a
ti
o
n
F
o
u
n
d
a
ti
o
n
(KL
U),
In
d
ia
2
De
p
a
rtme
n
t
o
f
C
SE
,
NEC,
G
u
d
u
r
,
In
d
ia
3
De
p
a
rtme
n
t
o
f
C
SE
,
P
r
o
f
e
ss
o
r,
Ko
n
e
ru
L
a
k
sh
m
a
iah
Ed
u
c
a
ti
o
n
F
o
u
n
d
a
ti
o
n
(KL
U),
In
d
ia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Oct
24
,
2
0
1
7
R
ev
i
s
ed
J
an
7
,
2
0
1
8
A
cc
ep
ted
J
an
1
5
,
2
0
1
8
Da
ta
a
n
a
l
y
sis
p
la
y
s
a
p
ro
m
in
e
n
t
ro
le
in
in
ter
p
re
ti
n
g
v
a
rio
u
s
p
h
e
n
o
m
e
n
a
.
Da
ta
m
in
in
g
is
th
e
p
ro
c
e
ss
to
h
y
p
o
th
e
siz
e
u
se
f
u
l
k
n
o
w
l
e
d
g
e
f
ro
m
th
e
e
x
ten
siv
e
d
a
ta.
Ba
se
d
u
p
o
n
t
h
e
c
las
sic
a
l
sta
ti
stica
l
p
ro
to
ty
p
e
s
th
e
d
a
t
a
c
a
n
b
e
e
x
p
lo
it
e
d
b
e
y
o
n
d
th
e
sto
ra
g
e
a
n
d
m
a
n
a
g
e
m
e
n
t
o
f
th
e
d
a
ta.
Clu
ste
r
a
n
a
ly
sis
a
p
rim
a
r
y
in
v
e
stig
a
ti
o
n
w
it
h
li
tt
le
o
r
n
o
p
rio
r
k
n
o
w
led
g
e
,
c
o
n
sists
o
f
r
e
se
a
rc
h
a
n
d
d
e
v
e
lo
p
m
e
n
t
a
c
ro
ss
a
w
id
e
v
a
ri
e
t
y
o
f
c
o
m
m
u
n
it
ies
.
Clu
ste
r
e
n
se
m
b
les
a
re
m
e
lan
g
e
o
f
in
d
iv
i
d
u
a
l
so
lu
t
io
n
s
o
b
tain
e
d
f
ro
m
d
i
ff
e
re
n
t
c
lu
s
terin
g
s
to
p
ro
d
u
c
e
f
in
a
l
q
u
a
li
ty
c
lu
ste
rin
g
w
h
ich
is
re
q
u
ired
i
n
w
id
e
r
a
p
p
li
c
a
ti
o
n
s.
T
h
e
m
e
th
o
d
a
rise
s
in
th
e
p
e
rsp
e
c
ti
v
e
o
f
in
c
re
a
sin
g
ro
b
u
st
n
e
ss
,
sc
a
lab
il
it
y
a
n
d
a
c
c
u
ra
c
y
.
T
h
is
p
a
p
e
r
g
iv
e
s
a
b
rief
o
v
e
rv
i
e
w
o
f
th
e
g
e
n
e
ra
ti
o
n
m
e
th
o
d
s
a
n
d
c
o
n
se
n
su
s
f
u
n
c
ti
o
n
s
in
c
l
u
d
e
d
in
c
lu
ste
r
e
n
se
m
b
le.
T
h
e
su
rv
e
y
is
to
a
n
a
ly
z
e
th
e
v
a
rio
u
s tec
h
n
iq
u
e
s an
d
c
lu
ste
r
e
n
se
m
b
le
m
e
th
o
d
s.
K
ey
w
o
r
d
:
C
lu
s
ter
i
n
g
C
lu
s
ter
i
n
g
e
n
s
e
m
b
le
C
o
n
s
e
n
s
u
s
m
et
h
o
d
Gen
er
ati
o
n
m
et
h
o
d
Un
s
u
p
er
v
i
s
ed
class
i
f
icatio
n
Co
p
y
rig
h
t
©
2
0
1
8
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
:
Y
.
L
ee
la
Sa
n
d
h
y
a
R
a
n
i,
Dep
ar
te
m
en
t o
f
C
o
m
p
u
ter
Sci
en
ce
an
d
E
n
g
i
n
ee
r
in
g
,
Sir
C
R
R
ed
d
y
C
o
lle
g
e
o
f
E
n
g
in
ee
r
in
g
,
E
lu
r
u
,
W
G
Dt,
An
d
h
r
a
P
r
r
ad
esh
,
I
n
d
ia.
E
m
ail:
l
s
r
an
i
y
ala
m
ar
th
i1
8
@
g
m
ail.
co
m
1.
I
NT
RO
D
UCT
I
O
N
Data
m
in
i
n
g
r
e
f
er
s
to
d
r
ed
g
i
n
g
o
f
k
n
o
w
led
g
e
f
r
o
m
ex
ten
s
iv
e
d
ata.
Data
m
i
n
i
n
g
d
ea
ls
w
it
h
lar
g
e
a
m
o
u
n
t
o
f
d
ata
s
ets.
C
l
u
s
ter
i
n
g
is
th
e
m
aj
o
r
f
u
n
ctio
n
al
it
y
i
n
d
ata
m
i
n
in
g
.
C
l
u
s
ter
i
n
g
is
a
w
a
y
o
f
o
r
g
an
izi
n
g
th
e
d
ata
i
n
to
s
i
m
ilar
g
r
o
u
p
s
t
h
at
h
av
e
s
a
m
e
f
ea
tu
r
e
s
an
d
i
t
d
o
es
n
o
t
o
cc
u
r
i
n
o
th
er
g
r
o
u
p
s
.
B
ec
au
s
e
o
f
th
i
s
p
r
o
p
er
ty
o
f
cl
u
s
ter
i
n
g
f
o
r
lar
g
e
d
ata
s
et
s
,
s
ca
lab
le
al
g
o
r
it
h
m
s
ar
e
r
eq
u
ir
ed
.
B
u
t
f
o
r
m
o
s
t
ap
p
licatio
n
s
o
f
clu
s
ter
i
n
g
t
h
e
y
ar
e
n
o
t
s
ca
led
w
ell
d
u
e
to
s
m
al
l
d
ata
s
et
s
.
E
ac
h
al
g
o
r
ith
m
h
as
i
ts
p
r
o
s
an
d
co
n
s
,
f
o
r
d
if
f
er
en
t
alg
o
r
ith
m
s
w
it
h
s
a
m
e
d
ata
s
e
t
o
r
/an
d
d
if
f
er
en
t
d
ata
s
et
s
w
i
t
h
s
a
m
e
alg
o
r
it
h
m
p
r
o
d
u
ce
s
d
i
s
tin
c
t
s
o
l
u
tio
n
s
.
I
t
i
s
to
u
g
h
to
k
n
o
w
w
h
ic
h
al
g
o
r
ith
m
is
s
u
itab
le
f
o
r
g
i
v
en
d
ata
s
et
.
C
lu
s
ter
i
n
g
p
la
y
s
a
m
aj
o
r
r
o
le
in
d
ata
m
i
n
i
n
g
,
m
ac
h
i
n
e
lea
r
n
in
g
,
b
io
in
f
o
r
m
at
ics
i
m
a
g
e
p
r
o
ce
s
s
in
g
,
in
f
o
r
m
atio
n
r
etr
ie
v
al,
m
ar
k
et
s
eg
m
e
n
tatio
n
,
b
i
g
d
ata
an
al
y
ti
cs
an
d
m
an
y
m
o
r
e
ar
ea
s
.
O
n
e
o
f
i
m
p
o
r
tan
t
tas
k
s
in
ca
n
ce
r
cla
s
s
i
f
icatio
n
ca
ll
class
d
i
s
co
v
er
y
b
y
m
icr
o
ar
r
ay
is
p
r
ev
io
u
s
l
y
d
o
n
e
u
s
in
g
s
in
g
le
c
lu
s
ter
in
g
alg
o
r
ith
m
s
.
U
s
in
g
clu
s
ter
i
n
g
tech
n
iq
u
es
w
e
ca
n
id
en
ti
f
y
t
h
e
co
-
lo
ca
tio
n
p
atter
n
s
th
at
a
r
e
u
s
u
all
y
ar
is
e
in
s
p
atial
d
ata
b
ases
u
s
i
n
g
s
o
m
e
d
ata
m
i
n
i
n
g
alg
o
r
ith
m
s
[
1
]
.
Gr
o
u
p
in
g
o
f
u
n
s
t
r
u
ct
u
r
ed
d
ata
b
ased
o
n
its
co
n
ten
t
is
d
o
n
e
b
y
d
o
cu
m
en
t
cl
u
s
ter
i
n
g
w
h
ic
h
i
s
o
n
e
o
f
th
e
m
o
s
t
p
o
p
u
lar
m
ac
h
i
n
e
lear
n
i
n
g
tech
n
i
q
u
es
a
n
d
it
f
u
r
th
e
r
an
al
y
s
e
t
h
e
d
ata
to
u
n
d
er
s
ta
n
d
p
atter
n
s
i
n
it.
Sep
ar
ati
n
g
o
f
p
ix
els
i
n
to
cl
u
s
ter
s
is
d
o
n
e
b
y
cl
u
s
ter
i
n
g
b
ased
i
m
a
g
e
s
eg
m
e
n
t
ap
p
r
o
ac
h
in
im
ag
e
p
r
o
ce
s
s
i
n
g
.
No
w
d
a
y
s
it
is
v
er
y
d
if
f
ic
u
lt
to
s
ea
r
ch
in
i
n
ter
n
et
as
t
h
er
e
ar
e
m
an
y
d
o
cu
m
e
n
ts
av
ai
lab
le
i
n
t
h
e
i
n
ter
n
et.
Sear
ch
i
n
g
ca
n
b
e
d
o
n
e
e
f
f
ec
tiv
e
l
y
u
s
in
g
s
o
m
e
k
e
y
w
o
r
d
s
b
y
clu
s
ter
i
n
g
al
g
o
r
ith
m
s
.
T
ex
t
b
ased
clu
s
ter
in
g
p
la
y
s
a
m
aj
o
r
r
o
le
in
b
r
o
w
s
i
n
g
an
d
n
av
i
g
atio
n
p
r
o
ce
s
s
[
2
]
.
C
lu
s
ter
i
n
g
p
la
y
a
m
aj
o
r
r
o
le
in
s
o
cial
n
et
w
o
r
k
s
al
s
o
.
I
t
is
u
s
ed
to
an
al
y
ze
p
s
y
ch
o
lo
g
y
o
f
h
u
m
a
n
s
an
d
t
h
eir
r
elatio
n
s
h
ip
s
[
3
]
.
