I
nte
rna
t
io
na
l J
o
urna
l o
f
I
nfo
r
m
a
t
ics a
nd
Co
mm
u
n
ica
t
io
n T
ec
hn
o
lo
g
y
(
I
J
-
I
CT
)
Vo
l.
8
,
No
.
1
,
A
p
r
il
201
9
,
p
p
.
39
~
49
I
SS
N:
2252
-
8776
,
DOI
: 1
0
.
1
1
5
9
1
/i
j
ict.
v
8
i1
.
p
p
3
9
-
49
39
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
JI
C
T
Recent
tr
ends
in
big
da
ta
using
ha
do
o
p
Chet
na
K
a
us
ha
l,
Dee
pi
k
a
K
o
un
da
l
De
p
a
rtme
n
t
of
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
E
n
g
in
e
e
rin
g
,
C
h
it
k
a
ra
Un
iv
e
rsity
,
In
d
ia
.
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
No
v
9,
2018
R
ev
i
s
ed
Dec
25
,
2
0
1
8
A
cc
ep
ted
J
an
11,
2
0
1
9
Big
d
a
ta
re
f
e
r
s
to
h
u
g
e
se
t
of
d
a
ta
w
h
ich
is
v
e
r
y
c
o
m
m
o
n
th
e
se
d
a
y
s
due
to
th
e
in
c
re
a
se
of
in
tern
e
t
u
ti
li
ti
e
s.
Da
ta
g
e
n
e
ra
ted
f
ro
m
so
c
ial
m
e
d
ia
is
a
v
e
r
y
c
o
m
m
o
n
e
x
a
m
p
le
f
o
r
th
e
sa
m
e
.
T
h
is
p
a
p
e
r
d
e
p
icts
t
h
e
su
m
m
a
r
y
on
b
ig
d
a
ta
a
n
d
w
a
y
s
in
w
h
ich
it
h
a
s
b
e
e
n
u
ti
li
z
e
d
in
a
ll
a
sp
e
c
ts.
Da
ta
m
in
in
g
is
ra
d
ica
ll
y
a
m
o
d
e
of
d
e
riv
in
g
th
e
in
d
isp
e
n
sa
b
le
k
n
o
w
led
g
e
f
ro
m
e
x
t
e
n
siv
e
l
y
v
a
st
f
ra
c
ti
o
n
s
of
d
a
ta
w
h
ich
is
q
u
it
e
c
h
a
ll
e
n
g
in
g
to
be
in
terp
re
ted
by
c
o
n
v
e
n
ti
o
n
a
l
m
e
th
o
d
s.
T
h
e
p
a
p
e
r
m
a
in
l
y
f
o
c
u
se
s
on
th
e
issu
e
s
re
lat
e
d
to
th
e
c
lu
ste
rin
g
tec
h
n
iq
u
e
s
in
b
ig
d
a
ta.
F
o
r
th
e
c
las
si
f
ica
ti
o
n
p
u
r
p
o
se
of
th
e
b
ig
d
a
ta,
th
e
e
x
isti
n
g
c
las
si
f
ica
ti
o
n
a
lg
o
rit
h
m
s
a
re
c
o
n
c
ise
l
y
a
c
k
n
o
w
led
g
e
d
a
n
d
a
f
ter
th
a
t,
k
-
n
e
a
re
st
n
e
ig
h
b
o
u
r
a
lg
o
rit
h
m
is
d
isc
re
e
tl
y
c
h
o
se
n
a
m
o
n
g
th
e
m
a
n
d
d
e
sc
rib
e
d
a
lo
n
g
w
it
h
an
e
x
a
m
p
le.
K
ey
w
o
r
d
s
:
B
ig
D
ata
C
las
s
i
f
icatio
n
C
lu
s
ter
i
n
g
Kn
o
w
led
g
e
D
i
s
co
v
er
y
Min
i
n
g
Co
p
y
rig
h
t
©
201
9
In
stit
u
te
of
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
.
All
rig
h
ts
re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
C
h
et
n
a
Ka
u
s
h
a
l
Dep
ar
t
m
en
t
of
C
o
m
p
u
ter
Scie
n
ce
an
d
E
n
g
i
n
ee
r
in
g
,
C
h
i
tk
ar
a
U
n
iv
er
s
it
y
,
P
u
n
j
ab
,
I
n
d
ia
1.
I
NT
RO
D
UCT
I
O
N
T
h
e
B
ig
d
ata
is
p
r
i
m
ar
il
y
d
ef
in
ed
as
a
ter
m
th
at
g
en
er
all
y
d
escr
ib
es
th
e
lar
g
e
d
i
m
e
n
s
io
n
s
of
h
ig
h
v
elo
cit
y
,
d
i
f
f
ic
u
lt
a
n
d
v
ar
iab
le
d
ata
th
at
in
v
o
lv
e
in
n
o
v
a
tiv
e
t
ec
h
n
iq
u
es
a
n
d
eq
u
ip
m
e
n
t
to
f
a
cilitate
t
h
e
ca
p
tu
r
e,
s
to
r
ag
e,
s
h
ar
i
n
g
,
ad
m
i
n
i
s
tr
atio
n
,
an
d
an
al
y
s
is
of
t
h
e
d
ata
or
i
n
f
o
r
m
atio
n
[
1
]
.
B
ig
d
ata
u
lti
m
atel
y
s
u
r
p
ass
es
t
h
e
h
an
d
li
n
g
ab
ilit
y
of
tr
ad
itio
n
al
d
atab
ases
an
d
is
to
o
b
ig
to
be
m
a
n
ag
ed
by
a
s
in
g
le
m
ac
h
i
n
e.