C
lu
s
ter
in
g
[
4
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2088
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
4
,
A
u
g
u
s
t
201
8
:
2
3
5
1
–
2
3
5
7
2352
So
m
e
o
f
t
h
e
d
ata
s
ets
co
n
tai
n
s
a
n
y
t
y
p
e
o
f
d
ata
s
u
c
h
as
n
u
m
er
ic
o
r
ca
teg
o
r
ical
o
r
b
o
th
an
d
d
if
f
er
s
f
r
o
m
t
h
eir
attr
ib
u
te
s
.
C
o
n
v
e
n
t
io
n
al
clu
s
ter
i
n
g
al
g
o
r
ith
m
s
ca
n
n
o
t
p
er
f
o
r
m
w
ell
w
h
e
n
th
e
d
ata
s
ets
ar
e
m
i
x
ed
t
y
p
e.
I
n
o
r
d
er
to
g
r
o
u
p
th
e
d
is
s
i
m
ilar
d
ata
ty
p
es
t
h
er
e
is
a
clu
s
ter
i
n
g
e
n
s
e
m
b
le
a
p
p
r
o
ac
h
in
w
h
ich
a
co
m
b
in
ed
s
o
lu
tio
n
i
s
o
b
tain
ed
f
r
o
m
a
g
r
o
u
p
o
f
i
n
d
i
v
id
u
al
s
o
lu
tio
n
s
to
p
r
o
d
u
ce
a
q
u
alit
y
cl
u
s
ter
i
n
g
.
C
lu
s
ter
i
n
g
en
s
e
m
b
le
i
m
p
r
o
v
es
t
h
e
s
tr
en
g
th
a
n
d
s
ta
b
ilit
y
o
f
c
lu
s
ter
in
g
s
o
l
u
tio
n
d
u
e
to
its
co
n
s
o
lid
ati
n
g
an
d
d
iv
i
d
in
g
n
at
u
r
e.
C
l
u
s
ter
E
n
s
e
m
b
le
is
a
n
ap
p
r
o
ac
h
th
at
co
n
s
o
lid
ates
v
ar
io
u
s
f
in
d
i
n
g
s
o
f
d
is
s
i
m
ilar
cl
u
s
ter
i
n
g
s
to
b
r
in
g
o
u
t
t
h
e
f
in
al
q
u
alit
y
cl
u
s
ter
in
g
o
f
o
r
ig
i
n
al
d
ata
s
et.
C
lu
s
ter
i
n
g
e
n
s
e
m
b
les ar
e
m
o
r
e
ad
v
an
ta
g
eo
u
s
t
h
a
n
a
s
i
n
g
le
c
lu
s
ter
i
n
g
alg
o
r
it
h
m
in
s
o
m
an
y
s
tr
an
d
s
R
o
b
us
tne
s
s
:
T
h
e
clu
s
ter
in
g
e
n
s
e
m
b
le
i
m
p
r
o
v
es
th
e
av
er
a
g
e
p
er
f
o
r
m
an
ce
o
n
d
i
f
f
er
en
t
s
t
r
ea
m
s
a
n
d
d
ataset
s
m
o
r
e
t
h
a
n
a
s
in
g
le
cl
u
s
ter
in
g
.
No
ve
lty
: T
h
e
co
m
b
i
n
ed
s
o
lu
ti
o
n
g
i
v
es
u
n
u
s
u
al
r
es
u
lt
s
w
h
ic
h
ca
n
n
o
t b
e
p
r
o
d
u
ce
d
b
y
o
n
e
cl
u
s
ter
i
n
g
al
g
o
r
it
h
m
.
Sta
b
ilit
y:
C
lu
s
ter
i
n
g
e
n
s
e
m
b
l
e
w
o
r
k
s
ef
f
icie
n
tl
y
an
d
ca
n
h
a
n
d
le
n
o
is
e
an
d
o
u
tlier
s
.
P
a
r
a
lleliz
a
tio
n
a
nd
Sca
la
b
ilit
y
:
P
ar
alleliza
tio
n
o
f
cl
u
s
ter
in
g
ca
n
b
e
ac
q
u
ir
ed
b
y
s
u
cc
ess
i
v
e
s
y
n
t
h
esi
s
o
f
r
esu
lt
s
.
I
t h
as t
h
e
ab
ilit
y
to
a
m
alg
a
m
ate
r
esu
lts
f
r
o
m
m
u
ltip
le
h
eter
o
g
en
eo
u
s
s
o
u
r
ce
s
o
f
d
ata
.
C
lu
s
ter
i
n
g
e
n
s
e
m
b
les
ar
e
u
s
ed
in
m
a
n
y
ar
ea
s
.
S
u
ch
a
s
b
io
in
f
o
r
m
at
ics,
m
ac
h
in
e
le
ar
n
in
g
a
n
d
in
f
o
r
m
atio
n
r
etr
iv
a
l.
T
h
e
en
s
e
m
b
les
ar
e
f
o
r
m
u
la
ted
w
it
h
d
i
f
f
er
en
t
t
y
p
es
o
f
o
p
ti
m
izat
io
n
alg
o
r
ith
m
s
s
u
c
h
a
s
g
en
et
ic
alg
o
r
ith
m
s
,
ev
o
lu
t
io
n
ar
y
al
g
o
r
ith
m
s
,
k
-
p
ar
ticle
s
w
a
r
m
o
p
ti
m
izatio
n
al
g
o
r
ith
m
,
k
-
m
u
s
cles
w
a
n
d
er
i
ng
o
p
tim
izatio
n
al
g
o
r
it
h
m
s
i
n
d
if
f
er
en
t a
s
p
ec
ts
i
n
d
i
f
f
er
e
n
t a
r
ea
s
ar
e
s
p
ec
if
ied
i
n
t
h
e
f
o
llo
w
in
g
s
ec
tio
n
s
clea
r
l
y
i
n
ac
co
r
d
an
ce
w
it
h
s
o
m
e
j
o
u
r
n
al
p
ap
er
s
.
T
h
e
d
if
f
ic
u
lt
y
w
i
th
cl
u
s
ter
i
n
g
e
n
s
e
m
b
le
i
s
to
p
er
ce
iv
e
a
co
n
s
en
s
u
s
f
u
n
ct
io
n
.
T
h
er
e
ar
e
d
if
f
er
en
t
co
n
s
en
s
u
s
es
f
u
n
ctio
n
s
ar
e
av
ailab
le
b
u
t
i
n
o
r
d
er
to
in
cr
ea
s
e
th
e
s
tab
ilit
y
a
n
d
r
o
b
u
s
t
n
es
s
g
e
n
etic
al
g
o
r
it
h
m
w
it
h
co
-
ass
o
ciat
io
n
m
atr
i
x
is
u
s
ed
as
a
co
n
s
e
n
s
u
s
f
u
n
c
tio
n
.
T
h
e
g
en
etic
al
g
o
r
it
h
m
co
m
p
o
s
ed
o
f
f
o
u
r
p
h
a
s
es
in
cl
u
d
es
f
it
n
es
s
f
u
n
ctio
n
,
s
e
lectio
n
m
et
h
o
d
,
cr
o
s
s
o
v
er
m
et
h
o
d
an
d
f
in
al
l
y
m
u
tatio
n
m
et
h
o
d
.
T
h
e
co
-
ass
o
ciatio
n
m
atr
ix
v
a
lu
e
s
ar
e
u
s
ed
to
o
b
tain
t
h
e
in
tr
a
an
d
ex
tr
a
clu
s
ter
f
it
n
es
s
b
y
ev
a
lu
at
in
g
a
v
er
ag
e
s
i
m
ilar
it
y
b
et
w
ee
n
all
clu
s
ter
s
in
f
ir
s
t
p
h
ase.
I
n
t
h
e
s
ec
o
n
d
p
h
ase
to
u
r
n
a
m
en
t
s
e
lectio
n
is
u
s
ed
in
w
h
i
ch
t
w
o
i
n
d
i
v
id
u
al
s
ar
e
ad
o
p
ted
ar
b
itra
r
ily
an
d
th
e
in
d
i
v
id
u
al
w
it
h
p
r
ef
er
ab
le
f
it
n
es
s
is
elec
ted
f
o
r
n
e
x
t
p
o
p
u
la
tio
n
.
I
n
th
ir
d
p
h
as
e
th
e
t
w
o
o
f
f
s
p
r
in
g
s
ar
e
g
en
er
ated
b
y
t
h
e
i
n
d
iv
id
u
als
ar
e
ex
ch
a
n
g
ed
w
it
h
a
r
an
d
o
m
cr
o
s
s
o
v
er
p
o
in
t
.
I
n
telli
g
en
ce
m
u
tatio
n
i
s
u
s
ed
in
f
o
u
r
th
p
h
a
s
e
[
5
]
.
T
h
e
u
tili
za
t
io
n
o
f
lo
ca
ll
y
ad
ap
tiv
e
cl
u
s
ter
i
n
g
al
g
o
r
ith
m
p
r
o
v
id
es
a
n
i
m
p
le
m
en
tatio
n
to
id
en
ti
f
y
a
p
ar
titi
o
n
th
at
f
i
n
d
s
s
o
l
u
tio
n
s
to
th
e
clu
s
ter
s
.
I
t
i
m
p
ar
ts
s
et
o
f
clu
s
ter
s
w
it
h
s
o
m
e
w
ei
g
h
t
s
th
e
n
ass
ig
n
a
s
p
ec
if
ied
p
r
o
b
ab
ilit
y
to
ea
c
h
clu
s
ter
.
Usi
n
g
J
ac
ca
r
d
co
ef
f
ic
ien
t
f
i
n
d
i
n
ter
cl
u
s
ter
s
i
m
ilar
i
t
y
b
ased
o
n
f
ea
t
u
r
e
an
d
o
b
j
ec
t.
T
h
is
t
w
o
-
o
b
j
ec
tiv
e
clu
s
ter
i
n
g
en
s
e
m
b
le
co
m
p
li
ca
te
in
s
etti
n
g
p
ar
a
m
eter
an
d
in
i
n
ter
p
r
etatio
n
o
f
r
esu
lt
s
.
So
s
i
n
g
le
o
b
j
ec
tiv
e
clu
s
ter
i
n
g
is
co
m
p
o
s
ed
b
o
th
f
e
atu
r
e
b
ased
an
d
o
b
j
ec
t
b
ased
as
a
w
h
o
le
w
h
ic
h
in
cr
ea
s
es a
cc
u
r
ac
y
[
6
]
.
Fo
r
s
tr
ea
m
m
in
in
g
cl
u
s
ter
i
n
g
en
s
e
m
b
le
is
i
m
p
ar
ted
.