T
h
er
ef
o
r
e,
n
o
v
el
an
d
ad
v
an
ce
d
w
a
y
s
ar
e
m
an
d
ato
r
y
to
p
r
o
ce
s
s
a
n
d
s
to
r
e
s
u
ch
an
e
n
o
r
m
o
u
s
s
ize
of
t
h
e
d
ata.
T
h
ese
d
ata
ar
e
p
r
o
d
u
ce
d
f
r
o
m
v
ir
t
u
al
tr
an
s
ac
tio
n
s
,
elec
tr
o
n
ic
m
ail
s
,
au
d
io
s
,
v
id
eo
s
,
p
ictu
r
e
s
,
to
r
r
en
ts
,
r
e
co
r
d
s
,
p
o
s
ts
,
s
ea
r
ch
r
eq
u
ests
,
f
i
tn
e
s
s
r
ec
o
r
d
s
,
s
o
cial
n
et
w
o
r
k
i
n
g
co
n
n
ec
tio
n
s
,
s
cien
ce
d
ata,
s
en
s
o
r
s
an
d
ce
ll
-
p
h
o
n
es
an
d
th
e
ir
ap
p
licatio
n
s
[
2
]
.
T
h
ey
ar
e
d
ep
o
s
ited
in
d
atab
ases
t
h
at
r
i
s
e
en
o
r
m
o
u
s
l
y
a
n
d
tu
r
n
o
u
t
to
be
c
h
alle
n
g
i
n
g
in
o
r
d
er
to
ca
p
tu
r
e,
ar
r
an
g
e,
s
to
r
e,
m
a
n
ag
e,
s
h
ar
e
an
d
a
n
al
y
ze
t
h
e
d
at
ab
ase
w
ith
th
e
u
s
e
of
s
ta
n
d
ar
d
d
atab
ase
s
o
f
t
w
ar
e
to
o
ls
.
Data
b
ase
Ma
ch
i
n
e
is
an
i
m
p
o
r
tan
t
p
ar
t
of
B
ig
d
ata
p
r
o
ce
s
s
in
g
.
T
h
e
id
ea
of
th
e
“
d
atab
ase
m
ac
h
in
e”
w
a
s
f
ir
s
t
ap
p
ea
r
ed
in
th
e
late
1
9
7
0
’
s,
it
is
an
eq
u
ip
m
e
n
t
th
a
t
w
a
s
s
p
ec
iall
y
b
u
ilt
f
o
r
th
e
p
u
r
p
o
s
e
of
s
to
r
ag
e
an
d
an
al
y
s
is
of
d
ata.
A
s
o
le
m
ai
n
f
r
a
m
e
n
et
w
o
r
k
ar
r
an
g
e
m
e
n
t
tu
r
n
ed
in
s
u
f
f
icie
n
t
w
i
th
th
e
i
n
cr
e
m
en
t
of
d
ata
e
x
te
n
t
an
d
th
e
d
ata
r
ep
o
s
ito
r
y
.
W
it
h
th
e
i
n
cr
ea
s
i
n
g
d
e
m
a
n
d
of
tec
h
n
o
lo
g
y
,
T
er
ad
ata
s
y
s
te
m
e
m
er
g
ed
as
th
e
lead
in
g
co
m
m
er
ciall
y
ef
f
icie
n
t
d
atab
ase
w
h
ic
h
w
a
s
ba
s
ed
u
p
o
n
th
e
p
ar
allel
s
y
s
te
m
.
In
1986,
a
b
r
ea
k
th
r
o
u
g
h
e
v
en
t
h
ap
p
en
ed
w
h
i
le
T
er
ad
ata
in
iti
all
y
b
r
o
u
g
h
t
th
e
s
y
s
te
m
of
p
ar
allel
d
atab
ase
co
m
p
r
is
i
n
g
t
h
e
ca
p
ac
it
y
of
s
to
r
in
g
d
ata
f
r
o
m
1
T
B
up
to
Km
ar
t
in
o
r
d
er
to
p
r
o
v
id
e
co
n
v
en
i
en
ce
to
th
e
r
etail
co
m
p
a
n
ies
at
lar
g
e
-
s
ca
le.
T
h
e
b
en
ef
it
s
of
th
e
p
ar
allel
s
y
s
te
m
b
ased
d
atab
ase
s
to
o
d
b
r
o
a
d
ly
ac
k
n
o
w
led
g
ed
in
th
e
d
o
m
ai
n
of
d
atab
ases
,
d
u
r
in
g
1
9
9
0
'
s
[
3
]
.
Fig
u
r
e
1
d
ep
icts
a
g
en
er
aliz
ed
ar
ch
itect
u
r
e
o
f
b
ig
d
ata
.
Go
o
g
le
f
o
r
m
u
lated
p
r
o
g
r
a
m
m
i
n
g
p
ar
ad
ig
m
s
lik
e
Ma
p
R
ed
u
ce
an
d
G
FS
,
to
co
p
e
up
w
it
h
t
h
e
tr
ials
b
r
o
u
g
h
t
at
t
h
e
I
n
ter
n
et
by
d
ata
ad
m
in
is
tr
atio
n
an
d
in
ter
p
r
etatio
n
.