T
h
is
in
te
g
r
ates
b
o
th
cl
u
s
ter
s
an
d
class
i
f
ier
s
to
g
eth
er
a
n
d
e
m
p
lo
y
g
en
etic
a
lg
o
r
ith
m
a
n
d
h
as
h
ig
h
p
r
o
p
en
s
it
y
to
h
a
n
d
le
o
p
ti
m
izatio
n
[
7
].
T
h
e
clu
s
ter
i
n
g
en
s
e
m
b
le
is
d
esig
n
ed
as
a
n
o
p
ti
m
izatio
n
p
r
o
b
lem
o
n
m
u
ltip
le
o
b
j
ec
ts
b
y
ad
o
p
tin
g
ev
o
lu
tio
n
ar
y
al
g
o
r
ith
m
o
n
m
u
ltip
le
o
b
j
ec
ts
.
T
h
e
f
ir
s
t
cr
iter
ia
in
m
u
lti
o
b
j
ec
tiv
e
clu
s
ter
i
n
g
e
n
s
e
m
b
le
ar
e
to
m
ax
i
m
ize
th
e
s
i
m
ilar
it
y
m
ea
s
u
r
e
o
f
f
i
n
al
cl
u
s
te
r
i
n
g
f
r
o
m
all
in
p
u
t
c
lu
s
ter
in
g
s
.
T
h
e
s
i
m
ilar
it
y
m
ea
s
u
r
e
i
s
ca
lcu
lated
u
s
i
n
g
ad
j
u
s
ted
r
an
d
o
m
in
d
e
x
.
T
h
e
s
ec
o
n
d
cr
iter
io
n
i
n
m
u
lti
o
b
j
ec
tiv
e
cl
u
s
ter
in
g
en
s
e
m
b
le
is
t
o
r
ed
u
ce
th
e
s
i
m
ilar
it
y
m
ea
s
u
r
e
[
8
].
T
h
e
clu
s
ter
in
g
en
s
e
m
b
le
is
d
esig
n
ed
u
s
i
n
g
th
r
ee
d
if
f
er
e
n
t
alg
o
r
ith
m
s
k
-
m
ea
n
s
,
k
-
p
ar
ticle
s
w
ar
m
o
p
tim
izatio
n
an
d
k
-
m
u
s
cles
w
a
n
d
er
in
g
o
p
ti
m
izatio
n
.
T
h
e
co
m
b
i
n
atio
n
o
f
k
-
m
ea
n
s
w
i
th
m
u
s
cle
s
w
a
n
d
er
in
g
o
p
tim
izatio
n
o
v
er
co
m
es
t
h
e
s
h
o
r
tco
m
i
n
g
s
o
f
k
-
m
ea
n
s
al
g
o
r
ith
m
.
I
t
i
m
p
le
m
e
n
t
s
s
i
m
i
lar
it
y
b
ased
cl
u
s
ter
i
n
g
alg
o
r
ith
m
u
s
i
n
g
w
eig
h
t
s
o
n
in
p
u
t
d
ata.
Sa
m
p
les
t
h
e
d
ataset
f
ir
s
t
an
d
t
h
en
ap
p
l
y
cl
u
s
ter
i
n
g
al
g
o
r
it
h
m
s
s
p
ec
if
ied
o
n
s
u
b
s
a
m
p
le
s
w
h
ic
h
g
iv
e
clu
s
ter
in
g
r
e
s
u
lt
s
.
Fro
m
t
h
at
s
i
m
ilar
it
y
m
atr
ices
ar
e
g
en
er
ated
.
B
ased
o
n
v
ar
io
u
s
m
etr
ics
o
f
clu
s
ter
i
n
g
b
est
clu
s
ter
i
n
g
ca
n
b
e
d
er
iv
ed
.
R
ed
u
ce
th
e
w
ei
g
h
ts
o
f
t
h
e
s
a
m
p
les.
R
ep
ea
t
th
e
p
r
o
ce
s
s
u
n
ti
l b
est r
esu
l
tan
t c
l
u
s
ter
in
g
f
o
u
n
d
[
9
]
.
T
h
e
clu
s
ter
i
n
g
e
n
s
e
m
b
le
i
s
i
n
tr
o
d
u
ce
d
b
ased
o
n
p
ar
ticle
s
w
ar
m
cl
u
s
ter
in
g
.
T
h
e
p
ar
ticle
s
w
ar
m
clu
s
ter
i
n
g
is
ac
t
as
a
b
ase
clu
s
ter
er
an
d
as
w
ell
as
co
n
s
e
n
s
u
s
f
u
n
ctio
n
is
a
ch
allen
g
i
n
g
ele
m
en
t.
T
h
e
co
n
s
en
s
u
s
f
u
n
ctio
n
allo
w
s
t
h
e
b
ase
p
ar
titi
o
n
s
w
i
th
d
i
f
f
er
en
t
n
u
m
b
er
o
f
clu
s
ter
s
a
n
d
p
er
m
it
s
b
o
th
d
is
j
o
in
t
an
d
o
v
er
lap
p
in
g
p
ar
titi
o
n
s
.
P
r
o
p
o
s
ed
en
s
e
m
b
le
p
r
o
d
u
ce
s
tati
s
tica
ll
y
b
etter
p
ar
titi
o
n
s
[
10
].
T
h
e
n
ex
t
p
ar
t
o
f
th
e
p
ap
er
is
o
r
g
an
ized
as
f
o
llo
w
s
.
Sectio
n
2
g
iv
e
s
c
o
n
ce
p
t
o
f
cl
u
s
ter
i
n
g
en
s
e
m
b
le,
S
ec
tio
n
3
ex
p
lai
n
s
t
ax
o
n
o
m
y
o
f
g
e
n
er
atio
n
m
et
h
o
d
s
,
S
ec
tio
n
4
s
p
ec
if
ies
ta
x
o
n
o
m
y
o
f
co
n
s
en
s
u
s
m
e
th
o
d
s
a
n
d
S
ec
tio
n
5
p
r
esen
ts
co
n
clu
s
io
n
an
d
f
u
tu
r
e
w
o
r
k
o
n
cl
u
s
ter
in
g
en
s
e
m
b
le.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
E
xten
s
ive
A
n
a
lysi
s
o
n
Gen
era
tio
n
a
n
d
C
o
n
s
en
s
u
s
Mech
a
n
is
ms o
f Cl
u
s
te
r
in
g
…
(
Y
.
Leela
S
a
n
d
h
ya
R
a
n
i
)
2353
2.
CL
US
T
E
RIN
G
E
NS
E
M
B
L
E
C
lu
s
ter
i
n
g
en
s
e
m
b
le
is
a
n
ap
p
r
o
ac
h
th
at
co
n
ca
te
n
ates
t
h
e
s
u
b
s
et
cl
u
s
ter
i
n
g
s
o
l
u
tio
n
s
i
n
to
q
u
alit
y
clu
s
ter
i
n
g
f
o
r
o
r
ig
i
n
al
d
ata
s
e
t.
Fro
m
F
ig
u
r
e
1
th
e
d
ata
s
et
ca
n
b
e
d
iv
id
ed
in
to
N
n
u
m
b
er
o
f
s
a
m
p
les
a
n
d
ap
p
lied
clu
s
ter
in
g
alg
o
r
ith
m
o
n
ea
ch
s
a
m
p
le
s
et,
cl
u
s
ter
i
n
g
en
s
e
m
b
le
g
e
n
er
ates
N
clu
s
t
er
in
g
s
o
l
u
tio
n
s
an
d
f
i
n
all
y
co
m
b
i
n
ed
th
e
s
o
l
u
tio
n
s
to
g
et
f
i
n
al
q
u
alit
y
cl
u
s
te
r
in
g
u
s
in
g
a
co
n
s
e
n
s
u
s
f
u
n
ct
io
n
.
T
h
e
clu
s
ter
in
g
en
s
e
m
b
le
h
as
b
ee
n
p
r
o
ce
s
s
ed
in
t
w
o
s
tep
s
.
O
n
e
i
s
g
en
er
atio
n
s
tep
w
h
ic
h
g
e
n
er
ates
t
h
e
n
u
m
b
er
o
f
cl
u
s
ter
in
g
s
o
lu
tio
n
s
.
Seco
n
d
o
n
e
i
s
co
n
s
e
n
s
u
s
s
tep
w
h
ic
h
co
m
b
i
n
es t
h
e
s
o
lu
tio
n
in
to
a
f
in
al
cl
u
s
ter
in
g
.
2
.
1
.
G
ener
a
t
io
n
s
t
ep
T
h
e
w
a
y
o
f
co
m
b
i
n
in
g
all
t
h
e
in
d
iv
id
u
al
cl
u
s
ter
i
n
g
s
o
lu
tio
n
s
o
f
s
u
b
s
e
ts
g
en
er
ated
f
r
o
m
o
r
ig
in
a
l
d
ata
s
et
is
ca
lled
en
s
e
m
b
li
n
g
.
T
h
e
f
ir
s
t
s
tep
is
g
e
n
er
atio
n
o
f
all
in
d
iv
id
u
al
clu
s
ter
in
g
s
o
l
u
tio
n
s
.
T
h
e
clu
s
ter
i
n
g
en
s
e
m
b
le
is
t
h
e
co
m
b
i
n
atio
n
o
f
clu
s
ter
in
g
r
es
u
lt
s
.
Giv
e
n
a
d
ata
s
e
ts
o
f
m
o
b
j
ec
ts
P
={
P
1
,P
2
,P
3
….
P
m
},
th
e
clu
s
ter
i
n
g
en
s
e
m
b
le
g
e
n
er
ates
n
n
u
m
b
er
o
f
cl
u
s
ter
in
g
s
r
ep
r
es
en
ted
as
β={β
1
,
β
2
, β
3
….
.
β
n
}[
1
]
.
E
ac
h
clu
s
ter
in
g
s
o
lu
tio
n
β
i
is
o
n
e
p
ar
t o
f
t
h
e
o
r
ig
i
n
al
d
ata
s
et
P
in
to
K
i
d
is
s
i
m
ilar
g
r
o
u
p
s
o
f
o
b
j
ec
ts
,
d
en
o
ted
as β
i
= a
i
1
,
…a
i
k
.
Fig
u
r
e
1
.
Gen
er
al
m
ec
h
an
is
m
o
f
clu
s
ter
in
g
en
s
e
m
b
le
2
.
2
.
Co
ns
ens
us
s
t
e
p
T
h
e
co
n
s
en
s
u
s
s
tep
is
u
s
ed
to
co
m
b
in
e
th
e
s
o
lu
t
io
n
s
o
f
cl
u
s
ter
in
g
s
a
n
d
i
s
t
h
e
i
m
p
o
r
ta
n
t
s
tep
in
a
n
y
alg
o
r
ith
m
o
f
cl
u
s
ter
in
g
en
s
e
m
b
le.