B
esid
es,
th
e
lo
ad
g
en
er
ated
by
s
en
s
o
r
s
,
clien
ts
,
a
n
d
ad
d
itio
n
al
w
o
r
ld
w
id
e
r
eser
v
o
ir
s
of
d
ata,
f
u
r
t
h
er
p
o
w
er
ed
t
h
e
o
v
er
w
h
el
m
i
n
g
s
tr
ea
m
s
of
d
ata
t
h
at
lac
k
ed
ce
r
tain
a
m
en
d
m
en
t
on
t
h
e
co
m
p
u
ti
n
g
s
tr
u
ctu
r
e
an
d
f
ar
-
r
ea
c
h
i
n
g
d
ata
p
r
o
ce
s
s
i
n
g
m
ac
h
in
e
[
4
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8776
IJ
-
I
C
T
Vo
l.
8
,
No
.
1
,
A
p
r
il 2
0
1
9
:
39
–
49
40
Fig
u
r
e
1.
An
o
v
er
v
ie
w
of
b
ig
d
ata
P
r
ac
tically
,
all
t
h
e
f
o
r
e
m
o
s
t
e
s
tab
lis
h
m
e
n
t
s
h
a
v
e
in
itiated
t
h
eir
in
d
i
v
id
u
al
d
ev
e
lo
p
m
en
ts
co
n
ce
r
n
i
n
g
th
e
b
ig
d
ata
w
it
h
i
n
a
f
e
w
f
o
r
m
er
y
ea
r
s
,
co
m
p
r
i
s
in
g
Go
o
g
le,
Face
b
o
o
k
,
Mic
r
o
s
o
f
t,
E
MC,
Am
az
o
n
,
Or
ac
le,
an
d
I
B
M,
etc.
L
ik
e
w
i
s
e,
s
e
v
er
al
n
atio
n
-
w
id
e
g
o
v
er
n
m
en
ts
h
av
e
also
p
aid
ab
u
n
d
an
t
d
e
v
o
tio
n
to
b
ig
d
ata
an
d
m
ad
e
m
illi
o
n
s
of
f
u
n
d
s
to
i
n
itiate
t
h
e
P
r
o
j
ec
t
r
eg
ar
d
in
g
th
e
An
al
y
s
i
s
an
d
A
d
v
a
n
ce
m
en
t
of
th
e
B
ig
Data
[
5
].
T
h
e
co
n
clu
d
i
n
g
o
b
j
ec
tiv
e
of
b
ig
d
ata
s
ta
n
d
s
to
d
eliv
er
th
e
p
r
o
d
u
cti
v
it
y
as
s
o
m
e
co
m
m
er
cial
r
eso
lu
tio
n
s
t
h
at
can
co
m
f
o
r
t
a
co
m
p
an
y
to
g
ai
n
p
r
o
f
ess
io
n
al
s
o
l
u
tio
n
s
.
Fo
r
in
s
ta
n
ce
,
a
n
y
co
m
p
an
y
c
an
be
b
en
e
f
ited
if
t
h
e
y
co
u
ld
u
n
d
er
s
tan
d
t
h
at
if
cl
ien
t
p
u
r
ch
ase
s
“
X”
th
en
it
is
p
r
o
b
ab
le
th
at
h
e/
s
h
e
m
i
g
h
t
al
s
o
be
in
ter
ested
in
b
u
y
i
n
g
“
Y”.
T
h
is
t
y
p
e
of
an
al
y
s
is
at
r
u
n
-
ti
m
e
ca
n
g
r
ea
tl
y
b
e
n
ef
it
by
i
n
cr
ea
s
i
n
g
b
u
s
i
n
ess
.
T
h
e
w
eb
ac
co
u
n
ts
ar
e
an
al
y
s
ed
by
th
e
o
n
li
n
e
s
i
tes
o
f
f
er
in
g
h
u
m
a
n
in
ter
ac
tio
n
in
o
r
d
er
to
p
r
o
p
o
s
e
s
o
m
e
p
r
ef
er
en
ce
s
to
th
e
u
s
e
r
s
b
ased
u
p
o
n
th
eir
v
ested
i
n
t
er
ests
.
B
ig
d
ata
also
tar
g
ets
on
r
e
m
ar
k
ab
le
r
ed
u
ctio
n
in
e
x
p
en
s
es
a
n
d
n
ec
es
s
ar
y
d
ev
elo
p
m
e
n
ts
[
6
]
.
T
h
er
e
ar
e
th
r
ee
m
ai
n
k
e
y
s
f
o
r
b
ig
d
ata,
also
k
n
o
w
n
as
3
V’
s
of
b
ig
d
ata
.
[
7
]
i.
Vo
lu
m
e
-
P
r
esen
t
l
y
t
h
e
d
ata
s
ize
is
m
u
c
h
lar
g
er
in
co
m
p
a
r
is
o
n
to
p
ast
d
ata
s
izes,
i.e
.
e
x
ce
ed
in
g
-
ter
ab
y
te
s
a
n
d
p
eta
b
y
tes.
T
h
e
s
tr
ik
i
n
g
r
a
n
g
e
an
d
g
r
ad
u
al
s
u
r
g
e
in
th
e
d
ata
s
ize
ta
g
s
it
v
i
g
o
r
o
u
s
l
y
h
ar
d
to
s
av
e
an
d
r
e
v
ie
w
by
e
m
p
lo
y
in
g
t
h
e
co
n
v
e
n
tio
n
al
ap
p
r
o
ac
h
es.
Fo
r
in
s
ta
n
ce
,
Face
b
o
o
k
co
n
s
u
m
e
s
ap
p
r
o
x
im
a
tel
y
500
-
ter
ab
y
te
s
of
d
ata
on
a
d
ail
y
b
asis
.
ii.
Velo
cit
y
-
T
h
e
u
tili
za
t
io
n
of
t
h
e
b
ig
d
ata
is
m
u
s
t
as
it
s
tr
ea
m
s
t
h
e
d
ata
to
o
b
tain
th
e
o
p
ti
m
u
m
u
s
e
of
its
v
al
u
e
f
o
r
ti
m
e
r
estricte
d
p
r
o
ce
s
s
es.
iii.
Var
iet
y
-
Or
i
g
in
a
tio
n
of
t
h
e
b
ig
d
ata
is
p
r
i
m
ar
il
y
b
ased
on
t
h
e
d
iv
er
s
it
y
of
s
o
u
r
ce
s
.
T
h
e
C
o
n
v
e
n
tio
n
al
s
y
s
te
m
s
of
d
atab
ases
w
er
e
p
r
o
p
o
s
ed
to
m
ar
k
lo
w
er
ex
te
n
ts
of
clas
s
i
f
ied
d
ata,
s
m
all
er
am
o
u
n
t
u
p
d
ates
or
a
s
tead
y
a
n
d
f
ea
s
ib
le
d
ata
ar
r
an
g
e
m
e
n
t.
Ho
w
ev
er
,
th
e
s
p
atial
d
ata,
3
-
D
d
ata,
au
d
io
-
v
id
eo
,
an
d
th
e
clu
tter
ed
m
an
u
s
cr
ip
t,
co
m
p
r
is
in
g
ac
co
u
n
t
f
i
les
an
d
s
o
cial
m
ed
ia
ar
e
also
co
n
s
id
er
ed
as
b
ig
d
ata.
B
ig
Data
tec
h
n
o
lo
g
y
p
er
m
its
t
h
e
co
llectio
n
a
n
d
p
r
o
ce
s
s
i
n
g
of
lar
g
e
ex
te
n
t
s
of
d
ata,
i
n
cl
u
d
in
g
p
er
s
o
n
al
in
f
o
r
m
atio
n
or
in
f
o
r
m
atio
n
t
h
at
can
r
ec
o
g
n
ize
an
i
n
d
iv
id
u
al.
P
r
esen
tl
y
,
t
h
e
d
ata
h
as
tr
an
s
f
o
r
m
ed
as
an
i
m
p
er
ativ
e
co
n
s
tit
u
e
n
t
th
a
t
co
u
ld
be
an
alo
g
o
u
s
to
r
ea
l
ass
e
ts
an
d
in
d
i
v
id
u
al
r
eso
u
r
ce
s
.
Gen
er
all
y
,
t
h
er
e
ar
e
f
i
v
e
cu
s
to
m
w
a
y
s
t
h
r
o
u
g
h
w
h
ich
t
h
e
b
ig
d
ata
ca
n
be
u
s
ed
[
8
]
.
First,
it
can
cr
ea
te
i
n
f
o
r
m
atio
n
m
o
r
e
cr
y
s
tal
clea
r
an
d
r
ap
id
l
y
.
Seco
n
d
,
t
h
e
estab
lis
h
m
e
n
t
s
ca
n
a
s
s
e
m
b
le
a
n
d
ex
a
m
i
n
e
f
u
r
th
er
d
i
g
ital
d
at
a,
p
r
ec
is
el
y
.
T
h
ir
d
,
th
e
u
tili
za
tio
n
of
s
u
c
h
d
ata
can
g
en
er
ate
m
u
c
h
m
o
r
e
ac
c
u
r
atel
y
p
er
s
o
n
alize
d
g
o
o
d
s
or
f
ac
ilit
ies
f
o
r
co
n
s
u
m
er
s
.
Fo
u
r
th
,
p
o
o
led
w
ith
th
e
p
r
ec
is
e
an
al
y
tic
s
a
n
d
Data
Di
s
cip
li
n
e,
th
e
p
r
o
ce
s
s
of
d
ec
i
s
io
n
-
m
ak
in
g
co
n
s
id
er
ab
l
y
tu
r
n
s
i
n
to
m
o
r
e
p
r
o
f
icie
n
t.
F
if
th
,
it
ca
n
be
u
tili
ze
d
to
m
e
n
d
t
h
e
s
u
cc
ee
d
in
g
g
e
n
er
atio
n
of
a
m
en
ities
an
d
y
ield
s
f
o
r
a
co
m
p
an
y
’
s
clie
n
t
b
ase.
C
u
r
r
en
tl
y
,
b
ig
d
ata
h
a
s
b
ee
n
u
til
ized
in
p
r
ac
ticall
y
e
v
er
y
s
in
g
l
e
f
ield
[
9
]
.
So
m
e
of
th
e
f
ield
s
th
a
t
ar
e
co
n
s
u
m
i
n
g
b
ig
d
ata
s
er
v
ices
ar
e
d
ef
i
n
ed
b
elo
w
:
i.
R
etail:
T
h
e
f
o
r
e
m
o
s
t
tas
k
of
b
u
s
i
n
es
s
in
d
u
s
tr
y
is
b
u
ild
in
g
cli
en
t
r
elatio
n
s
h
ip
w
it
h
t
h
e
as
s
o
ciatio
n
s
or
o
r
g
an
izatio
n
s
.