I
t
is
th
e
f
u
n
ctio
n
w
h
ic
h
i
m
p
r
o
v
es
t
h
e
r
esu
lts
o
f
s
i
n
g
le
clu
s
ter
i
n
g
alg
o
r
it
h
m
.
T
h
er
e
ar
e
t
w
o
w
a
y
s
to
ap
p
l
y
co
n
s
e
n
s
u
s
f
u
n
ctio
n
o
n
e
i
s
c
o
r
r
elatio
n
b
et
w
ee
n
o
b
j
ec
ts
an
d
o
p
tim
izatio
n
f
o
r
p
ar
titi
o
n
.
T
h
e
f
ir
s
t
o
n
e
is
to
an
al
y
ze
t
h
e
n
u
m
b
er
o
f
o
n
e
in
s
t
an
ce
b
elo
n
g
i
n
g
to
o
n
e
clu
s
ter
an
d
n
u
m
b
er
o
f
t
w
o
in
s
t
a
n
ce
s
b
elo
n
g
i
n
g
to
th
e
s
a
m
e
clu
s
ter
.
I
t
is
d
o
n
e
th
r
o
u
g
h
v
o
tin
g
ap
p
r
o
ac
h
an
d
C
o
-
As
s
o
ciatio
n
Ma
tr
ix
b
ased
m
et
h
o
d
s
.
I
n
th
e
s
ec
o
n
d
ap
p
r
o
ac
h
o
f
co
n
s
e
n
s
u
s
f
u
n
c
tio
n
,
t
h
e
f
ea
tu
r
e
p
ar
titi
o
n
is
ac
q
u
ir
ed
in
as
s
o
ci
atio
n
w
it
h
o
p
tim
izatio
n
p
r
o
b
lem
[
11
]
.
T
h
e
p
ar
titi
o
n
c
a
n
b
e
f
i
n
d
b
y
u
s
in
g
s
o
m
e
s
i
m
ilar
it
y
b
et
w
ee
n
th
e
f
ea
tu
r
es
is
t
h
e
m
ai
n
p
r
o
b
lem
w
i
th
r
esp
ec
t to
t
h
e
clu
s
ter
en
s
e
m
b
le.
Fo
r
m
a
ll
y
,
th
e
f
ea
t
u
r
e
p
ar
titi
o
n
is
d
ef
in
e
d
as
:
∑
Her
e
ɼ
is
a
s
i
m
ilar
it
y
m
ea
s
u
r
e
b
et
w
ee
n
p
ar
titi
o
n
s
.
T
h
e
f
ea
t
u
r
e
p
ar
titi
o
n
is
th
e
m
ax
i
m
iza
tio
n
p
r
o
b
le
m
w
h
ic
h
is
g
iv
e
n
as
t
h
e
s
u
b
g
r
o
u
p
th
at
in
cr
ea
s
e
s
th
e
s
i
m
ilar
it
y
w
it
h
all
s
u
b
g
r
o
u
p
s
in
t
h
e
cl
u
s
ter
en
s
e
m
b
le.
T
h
e
f
o
llo
w
in
g
ar
e
t
h
e
e
x
a
m
p
les
u
s
e
f
ea
t
u
r
e
p
ar
titi
o
n
ar
e
k
er
n
el
b
ased
m
eth
o
d
s
a
n
d
n
o
n
-
n
eg
a
tiv
e
m
a
tr
ix
f
ac
to
r
izatio
n
.
3.
T
AXO
NAM
Y
O
F
CL
UST
E
RIN
G
E
N
SE
M
B
L
E
G
E
N
E
R
AT
I
O
N
M
E
T
H
O
DS
T
h
e
clu
s
ter
in
g
en
s
e
m
b
le
g
e
n
er
ates
th
e
s
et
o
f
cl
u
s
ter
in
g
s
o
lu
tio
n
s
b
y
ap
p
l
y
i
n
g
s
o
m
e
clu
s
ter
i
n
g
alg
o
r
ith
m
o
n
s
et
o
f
s
a
m
p
les
an
d
co
m
b
i
n
es
t
h
e
clu
s
ter
in
g
s
o
lu
tio
n
s
to
g
et
f
i
n
al
q
u
alit
y
cl
u
s
ter
in
g
.
T
h
e
m
ai
n
co
n
ce
p
t
is
to
h
an
d
le
d
if
f
er
en
t
ty
p
es
o
f
f
ea
tu
r
e
s
.
T
h
is
ca
n
b
e
s
o
lv
ed
b
y
r
an
d
o
m
l
y
s
e
lect
in
g
t
h
e
f
ea
t
u
r
es
o
n
b
asis
o
f
clu
s
ter
an
al
y
s
is
.
T
h
e
clu
s
ter
i
n
g
e
n
s
e
m
b
le
p
r
o
d
u
ce
s
ac
cu
r
ate
r
esu
l
ts
as
it
f
i
n
d
s
o
n
e
f
i
n
al
clu
s
ter
i
n
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2088
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
4
,
A
u
g
u
s
t
201
8
:
2
3
5
1
–
2
3
5
7
2354
f
r
o
m
s
et
o
f
cl
u
s
ter
i
n
g
s
b
ase
d
o
n
d
if
f
er
en
t
s
a
m
p
les
w
it
h
d
if
f
er
e
n
t
al
g
o
r
ith
m
s
f
o
r
g
e
n
e
r
atio
n
p
r
o
ce
s
s
.
W
e
p
r
o
v
id
e
s
o
m
e
o
f
th
e
g
en
er
atio
n
m
et
h
o
d
s
w
h
ich
ar
e
r
ev
ie
w
ed
f
r
o
m
p
r
ev
io
u
s
p
ap
er
s
.
3
.
1
.
S
i
m
ila
r
e
n
s
e
m
bl
ing
T
h
e
en
s
e
m
b
le
g
e
n
er
atio
n
p
r
o
ce
s
s
u
s
es
s
i
m
ilar
al
g
o
r
ith
m
s
f
o
r
all
s
a
m
p
le
s
u
b
s
et
s
.
T
h
e
k
-
m
ea
n
s
alg
o
r
ith
m
f
ail
s
to
p
er
f
o
r
m
c
l
ass
d
is
co
v
er
y
e
f
f
ec
ti
v
el
y
o
n
d
ata
s
ets
b
ec
au
s
e
it
ass
u
m
es
th
at
t
h
e
d
ata
i
s
i
n
Gau
s
s
ia
n
d
is
tr
ib
u
tio
n
.
A
s
s
p
ec
tr
al
clu
s
ter
i
n
g
h
a
n
d
les
t
h
i
s
p
r
o
b
le
m
,
Sp
ec
tr
al
clu
s
ter
i
n
g
(
SC
)
alg
o
r
ith
m
i
s
u
s
ed
f
o
r
th
e
en
s
e
m
b
le
g
e
n
er
atio
n
i
n
Kn
o
w
led
g
e
b
ased
C
l
u
s
ter
E
n
s
e
m
b
le
(
KC
E
)
.
T
h
e
k
n
o
w
led
g
e
b
ased
c
lu
s
ter
en
s
e
m
b
le
r
an
d
o
m
l
y
g
e
n
er
ate
s
d
x
d
i
m
en
s
io
n
s
f
r
o
m
d
d
i
m
e
n
s
io
n
s
th
i
s
p
r
o
ce
s
s
co
n
tin
u
e
s
u
n
til
A
s
u
b
s
p
ac
es
g
e
n
er
ated
{D
1
,D
2
,
….
D
a
}.
A
s
p
ec
tr
al
cl
u
s
ter
i
n
g
alg
o
r
ith
m
i
s
ap
p
lied
to
ea
ch
s
u
b
s
p
ac
e
an
d
g
e
n
er
ates
A
cl
u
s
ter
i
n
g
s
o
lu
tio
n
s
.
Sp
ec
tr
al
clu
s
ter
i
n
g
p
ar
titi
o
n
t
h
e
d
ata
p
o
in
ts
in
to
K
class
es.
First
SC
co
n
s
tr
u
cts
a
n
af
f
i
n
it
y
m
atr
i
x
,
an
d
o
b
tain
s
a
n
o
r
m
alize
d
m
atr
i
x
X
,
an
d
t
h
en
ap
p
lies
k
-
m
ea
n
s
f
o
r
ea
ch
r
o
w
o
f
X
to
g
et
t
h
ese
p
o
in
ts
i
n
to
K
cl
u
s
ter
s
.
I
n
t
h
is
w
a
y
S
C
is
ap
p
lied
r
ep
ea
ted
ly
t
o
all
s
a
m
p
le
s
to
g
e
t
s
o
lu
tio
n
f
o
r
ea
ch
s
a
m
p
le
s
u
b
s
p
ac
e.
Nex
t
a
co
n
f
id
en
ce
f
ac
to
r
w
as
ca
lcu
la
ted
f
o
r
all
t
h
e
cl
u
s
ter
in
g
s
o
l
u
tio
n
s
b
y
co
n
s
tr
u
ct
in
g
ad
j
ac
en
c
y
m
atr
i
x
o
n
p
air
w
i
s
e
co
n
s
tr
ai
n
t
s
.
I
f
t
h
e
m
o
s
t
o
f
th
e
p
air
w
is
e
co
n
s
tr
ain
ts
ar
e
s
ati
s
f
ied
b
y
t
h
e
clu
s
ter
in
g
s
o
lu
tio
n
it
w
i
ll h
a
v
e
h
ig
h
co
n
f
id
en
ce
m
ea
s
u
r
e
o
th
er
w
i
s
e
les
s
co
n
f
id
en
ce
m
ea
s
u
r
e
[
12
].
3
.
2
.
Ra
nd
o
m
ini
t
ia
liza
t
io
n o
f
inp
ut
pa
ra
m
et
er
s
T
h
e
“
P
r
o
j
ec
tiv
e
C
lu
s
ter
in
g
E
n
s
e
m
b
le”
(
P
C
E
)
is
b
ased
o
n
a
s
et
o
f
h
eter
o
g
en
eo
u
s
g
e
n
e
-
to
-
cl
u
s
ter
ass
i
g
n
m
e
n
ts
an
d
s
a
m
p
le
-
to
-
cl
u
s
ter
a
s
s
i
g
n
m
e
n
t
s
.
I
n
p
u
t
to
t
h
e
P
C
E
is
ta
k
e
n
f
r
o
m
g
e
n
e
ex
p
r
ess
io
n
d
ata
G.
E
ac
h
en
tr
y
o
f
G
r
ep
r
esen
ts
a
g
e
n
e
ex
p
r
ess
io
n
lev
e
l
o
f
a
p
ar
ti
c
u
lar
g
en
e.
I
f
w
e
g
r
o
u
p
s
a
m
p
le
s
in
to
clu
s
ter
s
u
s
e
s
a
m
p
le
-
to
-
c
lu
s
ter
ass
i
g
n
m
e
n
t.