T
h
e
o
p
ti
m
u
m
w
a
y
to
g
r
asp
a
n
d
d
o
m
i
n
ate
clien
t
s
is
co
n
d
u
ct
d
ea
li
n
g
s
a
n
d
tactic
s
ef
f
icien
tl
y
in
o
r
d
er
to
p
r
o
cu
r
e
b
ac
k
th
e
u
n
s
u
cc
e
s
s
f
u
l
g
o
o
d
s
an
d
p
r
o
g
r
ess
io
n
of
t
h
e
p
r
e
m
i
u
m
g
o
o
d
s
.
ii.
Ma
n
u
f
ac
t
u
r
i
n
g
:
T
h
e
co
m
p
a
n
i
es
can
i
m
p
r
o
v
e
t
h
e
s
u
p
er
io
r
it
y
an
d
e
f
f
icie
n
c
y
of
t
h
e
m
a
n
u
f
ac
tu
r
ed
g
o
o
d
s
by
m
i
n
i
m
izi
n
g
t
h
e
lef
t
o
v
er
w
it
h
t
h
e
a
w
ar
e
n
es
s
in
f
o
r
m
atio
n
d
eli
v
er
ed
by
b
ig
d
at
a.
Sev
er
al
co
m
p
a
n
i
es
ar
e
p
r
esen
tl
y
p
r
o
v
i
d
in
g
s
tr
ess
to
a
n
al
y
tics
-
b
ased
p
o
licy
f
o
r
r
eso
lv
i
n
g
d
if
f
ic
u
lt
a
n
d
f
le
x
ib
le
d
ec
is
io
n
m
a
k
i
n
g
.
iii.
E
d
u
ca
tio
n
:
E
d
u
ca
tio
n
co
m
p
le
tel
y
e
x
a
m
in
e
s
t
h
e
d
ata
o
cc
u
p
ied
f
r
o
m
t
h
e
s
c
h
o
o
l
f
ac
u
lt
y
a
s
s
o
ciatio
n
can
cr
ea
te
d
o
m
i
n
a
n
t
i
m
p
ac
t
on
o
r
g
a
n
izin
g
e
n
d
an
g
er
ed
lear
n
er
s
a
n
d
o
b
s
er
v
i
n
g
th
e
s
u
f
f
icie
n
t
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
I
C
T
I
SS
N:
2252
-
8776
R
ec
en
t
tr
en
d
s
in
b
ig
d
a
ta
u
s
in
g
h
a
d
o
o
p
(
C
h
etn
a
K
a
u
s
h
a
l)
41
i
m
p
r
o
v
e
m
en
t
of
s
t
u
d
en
t
s
.
A
s
s
ess
m
e
n
t
of
t
h
e
s
tu
d
e
n
t’
s
d
ev
elo
p
m
e
n
t
ca
n
be
m
ad
e
w
it
h
t
h
e
s
c
h
o
o
l
f
ac
u
lt
y
s
y
n
c
h
r
o
n
izatio
n
en
h
an
ce
d
s
y
s
te
m
.
iv
.
Hea
lth
ca
r
e:
Mi
s
ce
lla
n
eo
u
s
p
atien
t
r
ec
o
r
d
s
,
tr
ea
t
m
e
n
t
d
ata
an
d
p
r
o
ce
s
s
es
f
o
r
th
er
ap
y
ar
e
ac
co
m
p
li
s
h
ed
ef
f
ec
tiv
e
l
y
w
i
th
th
e
a
w
ar
en
es
s
of
i
n
f
o
r
m
atio
n
;
h
ea
lt
h
ca
r
e
p
r
o
v
id
er
s
can
co
m
p
r
e
h
e
n
d
an
d
p
r
o
f
icien
tl
y
r
ec
o
v
er
p
atien
t
’
s
f
it
n
ess
.
v.
Me
d
ia/en
ter
tai
n
m
e
n
t:
Fro
m
t
h
e
p
ast
f
iv
e
y
ea
r
s
,
t
h
e
i
n
d
u
s
tr
y
of
s
o
cial
m
ed
ia/en
ter
tai
n
m
e
n
t
h
a
s
s
h
i
f
t
ed
to
t
h
e
d
ig
ita
l
m
ea
n
s
of
p
r
o
d
u
ctio
n
,
r
ec
o
r
d
in
g
,
an
d
cir
cu
latio
n
is
cu
r
r
e
n
tl
y
ac
c
u
m
u
lat
in
g
en
o
r
m
o
u
s
a
m
o
u
n
t
s
of
u
s
er
s
o
b
s
er
v
i
n
g
ac
tio
n
s
a
n
d
th
e
r
ich
co
n
ten
t.
v
i.
L
i
f
e
s
c
ien
ce
s
:
Nea
r
l
y
to
n
n
e
s
of
in
f
o
r
m
a
tio
n
(
m
ea
s
u
r
ed
in
t
er
r
a
-
b
y
tes)
ar
e
p
r
o
d
u
ce
d
by
l
ess
er
p
r
ice
DN
A
s
eq
u
e
n
ci
n
g
w
h
ic
h
is
r
eq
u
ir
ed
to
be
ex
a
m
i
n
ed
in
o
r
d
er
to
s
ca
n
t
h
e
h
er
ed
itar
y
m
o
d
i
f
ic
atio
n
s
a
n
d
p
o
s
s
ib
le
p
r
o
f
icien
c
y
of
th
e
c
u
r
e.
v
ii.