T
h
e
p
r
o
b
ab
ilit
y
o
f
a
s
a
m
p
le
t
h
at
b
elo
n
g
s
to
a
clu
s
ter
is
n
o
t
h
in
g
b
u
t
s
a
m
p
le
-
to
-
clu
s
ter
a
s
s
i
g
n
m
e
n
t.
I
f
w
e
g
r
o
u
p
g
e
n
es
in
to
cl
u
s
ter
s
u
s
e
g
e
n
e
-
to
-
cl
u
s
ter
a
s
s
i
g
n
m
e
n
t.
T
h
e
p
r
o
b
ab
il
ity
o
f
g
e
n
e
b
elo
n
g
i
n
g
to
a
clu
s
ter
is
n
o
th
i
n
g
b
u
t g
e
n
e
-
to
-
cl
u
s
ter
ass
i
g
n
m
en
t.
T
h
ese
ass
i
g
n
m
e
n
t
s
ar
e
p
r
o
d
u
ce
d
b
y
ap
p
l
y
in
g
co
n
t
in
u
o
u
s
p
r
o
j
ec
tiv
e
clu
s
ter
in
g
N
t
i
m
e
s
w
it
h
d
if
f
er
e
n
t
r
an
d
o
m
in
itia
lizatio
n
s
f
o
r
in
p
u
t
p
ar
am
eter
s
to
p
r
o
d
u
ce
N
clu
s
ter
i
n
g
s
o
l
u
tio
n
s
,
w
h
ic
h
ar
e
u
s
ed
as
m
ai
n
clu
s
ter
i
n
g
f
o
r
co
n
s
e
n
s
u
s
cl
u
s
te
r
in
g
[
13
].
3
.
3
.
F
ea
t
ure
s
elec
t
io
n f
o
r
s
a
m
pli
ng
Set
o
f
s
a
m
p
le
s
u
b
s
et
ca
n
b
e
g
en
er
ated
b
ased
o
n
r
an
d
o
m
s
a
m
p
li
n
g
tech
n
iq
u
es
to
g
e
n
er
ate
s
et
o
f
clu
s
ter
i
n
g
s
o
lu
t
io
n
s
.
No
w
a
d
a
y
s
lar
g
e
d
i
m
en
s
io
n
al
d
ata
s
et
s
ar
e
u
s
ed
f
o
r
d
ata
an
al
y
s
i
s
.
S
o
f
ea
tu
r
e
s
elec
tio
n
p
lay
s
a
n
i
m
p
o
r
tan
t
r
o
le
in
g
e
n
er
atio
n
o
f
s
a
m
p
le
s
u
b
s
e
ts
.
I
n
“
Do
u
b
le
Selec
tio
n
Se
m
i
Su
p
er
v
is
ed
C
l
u
s
ter
i
n
g
E
n
s
e
m
b
le”
(
DSSS
C
E
)
th
e
y
u
s
ed
f
ea
tu
r
e
s
elec
tio
n
m
e
th
o
d
s
t
o
r
em
o
v
e
n
o
is
e
a
n
d
o
u
tlier
s
.
T
h
e
DSSS
C
E
u
s
e
in
p
u
t
f
r
o
m
g
en
e
e
x
p
r
ess
io
n
d
ata.
I
t
f
ir
s
t
a
p
p
lies
a
s
et
o
f
f
ea
tu
r
e
s
elec
t
io
n
m
et
h
o
d
s
s
u
c
h
as
M
u
tu
a
l
I
n
f
o
r
m
atio
n
Ma
x
i
m
izatio
n
(
MI
M)
,
Mu
t
u
al
I
n
f
o
r
m
a
tio
n
Feat
u
r
e
Selec
tio
n
(
MI
FS
)
,
J
o
in
t
Mu
t
u
al
I
n
f
o
r
m
atio
n
(
J
MI
)
,
C
o
n
d
itio
n
a
l
I
n
f
o
m
ax
Feat
u
r
e
E
x
tr
ac
tio
n
(
C
I
FE)
,
C
o
n
d
itio
n
al
R
ed
u
n
d
an
c
y
(
C
ONDRE
D)
,
I
n
ter
ac
tio
n
C
a
p
p
in
g
(
I
C
A
P
)
,
Do
u
b
le
I
n
p
u
t
S
y
m
m
etr
ical
R
ele
v
a
n
ce
(
DI
SR
)
,
Ma
x
-
R
ele
v
an
c
e
Min
-
R
ed
u
n
d
a
n
c
y
(
M
R
MR)
to
s
elec
t set o
f
s
u
b
s
a
m
p
les.
L
ater
DS
SS
C
E
ap
p
lies
P
C
-
K
m
ea
n
s
to
id
e
n
ti
f
y
th
e
lab
el
s
o
f
t
h
e
ca
n
ce
r
d
ataset.
T
h
is
alg
o
r
ith
m
co
n
s
id
er
s
th
e
n
u
m
b
er
o
f
m
u
s
t
-
li
n
k
a
n
d
ca
n
n
o
t
-
li
n
k
co
n
s
tr
ai
n
ts
b
et
w
ee
n
p
air
s
o
f
ca
n
ce
r
s
a
m
p
les
w
h
ich
lead
s
to
clu
s
ter
in
g
s
o
lu
tio
n
.
Us
in
g
f
ea
t
u
r
e
s
elec
tio
n
m
et
h
o
d
s
a
s
a
s
e
lectio
n
s
tr
ate
g
y
it
s
ele
cts
s
et
o
f
c
lu
s
ter
in
g
s
o
lu
tio
n
s
a
n
d
ag
g
r
e
g
ate
all
t
h
e
s
o
lu
tio
n
s
b
y
b
u
ild
i
n
g
m
atr
i
x
in
t
h
e
f
ir
s
t
p
h
a
s
e.
Nex
t,
D
SS
S
C
E
d
iv
id
es
t
h
e
ag
g
r
e
g
ated
s
o
l
u
tio
n
i
n
to
s
et
o
f
cl
u
s
ter
i
n
g
s
o
l
u
tio
n
s
a
n
d
c
alcu
late
th
e
co
n
f
id
en
ce
f
ac
to
r
s
f
o
r
th
e
cl
u
s
ter
i
n
g
s
o
lu
tio
n
s
b
ased
o
n
p
r
io
r
k
n
o
wled
g
e
o
f
th
e
d
ata
s
et
w
h
ic
h
is
s
p
ec
if
ied
b
y
p
air
w
i
s
e
co
n
s
tr
ai
n
ts
[
14
]
.
3
.
4
.
I
ncre
m
e
nta
l
e
ns
e
m
b
lin
g
I
n
“
I
n
cr
e
m
e
n
tal
Se
m
i
S
u
p
er
v
is
ed
C
l
u
s
ter
in
g
E
n
s
e
m
b
le”
(
I
SS
C
E
)
f
ir
s
t
o
n
e
o
r
ig
i
n
al
e
n
s
e
m
b
le
is
g
en
er
ated
.
T
h
en
th
e
f
i
n
al
n
e
w
en
s
e
m
b
le
is
p
r
o
d
u
ce
d
w
it
h
th
e
h
elp
o
f
s
et
o
f
s
elec
tio
n
m
e
m
b
er
s
.
I
t
g
en
er
ates
t
w
o
e
n
s
e
m
b
les
u
s
i
n
g
r
an
d
o
m
s
u
b
s
p
ac
e
g
e
n
er
atio
n
m
et
h
o
d
a
s
a
s
u
b
s
p
ac
e
g
en
er
ato
r
,
C
o
n
s
tr
ain
t
P
r
o
p
ag
atio
n
ap
p
r
o
ac
h
as a
clu
s
ter
in
g
al
g
o
r
ith
m
.
T
h
e
Do
u
b
le
Selectio
n
Se
m
i
S
u
p
er
v
is
ed
C
l
u
s
ter
i
n
g
E
n
s
e
m
b
l
e
f
ea
t
u
r
e
s
elec
t
io
n
m
et
h
o
d
s
ar
e
u
s
ed
a
s
a
s
u
b
s
et
s
elec
tio
n
an
d
clu
s
ter
in
g
ap
p
lied
in
t
w
o
p
h
ases
.
T
h
e
I
SS
C
E
also
u
s
ed
t
w
o
en
s
e
m
b
l
es
in
th
e
d
esi
g
n
.
T
o
h
an
d
le
h
i
g
h
d
i
m
en
s
io
n
al
d
ata
s
p
ac
e
u
s
e
r
a
n
d
o
m
s
u
b
s
p
ac
e
m
et
h
o
d
o
lo
g
y
to
g
e
n
er
ate
s
et
o
f
s
u
b
s
p
ac
es.
A
p
p
l
y
co
n
s
tr
ain
t
p
r
o
p
ag
atio
n
m
et
h
o
d
o
lo
g
y
o
n
s
et
o
f
s
u
b
s
p
ac
es
to
p
r
o
d
u
ce
s
et
o
f
clu
s
ter
i
n
g
s
o
lu
tio
n
s
.
T
h
e
I
SS
C
E
in
co
r
p
o
r
ated
in
cr
e
m
e
n
tal
m
e
m
b
er
s
elec
t
io
n
p
r
o
ce
s
s
b
ased
o
n
lo
ca
l
an
d
g
lo
b
al
co
s
t
f
u
n
ctio
n
a
n
d
p
r
o
d
u
ce
d
n
e
w
en
s
e
m
b
le
w
i
th
s
a
m
e
al
g
o
r
ith
m
u
s
ed
i
n
f
ir
s
t e
n
s
e
m
b
le
[
15
].
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
E
xten
s
ive
A
n
a
lysi
s
o
n
Gen
era
tio
n
a
n
d
C
o
n
s
en
s
u
s
Mech
a
n
is
ms o
f Cl
u
s
te
r
in
g
…
(
Y
.
Leela
S
a
n
d
h
ya
R
a
n
i
)
2355
3
.
5
.
Dis
s
i
m
ila
r
e
ns
e
m
bli
ng
T
h
e
g
en
er
atio
n
s
tep
i
n
d
is
s
i
m
ilar
e
n
s
e
m
b
li
n
g
in
v
o
l
v
es
d
i
f
f
er
en
t
cl
u
s
ter
in
g
al
g
o
r
ith
m
s
an
d
f
i
n
all
y
u
s
e
s
a
b
ase
cl
u
s
ter
i
n
g
.
T
h
e
n
u
m
b
er
o
f
clu
s
ter
s
i
s
g
en
er
ate
d
b
y
u
s
i
n
g
d
i
f
f
er
e
n
t
cl
u
s
ter
i
n
g
al
g
o
r
ith
m
w
it
h
i
ts
d
if
f
er
e
n
t
p
ar
a
m
eter
s
t
h
at
is
r
an
d
o
m
izatio
n
o
f
th
e
s
a
m
p
le
d
ata.