Vid
eo
s
u
r
v
eilla
n
ce
:
Vid
eo
s
u
r
v
eilla
n
ce
is
d
ev
elo
p
in
g
f
r
o
m
C
C
T
V
to
w
ar
d
I
PT
V
r
ec
o
r
d
in
g
s
y
s
te
m
s
an
d
ca
p
tu
r
i
n
g
d
e
v
ices
li
k
e
ca
m
er
as
t
h
at
ar
e
u
s
ed
by
t
h
e
o
r
g
an
izat
io
n
s
as
p
er
t
h
e
n
ee
d
to
an
al
y
s
e
p
atter
n
s
of
ac
ti
v
itie
s
an
d
ac
tio
n
s
(
en
h
a
n
ce
m
e
n
t
of
s
er
v
ice
a
n
d
s
ec
u
r
it
y
)
.
v
iii.
T
r
an
s
p
o
r
tatio
n
,
u
tili
tie
s
,
s
er
v
ices,
telec
o
m
m
u
n
icatio
n
a
n
d
lo
g
is
tics
:
At
h
i
g
h
r
ate
s
en
s
o
r
d
ata
is
g
en
er
ated
f
r
o
m
t
h
e
GP
S
tr
an
s
ce
iv
er
s
,
s
m
ar
t
m
eter
s
a
n
d
m
o
b
ile
d
ev
ices
(
ce
ll
p
h
o
n
es)
w
h
i
ch
is
t
h
e
n
u
s
ed
f
o
r
o
p
tim
izin
g
th
e
o
p
er
atio
n
s
an
d
f
i
n
d
th
e
r
elatio
n
s
h
i
p
b
etw
ee
n
th
e
d
ata
w
h
ic
h
f
o
r
m
r
ele
v
a
n
t
in
f
o
r
m
atio
n
f
o
r
b
u
s
i
n
es
s
in
te
l
lig
e
n
ce
(
B
I
)
to
m
ak
e
t
h
e
ap
p
r
o
p
r
iate
d
ec
is
io
n
s
f
o
r
d
if
f
er
e
n
t
b
u
s
i
n
ess
o
p
p
o
r
tu
n
itie
s
.
2.
DATA
M
I
NIN
G
B
ig
d
ata
on
clo
u
d
co
n
tain
s
all
th
e
r
a
w
d
ata
w
h
ic
h
is
g
a
th
er
ed
in
cl
u
s
ter
s
on
t
h
e
b
as
is
of
t
h
eir
r
elatio
n
s
h
ip
.
B
u
t
t
h
e
u
s
er
or
o
r
g
an
izat
io
n
n
ev
er
w
a
n
ted
to
w
a
s
te
t
h
eir
ti
m
e
in
g
ath
er
i
n
g
th
e
d
ata
d
etails
an
d
cr
ea
tin
g
s
tr
u
c
tu
r
al
i
n
f
o
r
m
atio
n
as
it
tak
e
s
a
lo
t
of
t
i
m
e.
He
n
c
e,
Data
Min
i
n
g
is
r
e
f
er
r
ed
as
tak
in
g
out
in
f
o
f
r
o
m
v
ast
g
r
o
u
p
s
of
r
ec
o
r
d
s
of
d
ata.
In
o
th
er
w
a
y
,
t
h
e
p
r
o
ce
s
s
of
d
ata
m
i
n
in
g
is
to
m
i
n
e
k
n
o
w
led
g
e
f
r
o
m
t
h
e
d
atab
ase
[
1
0
]
.
T
h
er
e
is
a
v
a
s
t
q
u
an
tit
y
of
d
ata
ex
i
s
ti
n
g
in
IT
I
n
d
u
s
tr
y
.
S
u
c
h
d
ata
ca
n
n
o
t
be
u
tili
ze
d
f
u
r
th
er
f
o
r
p
r
o
ce
s
s
in
g
,
u
n
les
s
th
at
d
ata
is
tr
an
s
f
o
r
m
ed
in
to
v
al
u
ab
le
in
f
o
.
It
is
in
d
is
p
en
s
ab
le
to
an
al
y
ze
en
o
r
m
o
u
s
v
o
l
u
m
e
of
d
ata
an
d
m
i
n
e
th
e
v
al
u
ab
l
e
in
f
o
r
m
atio
n
f
r
o
m
t
h
e
d
ata.
Min
i
n
g
of
t
h
e
i
n
f
o
r
m
atio
n
is
n
o
t
o
n
l
y
p
r
o
ce
d
u
r
e
w
h
ic
h
is
p
ar
tic
u
lar
l
y
r
eq
u
ir
ed
to
be
p
er
f
o
r
m
ed
;
t
h
er
e
ar
e
als
o
o
th
er
p
r
o
ce
s
s
es
th
a
t
ar
e
in
v
o
lv
ed
in
d
ata
m
i
n
i
n
g
lik
e
Data
C
lea
n
i
n
g
,
Data
Sele
ctio
n
,
Data
I
n
te
g
r
atio
n
,
Da
ta
T
r
an
s
f
o
r
m
atio
n
,
Da
ta
Min
in
g
,
Data
P
r
esen
tatio
n
an
d
P
atter
n
E
v
al
u
atio
n
is
d
es
cr
ib
ed
in
F
i
g
u
r
e
2
[
1
1
]
.
On
c
e
all
th
e
s
e
j
o
b
s
ar
e
co
m
p
lete
l
y
ter
m
i
n
ated
,
th
i
s
in
f
o
r
m
atio
n
co
u
ld
be
ad
ap
ted
f
u
r
th
er
in
v
ar
io
u
s
ap
p
licatio
n
s
as
Fra
u
d
E
x
p
o
s
u
r
e,
Ma
r
k
e
t
A
n
al
y
s
is
,
Scien
ce
E
x
p
lo
r
atio
n
an
d
C
o
n
tr
o
l
in
P
r
o
d
u
ctio
n
etc.