Dif
f
er
e
n
t
clu
s
ter
in
g
al
g
o
r
ith
m
s
m
a
y
g
i
v
e
d
if
f
er
e
n
t c
l
u
s
ter
i
n
g
r
es
u
lt
s
d
u
e
to
p
r
o
p
er
ties
o
f
d
ata.
I
f
w
e
ap
p
l
y
d
i
f
f
er
en
t
n
u
m
b
er
o
f
cl
u
s
ter
i
n
g
alg
o
r
it
h
m
s
o
n
d
ata
s
et
t
h
en
s
et
o
f
d
if
f
er
e
n
t
clu
s
ter
i
n
g
s
o
lu
tio
n
s
m
a
y
o
cc
u
r
.
Am
o
n
g
th
e
m
a
co
m
p
r
o
m
is
ed
clu
s
ter
i
n
g
s
o
lu
tio
n
s
h
o
u
ld
b
e
id
en
ti
f
i
ed
.
A
ll
th
e
cl
u
s
ter
s
ca
n
b
e
id
en
tif
ied
b
y
u
s
in
g
an
y
o
n
e
o
f
th
e
ca
n
d
id
ate
clu
s
ter
i
n
g
alg
o
r
ith
m
.
So
it
m
u
s
t
f
o
llo
w
th
at
an
y
d
ata
p
o
in
t
m
u
s
t
as
s
i
g
n
ed
to
o
n
l
y
o
n
e
cl
u
s
ter
an
d
ev
er
y
d
ata
p
o
in
t
in
t
h
e
s
et
m
u
s
t
as
s
i
g
n
ed
to
an
y
cl
u
s
ter
.
I
t
is
n
ec
ess
ar
y
to
in
ter
p
r
et
all
t
h
e
p
ar
titi
o
n
s
w
h
et
h
er
th
e
y
f
o
llo
w
ab
o
v
e
m
en
tio
n
ed
cr
iter
ia
o
r
n
o
t.
Use
g
o
o
d
n
e
s
s
f
u
n
ctio
n
to
ev
alu
a
te
th
e
q
u
a
lit
y
o
f
th
e
cl
u
s
ter
.
Select
ce
r
tain
clu
s
ter
i
n
g
s
o
lu
tio
n
s
u
s
in
g
g
o
o
d
n
ess
f
u
n
cti
o
n
[
16
]
.
4.
T
AXO
NAM
Y
O
F
CL
UST
E
RIN
G
E
N
SE
M
B
L
E
CO
NS
E
NSUS M
E
T
H
O
DS
T
h
er
e
ar
e
v
ar
io
u
s
t
y
p
e
s
o
f
co
n
s
e
n
s
u
s
f
u
n
ctio
n
s
t
h
e
y
ar
e
H
y
p
er
g
r
ap
h
P
ar
titi
o
n
i
n
g
,
C
o
-
a
s
s
o
ciatio
n
b
ased
f
u
n
c
tio
n
s
,
Mu
t
u
al
I
n
f
o
r
m
atio
n
A
l
g
o
r
ith
m
,
Fi
n
ite
Mi
x
tu
r
e
m
o
d
el
an
d
Vo
ti
n
g
A
p
p
r
o
ac
h
.
W
e
p
r
o
v
id
e
s
o
m
e
o
f
t
h
e
co
n
s
e
n
s
u
s
f
u
n
ctio
n
s
w
h
ic
h
ar
e
r
ev
ie
w
ed
f
r
o
m
p
r
ev
io
u
s
p
ap
er
s
.
4
.
1
.
Sp
ec
t
ra
l g
ra
ph
pa
rt
it
io
nin
g
Sp
ec
tr
al
C
lu
s
te
r
in
g
ch
o
o
s
es
a
s
p
ec
tr
al
g
r
ap
h
p
ar
titi
o
n
in
g
al
g
o
r
ith
m
,
w
h
ic
h
u
s
ed
to
o
p
tim
i
ze
th
e
cu
t
s
ca
le.
First
K
C
E
co
n
s
tr
u
cts
a
m
atr
i
x
b
y
co
n
s
id
er
i
n
g
all
th
e
g
en
er
ated
m
a
tr
ices
o
f
t
h
e
cl
u
s
ter
in
g
s
o
l
u
tio
n
s
an
d
r
esp
ec
ti
v
e
co
n
f
id
en
ce
f
ac
to
r
s
s
i
m
p
l
y
co
n
ca
te
n
atio
n
o
f
all
m
a
tr
ices.
Fin
a
ll
y
b
ased
o
n
s
p
ec
tr
al
cl
u
s
ter
i
n
g
alg
o
r
ith
m
it
p
ar
titi
o
n
s
th
e
n
e
w
f
ea
tu
r
es
i
n
to
K
clas
s
es.
Fo
r
t
h
e
m
aj
o
r
ity
o
f
th
e
ca
n
ce
r
d
ataset
s
K
C
E
o
u
tp
er
f
o
r
m
s
t
h
e
o
th
er
cl
u
s
ter
i
n
g
e
n
s
e
m
b
les.
KC
E
co
n
s
tr
u
cts
a
m
atr
i
x
b
y
s
p
ec
if
y
i
n
g
all
th
e
m
e
m
b
er
s
h
ip
m
atr
ice
s
o
f
t
h
e
clu
s
ter
in
g
s
o
l
u
tio
n
s
an
d
th
e
r
esp
ec
tiv
e
co
n
fi
d
en
ce
f
ac
t
o
r
s
as f
o
llo
w
s
:
W
h
er
e
is
is
th
e
r
ep
r
esen
tat
io
n
o
f
all
m
e
m
b
er
s
h
ip
m
atr
i
ce
s
o
f
t
h
e
clu
s
ter
in
g
s
o
l
u
tio
n
s
,
a
n
d
s
et
o
f
co
n
f
id
en
ce
v
ec
to
r
s
o
f
cl
u
s
ter
i
n
g
s
o
lu
tio
n
s
.
is
u
s
ed
to
co
n
ca
ten
ate
t
h
ese
t
w
o
.
Usi
n
g
s
p
ec
tr
al
clu
s
ter
i
n
g
p
ar
titi
o
n
t
h
e
n
e
w
f
e
atu
r
es o
f
co
n
ca
ten
ated
r
es
u
lt i
n
to
K
class
e
s
[
12
]
.
4
.
2
.
O
pti
m
iza
t
io
n a
lg
o
rit
h
m
I
t
is
n
ec
es
s
ar
y
f
o
r
a
cl
u
s
ter
in
g
e
n
s
e
m
b
le
to
f
i
n
d
a
co
n
s
e
n
s
u
s
f
u
n
ctio
n
th
at
m
in
i
m
izes
th
e
d
is
tan
c
e
f
r
o
m
all
cl
u
s
ter
s
s
o
th
e
f
o
llo
w
i
n
g
f
u
n
ct
io
n
is
o
p
ti
m
ized
.
{
}
Her
e
is
u
s
ed
as
a
d
is
tan
ce
f
u
n
ctio
n
f
o
r
th
e
clu
s
ter
in
g
s
.
P
C
E
o
p
tim
ize
th
e
f
o
r
t
w
o
r
eq
u
ir
e
m
en
ts
g
en
e
-
to
-
clu
s
ter
an
d
s
a
m
p
le
-
to
-
clu
s
ter
ass
ig
n
m
en
t.
So
E
x
p
ec
tatio
n
Ma
x
i
m
izat
io
n
o
f
P
r
o
j
ec
tiv
e
C
lu
s
ter
i
n
g
E
n
s
e
m
b
le
(
E
M
-
P
C
E
)
i
s
u
s
ed
as
a
co
n
s
en
s
u
s
f
u
n
ct
io
n
.
T
h
e
m
ai
n
ai
m
o
f
E
M
-
P
C
E
is
to
m
i
n
i
m
ize
t
h
e
er
r
o
r
th
at
co
r
r
esp
o
n
d
s
to
b
o
th
s
am
p
le
to
clu
s
ter
a
n
d
g
en
e
to
cl
u
s
ter
as
s
i
g
n
m
e
n
t [
13
].
4
.
3
.
G
ra
ph
pa
rt
i
t
io
nin
g
I
n
“
d
o
u
b
le
s
elec
tio
n
s
e
m
i
s
u
p
er
v
is
ed
clu
s
ter
i
n
g
en
s
e
m
b
le
”
th
e
y
d
esi
g
n
ed
co
n
s
e
n
s
u
s
f
u
n
ct
io
n
b
y
co
m
b
i
n
i
n
g
all
t
h
e
m
e
m
b
er
s
h
i
p
m
atr
ices
o
f
t
h
e
clu
s
ter
in
g
s
o
lu
tio
n
s
an
d
co
r
r
esp
o
n
d
in
g
co
n
f
id
e
n
ce
f
ac
to
r
s
to
o
n
e
m
atr
i
x
A
.
B
ased
o
n
th
e
s
a
m
p
le
s
et
Y
a
g
r
ap
h
is
co
n
s
tr
u
cted
o
n
Y
an
d
A
.
Us
in
g
th
e
n
o
r
m
alize
d
cu
t
ap
p
r
o
ac
h
o
n
th
e
co
n
s
tr
u
c
ted
g
r
ap
h
,
th
e
f
i
n
al
cl
u
s
ter
i
n
g
o
f
t
h
e
o
r
ig
in
al
d
ata
s
et
i
s
o
b
tain
ed
[
14
]
,
[
15
]
.
4
.
4
.
H
ill cl
i
m
bi
ng
B
ased
u
p
o
n
t
h
e
g
o
o
d
n
es
s
f
u
n
ctio
n
t
h
e
n
u
m
b
er
o
f
cl
u
s
ter
i
n
g
s
o
l
u
tio
n
s
ca
n
b
e
o
b
tain
ed
.
T
o
g
en
er
ate
clu
s
ter
i
n
g
s
o
lu
t
io
n
s
t
h
er
e
ar
e
t
w
o
co
n
f
licts
,
o
n
e
i
s
ab
s
en
c
e
co
n
f
lict
a
n
d
o
th
er
is
co
v
er
ag
e
co
n
f
lict.
So
th
e
co
n
s
id
er
atio
n
o
f
co
n
f
lict
s
b
ec
o
m
e
s
NP
h
ar
d
.
B
ased
o
n
h
ill
cli
m
b
i
n
g
ap
p
r
o
ac
h
th
e
o
p
ti
m
i
za
tio
n
p
r
o
b
lem
ca
n
b
e
s
o
lv
ed
an
d
f
i
n
all
y
g
et
s
o
n
e
clu
s
ter
i
n
g
f
o
r
th
e
g
i
v
en
d
ata
s
et
[
16
].
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2088
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
8
,
No
.
4
,
A
u
g
u
s
t
201
8
:
2
3
5
1
–
2
3
5
7
2356
5.
P
RO
P
O
SE
D
O
B
J
E
CT
I
VE
C
lu
s
ter
i
n
g
e
n
s
e
m
b
le
i
s
a
f
r
am
e
w
o
r
k
t
h
at
co
m
b
i
n
e
s
th
e
s
o
lu
tio
n
s
f
r
o
m
i
n
d
iv
id
u
al
cl
u
s
ter
in
g
s
to
p
r
o
d
u
ce
a
q
u
alif
ied
clu
s
ter
in
g
.
Ou
r
o
b
j
ec
tiv
e
is
to
p
r
ep
r
o
ce
s
s
t
h
e
d
ata
b
y
u
s
in
g
h
y
b
r
id
f
u
zz
y
lo
g
ic
f
ea
tu
r
e
s
elec
tio
n
m
et
h
o
d
w
h
ic
h
is
o
u
r
n
e
x
t
f
u
t
u
r
e
w
o
r
k
.
Fo
r
t
h
e
r
esu
lta
n
t
s
a
m
p
les
w
e
ap
p
l
y
d
if
f
er
e
n
t
cl
u
s
ter
i
n
g
alg
o
r
ith
m
s
a
n
d
f
i
n
all
y
w
e
g
et
q
u
alif
ied
clu
s
ter
in
g
.
Fro
m
F
ig
u
r
e
2
,
c
-
1
c
-
2
c
-
3
c
-
4
s
p
ec
if
ie
s
d
if
f
er
en
t
clu
s
ter
i
n
g
alg
o
r
it
h
m
s
.
Fig
u
r
e
2
.
E
x
ten
d
ed
clu
s
ter
in
g
en
s
e
m
b
le
6.
CO
NCLU
SI
O
N
AND
F
U
T
U
RE
WO
RK
C
lu
s
ter
i
n
g
e
n
s
e
m
b
le
is
a
f
r
a
m
e
w
o
r
k
th
at
p
r
o
v
id
es
s
e
t
o
f
c
lu
s
ter
in
g
s
o
l
u
tio
n
s
a
n
d
m
er
g
e
s
s
o
lu
tio
n
s
to
g
et
a
q
u
ali
f
ied
cl
u
s
ter
i
n
g
o
u
tp
u
t
f
o
r
th
e
g
iv
e
n
d
ata
s
et.
De
f
i
n
itel
y
it p
r
o
d
u
ce
s
m
o
r
e
ac
cu
r
at
e
r
esu
lt
s
as
w
i
th
t
h
e
s
in
g
le
clu
s
ter
i
n
g
an
d
also
i
m
p
r
o
v
es
t
h
e
r
o
b
u
s
tn
e
s
s
,
s
ca
lab
ili
t
y
a
n
d
q
u
a
lit
y
o
f
t
h
e
cl
u
s
ter
in
g
.
I
n
t
h
is
p
ap
er
w
e
r
ev
ie
w
ed
s
o
m
e
p
ap
er
s
w
h
ic
h
u
s
e
d
i
f
f
er
en
t
g
e
n
er
atio
n
m
et
h
o
d
s
an
d
d
if
f
er
en
t
co
n
s
e
n
s
u
s
f
u
n
ct
io
n
s
to
g
et
f
in
a
l
clu
s
ter
i
n
g
.
T
h
e
m
ai
n
a
s
p
ec
t
i
s
g
e
n
er
atio
n
m
ec
h
a
n
is
m
.
I
n
s
o
m
e
p
ap
er
s
th
e
y
u
s
ed
s
i
m
ilar
alg
o
r
ith
m
s
a
n
d
in
s
o
m
e
t
h
e
y
u
s
ed
d
is
s
i
m
ilar
al
g
o
r
ith
m
s
f
o
r
g
e
n
er
atio
n
p
r
o
ce
s
s
.
Fo
r
s
a
m
p
l
in
g
o
f
s
u
b
s
p
ac
e
s
o
m
e
u
s
ed
f
ea
t
u
r
e
s
elec
tio
n
a
n
d
s
o
m
e
u
s
ed
R
a
n
d
o
m
Sa
m
p
li
n
g
.
C
u
r
r
en
t
tr
en
d
s
h
a
n
d
le
lar
g
e
d
i
m
e
n
s
io
n
al
d
ata
s
ets
s
o
w
e
u
s
e
f
ea
t
u
r
e
s
elec
tio
n
m
et
h
o
d
s
f
o
r
r
ed
u
cin
g
t
h
e
d
i
m
en
s
io
n
al
it
y
an
d
in
cr
ea
s
i
n
g
th
e
p
er
f
o
r
m
an
ce
.
L
ater
w
e
ap
p
l
y
d
if
f
er
e
n
t
clu
s
ter
i
n
g
al
g
o
r
ith
m
s
f
o
r
ea
ch
s
u
b
s
et
g
e
n
er
ated
f
r
o
m
t
h
e
ap
p
licatio
n
o
f
f
ea
t
u
r
e
s
elec
tio
n
m
et
h
o
d
s
.
T
h
is
n
e
w
g
e
n
er
atio
n
s
tep
o
f
o
u
r
n
e
w
en
s
e
m
b
le
i
n
cr
ea
s
es
th
e
p
er
f
o
r
m
an
ce
o
f
t
h
e
f
i
n
al
cl
u
s
ter
in
g
s
o
lu
tio
n
a
s
w
e
ap
p
l
y
in
g
h
y
b
r
id
f
u
zz
y
lo
g
ic
f
ea
t
u
r
e
s
elec
tio
n
m
e
th
o
d
a
n
d
d
if
f
er
en
t
cl
u
s
ter
i
n
g
al
g
o
r
ith
m
s
.
I
f
w
e
r
e
m
o
v
e
th
e
n
o
is
e
a
n
d
r
ed
u
n
d
an
t
d
ata
f
r
o
m
t
h
e
d
a
ta
s
et
it
w
i
ll
in
cr
e
ases
th
e
p
er
f
o
r
m
a
n
ce
o
f
d
ata
an
al
y
s
is
.
I
t
is
d
o
n
e
w
it
h
h
y
b
r
id
f
u
zz
y
lo
g
ic
f
ea
t
u
r
e
s
elec
t
io
n
m
et
h
o
d
.
I
f
w
e
ap
p
ly
d
if
f
er
en
t
cl
u
s
ter
i
n
g
a
lg
o
r
ith
m
s
d
if
f
er
en
t
clu
s
ter
i
n
g
s
o
lu
tio
n
s
w
ill
b
e
g
en
er
ated
f
r
o
m
t
h
e
m
w
h
ic
h
ar
e
h
av
in
g
th
e
h
ig
h
e
s
t
s
i
m
il
ar
it
y
th
o
s
e
w
ill
b
e
co
n
s
id
er
ed
as
b
est
c
lu
s
ter
in
g
s
o
lu
tio
n
s
a
n
d
also
u
n
co
v
er
ed
clu
s
ter
s
f
r
o
m
d
i
f
f
er
e
n
t
s
o
lu
t
io
n
s
ar
e
a
m
alg
a
m
ated
to
g
et
f
i
n
al
cl
u
s
ter
i
n
g
s
o
lu
t
io
n
.
T
h
is
is
th
e
f
u
t
u
r
e
s
co
p
e
o
f
o
u
r
w
o
r
k
.
RE
F
E
R
E
NC
E
S
[1
]
Na
v
e
e
n
Ku
m
a
r,
S
.
S
iv
a
S
a
th
y
a
,
“
Clu
ste
rin
g
A
ss
isted
Co
-
lo
c
a
ti
o
n
P
a
tt
e
rn
M
i
n
in
g
f
o
r
S
p
a
ti
a
l
Da
ta
”
,
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
A
p
p
li
e
d
En
g
i
n
e
e
rin
g
R
e
se
a
rc
h
,
IS
S
N 0
9
7
3
-
4
5
6
2
,
v
o
l.
1
1
,
n
o
.
2
,
p
p
.
1
3
8
6
-
1
3
9
3
,
2
0
1
6
.
[2
]
M
.
Jo
h
n
Ba
sh
a
,
K.
P
.
Ka
li
y
a
m
u
rth
ie,
“
A
n
I
m
p
ro
v
e
d
S
im
il
a
rit
y
M
a
tc
h
in
g
b
a
se
d
Cl
u
ste
rin
g
F
ra
m
e
w
o
rk
f
o
r
S
h
o
rt
a
n
d
S
e
n
ten
c
e
L
e
v
e
l
Tex
t
”
,
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
E
lec
trica
l
a
n
d
Co
mp
u
ter
En
g
in
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
7
,
n
o
.
1
,
p
p
.
5
5
1
-
5
5
8
,
F
e
b
ru
a
ry
2
0
1
7
.
[3
]
Ch
a
ru
V
irm
a
n
i,
A
n
u
ra
d
h
a
P
il
lai,
Dim
p
le
Ju
n
e
ja,
“
Clu
ste
rin
g
in
Ag
g
re
g
a
ted
U
se
r
P
ro
f
il
e
s
a
c
ro
ss
M
u
lt
i
p
le
S
o
c
ial
Ne
tw
o
rk
s
”
,
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
Co
m
p
u
ter
En
g
in
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
7
,
n
o
.
6
,
p
p
.
3
6
9
2
-
3
6
9
9
De
c
e
m
b
e
r
2
0
1
7
.
[4
]
K.R.
Nirm
a
l,
K.V
.
V
.
S
a
ty
a
n
a
ra
y
a
n
a
,
“
Iss
u
e
s
o
f
K
M
e
a
n
s
Cl
u
ste
rin
g
W
h
il
e
M
ig
ra
ti
n
g
t
o
M
a
p
Re
d
u
c
e
P
a
ra
d
ig
m
w
it
h
Big
Da
ta:
A
S
u
rv
e
y
”
,
In
ter
n
a
ti
o
n
a
l
J
o
u
r
n
a
l
o
f
El
e
c
trica
l
a
n
d
Co
mp
u
ter
En
g
in
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
6
,
n
o
.
6
,
,
p
p
.
3
0
4
7
-
3
0
5
1
,
De
c
e
m
b
e
r
2
0
1
6
.
[5
]
Ja
v
a
d
A
z
i
m
i,
M
e
h
d
i
M
o
h
a
m
m
a
d
i,
A
li
m
o
v
a
g
h
a
r,
M
o
rtez
a
A
n
a
lo
u
i,
“
Clu
ste
ri
n
g
En
se
m
b
les
Us
in
g
G
e
n
e
ti
c
A
l
g
o
rit
h
m
”
,
T
h
e
In
ter
n
a
ti
o
n
a
l
W
o
rk
sh
o
p
o
n
C
o
mp
u
ter
Arc
h
it
e
c
tu
re
fo
r
M
a
c
h
in
e
Per
c
e
p
ti
o
n
a
n
d
S
e
n
sin
g
,
S
e
p
tem
b
e
r
2
0
0
6
.