[
1
2
].
Fig
u
r
e
2
.
Data
m
i
n
in
g
in
Kn
o
w
led
g
e
Dis
co
v
er
y
p
r
o
ce
s
s
[
1
3
]
Data
m
i
n
in
g
,
o
f
te
n
r
ef
er
r
ed
to
Kn
o
w
led
g
e
d
is
co
v
er
y
(
KD
D)
in
v
o
l
v
es
m
in
in
g
of
i
n
f
o
r
m
atio
n
or
k
n
o
w
led
g
e
as
it
s
p
r
i
m
ar
y
a
n
d
th
e
u
t
m
o
s
t
c
h
alle
n
g
i
n
g
a
n
d
in
tr
i
g
u
in
g
s
tep
[
1
3
]
.
No
r
m
all
y
,
d
ata
m
i
n
i
n
g
d
is
clo
s
es
t
h
e
in
tr
i
g
u
in
g
p
att
er
n
s
a
n
d
i
n
f
er
e
n
ce
s
th
a
t
ar
e
co
n
ce
aled
co
v
er
tl
y
i
n
s
id
e
a
lar
g
e
v
o
lu
m
e
of
u
n
a
n
al
y
s
ed
or
p
r
i
m
ar
y
d
ata,
an
d
th
e
o
u
tco
m
e
s
w
h
ic
h
ar
e
ca
r
r
ied
o
u
t
m
a
y
p
o
s
s
ib
l
y
s
u
p
p
o
r
t
f
u
tu
r
e
o
b
s
er
v
at
io
n
s
in
th
e
ac
t
u
al
w
o
r
ld
.
Data
m
in
in
g
h
a
s
b
ee
n
ex
p
lo
ited
by
an
ex
te
n
s
i
v
e
v
ar
iet
y
of
ap
p
licatio
n
s
i.e
.
b
u
s
in
e
s
s
,
d
r
u
g
,
s
cien
c
e
an
d
en
g
i
n
ee
r
i
n
g
.
A
lt
h
o
u
g
h
,
th
e
d
ata
m
i
n
in
g
is
m
ain
p
h
a
s
e
in
k
n
o
w
led
g
e
d
is
co
v
er
y
p
r
o
ce
s
s
th
er
ef
o
r
e
it
is
also
u
s
ed
as
a
s
u
b
s
tit
u
te
f
o
r
en
tire
p
r
o
ce
s
s
of
tak
i
n
g
o
u
t
u
s
ef
u
l
i
n
f
o
f
r
o
m
d
atab
ases
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8776
IJ
-
I
C
T
Vo
l.
8
,
No
.
1
,
A
p
r
il 2
0
1
9
:
39
–
49
42
B
u
t
s
ti
ll
in
tec
h
n
ical
e
n
v
ir
o
n
m
en
t
(
in
d
u
s
tr
y
)
,
in
th
e
d
atab
ase
r
esear
ch
a
n
d
in
m
ed
ia,
d
at
a
m
i
n
i
n
g
is
b
ec
o
m
i
n
g
ex
tr
ap
r
ev
ale
n
t.
T
h
e
r
e
ar
e
n
u
m
b
er
of
s
tep
s
i
n
v
o
lv
ed
in
th
e
e
n
tire
p
r
o
ce
s
s
of
k
n
o
w
led
g
e
d
i
s
co
v
er
y
f
r
o
m
d
atab
ases
w
h
ic
h
is
d
escr
ib
ed
in
F
ig
u
r
e
2
[
1
4
]
.
T
h
e
f
ig
u
r
e
r
ep
r
esen
ts
th
e
s
eq
u
en
ce
of
th
e
in
d
iv
id
u
al
s
tep
w
it
h
i
n
t
h
e
p
r
o
ce
s
s
an
d
is
b
r
ief
l
y
d
escr
ib
ed
in
th
e
te
x
t
b
elo
w
:
−
Data
I
n
teg
r
atio
n
−
th
e
m
u
ltip
l
e
d
ata
f
r
o
m
d
i
s
ti
n
ct
s
o
u
r
ce
s
ar
e
in
itiall
y
j
o
in
ed
.
−
Selectio
n
of
d
ata
–
th
e
ap
p
r
o
p
r
iate
d
ata
as
p
er
th
e
an
al
y
s
i
s
of
t
h
e
tas
k
is
s
elec
ted
f
r
o
m
m
u
l
tip
le
d
ata
s
o
u
r
ce
s
.
−
Data
P
r
e
-
P
r
o
ce
s
s
in
g
−
t
h
e
n
o
i
s
e
an
d
in
co
n
s
i
s
ten
c
y
of
t
h
e
d
at
a
is
eli
m
in
ated
.
−
Data
T
r
an
s
f
o
r
m
at
io
n
−
d
ata
is
co
n
v
er
ted
or
m
er
g
ed
i
n
to
s
u
ch
f
o
r
m
s
t
h
at
ar
e
ap
p
licab
le
f
o
r
m
in
i
n
g
by
ca
r
r
y
i
n
g
o
u
t
s
u
m
m
ar
y
or
b
len
d
ed
p
r
o
ce
d
u
r
es.
−
Data
Min
i
n
g
–
in
tellect
u
al
p
r
o
ce
d
u
r
es
ar
e
ad
ap
ted
to
a
b
s
tr
ac
t
th
e
p
atter
n
s
of
d
ata.