[6
]
P
u
ru
s
h
o
t
h
a
m
a
n
B,
“
Clu
ste
rin
g
E
n
se
m
b
les
Us
in
g
Ev
o
lu
ti
o
n
a
ry
A
l
g
o
rit
h
m
”
,
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
f
o
r
Res
e
a
rc
h
in
Ap
p
li
e
d
S
c
ie
n
c
e
&
En
g
in
e
e
rin
g
T
e
c
h
n
o
l
o
g
y
,
I
S
S
N:
2
3
2
1
-
9
6
5
3
,
v
o
l
.
3
,
n
o
.
2,
F
e
b
r
u
a
ry
2
0
1
5
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
E
xten
s
ive
A
n
a
lysi
s
o
n
Gen
era
tio
n
a
n
d
C
o
n
s
en
s
u
s
Mech
a
n
is
ms o
f Cl
u
s
te
r
in
g
…
(
Y
.
Leela
S
a
n
d
h
ya
R
a
n
i
)
2357
[7
]
A
n
u
to
sh
P
ra
tap
S
in
g
h
,
Jiten
d
ra
Ag
ra
w
a
l,
V
a
rsh
a
S
h
a
rm
a
,
“
A
n
Eff
icie
n
t
A
p
p
ro
a
c
h
to
E
n
h
a
n
c
e
Clas
sif
ier
a
n
d
Clu
ste
r
En
se
m
b
les
Us
in
g
G
e
n
e
ti
c
a
lg
o
rit
h
m
s
f
o
r
M
in
i
n
g
Drif
ti
n
g
Da
ta
S
trea
m
s
M
u
lt
i
o
b
jec
c
ti
v
e
c
lu
ste
rin
g
”
,
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
C
o
mp
u
ter
Ap
p
l
ica
ti
o
n
s
IS
S
N 0
9
7
5
–
8
8
8
7
,
v
o
l.
4
4
,
n
o.
2
1
,
A
p
ril
2
0
1
2
.
[8
]
S
u
jo
y
Ch
a
tt
e
rjee
,
A
n
irb
a
n
M
u
k
h
o
p
a
d
h
y
a
y
,
“
Clu
ste
rin
g
En
se
m
b
le:
A
M
u
lt
io
b
jec
ti
v
e
G
e
n
e
ti
c
Alg
o
rit
h
m
b
a
se
d
A
p
p
ro
a
c
h
”
,
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
C
o
mp
u
ta
ti
o
n
a
l
In
telli
g
e
n
c
e
:
M
o
d
e
li
n
g
,
T
e
c
h
n
iq
u
e
s a
n
d
Ap
p
li
c
a
ti
o
n
,
2
0
1
3
.
[9
]
Qi
Ka
n
g
,
S
h
iYa
o
L
iu
,
M
e
n
g
C
h
u
Zh
o
u
,
S
i
S
i
L
i,
“
A
we
ig
h
t
-
in
c
o
rp
o
ra
ted
sim
il
a
rit
y
-
b
a
se
d
c
lu
st
e
rin
g
e
n
se
m
b
le
m
e
th
o
d
b
a
se
d
o
n
sw
a
r
m
in
telli
g
e
n
c
e
”
,
k
n
o
wled
g
e
b
a
se
d
sy
ste
ms
,
El
se
v
ier
.
[1
0
]
Jo
s´
e
V
a
len
te
d
e
Oliv
e
ira,
“
P
a
rti
c
le
S
w
a
r
m
Clu
ste
rin
g
in
c
lu
ste
rin
g
e
n
se
m
b
les
:
e
x
p
lo
it
in
g
p
r
u
n
i
n
g
a
n
d
a
li
g
n
m
e
n
t
f
re
e
c
o
n
se
n
su
s”
,
Ap
p
li
e
d
S
o
ft
C
o
mp
u
ti
n
g
,
F
e
b
1
3
,
2
0
1
6
.
[1
1
]
S
a
n
d
r
o
V
e
g
a
-
P
o
n
s an
d
Jo
se
Ru
iz
-
S
h
u
lcl
o
p
e
r,
“
A
S
u
rv
e
y
o
f
Clu
ste
ri
n
g
En
se
m
b
le A
l
g
o
rit
h
m
s”
,
2011.
[1
2
]
Zh
iw
e
n
Yu
,
Ha
u
-
S
a
n
W
o
n
g
b
,
Ja
n
e
Yo
u
,
Qin
m
in
Ya
n
g
a
n
d
Ho
n
g
y
in
g
L
iao
,
“
Kn
o
w
led
g
e
Ba
se
d
Clu
ste
r
En
se
m
b
le
f
o
r
Ca
n
c
e
r
Disc
o
v
e
r
y
f
ro
m
Bio
m
o
lec
u
lar Da
ta”
,
IEE
E
T
r
a
n
s
a
c
ti
o
n
s o
n
N
a
n
o
b
i
o
sc
ien
c
e
v
o
l.
1
0
,
n
o
2
,
Ju
n
e
2
0
1
1
.
[1
3
]
X
lan
x
u
e
Y
u
G
u
o
x
ian
Y
u
a
n
d
Ju
n
W
a
n
g
,
“
Clu
ste
ri
n
g
c
a
n
c
e
r
g
e
n
e
e
x
p
re
ss
io
n
d
a
ta
b
y
p
ro
jec
ti
v
e
c
lu
ste
rin
g
e
n
se
m
b
le”
,
Re
se
a
rc
h
a
rti
c
le,
Pl
o
s On
e
,
2
0
1
7
.
[1
4
]
Zh
iw
e
n
Yu
,
Ho
n
g
sh
e
n
g
Ch
e
n
Ja
n
e
Yu
,
Ha
u
S
a
n
W
o
n
g
,
Ji
m
in
g
L
iu
F
e
ll
o
w
L
e
L
i
G
u
o
q
ian
g
Ha
n
,
“
Do
u
b
le S
e
lec
ti
o
n
Ba
se
d
S
e
m
i
S
u
p
e
rv
ise
d
Clu
ste
rin
g
En
se
m
b
le
f
o
r
T
u
m
o
r
C
lu
ste
rin
g
f
ro
m
G
e
n
e
E
x
p
re
ss
io
n
P
r
o
f
il
e
s”
,
IEE
E/
ACM
T
ra
n
sa
c
ti
o
n
s
o
n
Co
m
p
u
t
a
ti
o
n
a
l
B
io
lo
g
y
a
n
d
Bi
o
in
f
o
rm
a
ti
c
s,
2
0
1
3
.
[1
5
]
Zh
iw
e
n
Yu
,
Ja
n
e
Yo
u
,
Ha
u
S
a
n
W
o
n
g
,
Ju
n
Zh
a
n
g
,
“
In
c
re
m
e
n
tal
se
m
i
su
p
e
rv
is
e
d
Clu
ste
rin
g
En
se
m
b
le
f
o
r
Hig
h
Dim
e
n
sio
n
a
l
d
a
ta Cl
u
ste
rin
g
”
,
a
rt
icle
,
IEE
E
T
r
a
n
sa
c
ti
o
n
s
o
n
Kn
o
w
led
g
e
a
n
d
Da
t
a
En
g
i
n
e
e
rin
g
,
2
0
1
5
.
[1
6
]
M
a
rti
n
H.C.
L
a
w
,
A
le
x
a
n
d
e
r
P
.
T
o
p
c
h
y
,
A
n
il
K.
Ja
in
“
M
u
lt
io
b
je
c
ti
v
e
Da
ta
Clu
ste
rin
g
”
,
IEE
E
Co
mp
u
ter
S
o
c
iety
Co
n
fer
e
n
c
e
o
n
Co
m
p
u
ter
Vi
si
o
n
a
n
d
Pa
tt
e
rn
Rec
o
g
n
it
io
n
,
2
0
4
.
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
Y.
L.
S
a
n
d
h
y
a
Ra
n
i
is
a
S
c
h
o
lar
in
K
L
Un
iv
e
r
sit
y
a
n
d
w
o
rk
in
g
a
s
a
n
a
ss
istan
t
p
ro
f
e
ss
o
r
in
S
ir
C
R
R
Co
ll
e
g
e
o
f
En
g
in
e
e
rin
g
,
El
u
ru
.
S
h
e
h
a
s
in
tere
ste
d
in
t
h
e
a
re
a
s
l
ik
e
Clu
ste
rin
g
in
d
a
ta
m
in
in
g
,
Da
ta
S
tru
c
tu
re
s
e
tc.
Cu
rre
n
t
ly
sh
e
w
a
s
w
o
rk
in
g
w
it
h
Clu
ste
rin
g
En
se
m
b
le
P
e
rf
o
rm
a
n
c
e
a
s
a
re
s
e
a
rc
h
wo
rk
.
S
h
e
h
a
s
p
u
b
li
sh
e
d
m
a
n
y
p
a
p
e
rs i
n
n
a
ti
o
n
a
l
a
n
d
in
tern
a
ti
o
n
a
l
j
o
u
r
n
a
l
s.
Dr
.
V
.
S
u
c
h
a
r
ita
is
a
p
ro
f
e
ss
o
r
in
Na
ra
y
a
n
a
En
g
in
e
e
rin
g
Co
ll
e
g
e
,
G
u
d
u
r.
S
h
e
h
a
s
p
u
b
li
sh
e
d
m
a
n
y
p
a
p
e
rs
in
n
a
ti
o
n
a
l
a
n
d
in
tern
a
ti
o
n
a
l
jo
u
r
n
a
ls.
A
lso
sh
e
is
re
v
ie
we
r
a
n
d
e
d
it
o
rial
b
o
a
rd
m
e
m
b
e
r
fo
r
re
p
u
ted
jo
u
rn
a
ls
Dr
.
K
.
V
.
V
.
S
a
ty
a
n
a
r
a
y
a
n
a
is
p
ro
f
e
ss
o
r
in
K
L
u
n
iv
e
rsit
y
.
He
h
a
s
p
u
b
l
ish
e
d
m
a
n
y
p
a
p
e
rs
in
n
a
ti
o
n
a
l
a
n
d
i
n
tern
a
ti
o
n
a
l
j
o
u
r
n
a
ls.
He
h
a
s
in
tere
ste
d
i
n
th
e
a
re
a
li
k
e
Bio
in
f
o
rm
a
ti
c
s an
d
c
lo
u
d
c
o
m
p
u
ti
n
g
.
Evaluation Warning : The document was created with Spire.PDF for Python.