−
E
v
alu
a
tio
n
of
P
atter
n
−
e
v
alu
a
t
io
n
s
of
d
ata
p
atter
n
s
w
h
ic
h
ar
e
ab
s
tr
ac
ted
.
−
Kn
o
w
led
g
e
P
r
esen
ta
tio
n
–
In
C
o
n
cl
u
s
io
n
,
k
n
o
w
led
g
e
is
r
ep
r
esen
ted
.
2
.
1
.
Cha
lleng
es
in
B
ig
Da
t
a
M
ini
ng
T
h
e
f
o
r
em
o
s
t
ch
alle
n
g
e
s
th
a
t
ar
is
e
in
b
ig
d
ata
m
i
n
i
n
g
ar
e
b
r
ief
l
y
d
e
f
in
ed
in
th
e
f
o
llo
w
in
g
p
o
in
ts
at
T
ab
le
1
[
1
5
]
:
T
ab
le
1.
C
h
allen
g
es
in
t
h
e
B
ig
Data
Min
i
n
g
C
h
a
l
l
e
n
g
e
s
D
e
scri
p
t
i
o
n
S
h
i
e
l
d
i
n
g
p
r
i
v
a
c
y
a
n
d
c
o
n
f
i
d
e
n
t
i
a
l
i
t
y
P
r
i
me
f
o
c
u
s
on
g
e
n
e
r
a
t
i
n
g
t
h
e
t
e
c
h
n
i
q
u
e
s
t
h
a
t
w
i
l
l
n
e
v
e
r
d
i
scl
o
se
t
h
e
d
e
si
g
n
s
a
n
d
a
l
so
e
n
su
r
e
safe
t
y
a
n
d
p
r
i
v
a
c
y
M
a
n
a
g
i
n
g
t
h
e
i
n
a
d
e
q
u
a
t
e
i
n
f
o
r
mat
i
o
n
A
b
se
n
t
v
a
l
u
e
s
t
h
a
t
r
e
l
a
t
e
s
to
d
e
f
i
c
i
e
n
c
y
of
f
e
a
t
u
r
e
s,
is
a
r
g
u
e
d
c
o
mp
r
e
h
e
n
s
i
v
e
l
y
f
o
r
o
f
f
l
i
n
e
,
st
a
t
i
c
se
t
t
i
n
g
s
U
n
d
e
f
i
n
e
d
d
a
t
a
M
o
st
a
p
p
l
i
c
a
t
i
o
n
s
do
n
o
t
p
o
sse
ss
su
f
f
i
c
i
e
n
t
d
a
t
a
f
o
r
a
r
i
t
h
me
t
i
c
p
r
o
c
e
d
u
r
e
s.
H
e
n
c
e
a
p
p
r
o
a
c
h
e
s
a
r
e
r
e
q
u
i
r
e
d
to
h
a
n
d
l
e
u
n
d
e
f
i
n
e
d
d
a
t
a
v
a
l
u
e
s
in
a
p
r
e
c
i
se
a
n
d
q
u
i
c
k
w
a
y
.
D
i
v
e
r
si
t
y
of
d
a
t
a
S
o
c
i
a
l
si
t
e
is
t
h
e
mo
st
c
a
p
t
i
v
a
t
i
n
g
i
mm
i
n
e
n
t
a
p
p
l
i
c
a
t
i
o
n
of
d
a
t
a
st
r
e
a
m
c
l
u
st
e
r
i
n
g
l
i
k
e
v
i
d
e
o
,
i
m
a
g
e
s,
t
e
x
t
a
n
d
a
u
d
i
o
.
S
y
n
o
p
si
s
a
n
d
su
m
marie
s
S
y
n
o
p
si
s
r
e
f
e
r
s
to
c
o
mp
r
e
sse
d
st
a
t
i
st
i
c
s
a
r
r
a
n
g
e
me
n
t
s
w
h
i
c
h
l
e
t
d
a
t
a
su
mm
a
r
i
z
a
t
i
o
n
f
o
r
a
d
v
a
n
c
e
q
u
e
st
i
o
n
i
n
g
l
i
k
e
t
h
e
h
i
s
t
o
g
r
a
ms,
w
a
v
e
l
e
t
s
f
o
r
ms
a
n
d
sam
p
l
e
s
d
e
f
i
n
e
t
h
e
e
n
o
r
mo
u
s
i
n
f
o
r
mat
i
o
n
in
t
h
e
c
o
mp
r
e
sse
d
w
a
y
.
D
i
st
r
i
b
u
t
e
d
s
t
r
e
a
ms
In
a
p
p
l
i
c
a
t
i
o
n
s
s
u
c
h
as
c
e
n
t
r
a
l
i
z
e
d
r
e
su
l
t
s
b
r
i
n
g
t
o
g
e
t
h
e
r
i
n
t
e
r
r
u
p
t
i
o
n
s
in
e
v
e
n
t
r
e
c
o
g
n
i
t
i
o
n
a
n
d
r
e
sp
o
n
se
t
h
a
t
can
c
r
e
a
te
m
i
n
i
n
g
sy
st
e
ms
u
n
s
u
c
c
e
ssf
u
l
Ev
a
l
u
a
t
i
o
n
of
d
a
t
a
st
r
e
a
m
p
r
o
c
e
d
u
r
e
s
Ex
i
st
i
n
g
t
o
o
l
s
su
c
h
as
i
d
e
a
i
m
p
l
i
c
a
t
i
o
n
,
r
e
st
r
i
c
t
e
d
p
r
o
c
e
ssi
n
g
i
n
t
e
r
v
a
l
,
a
u
t
h
e
n
t
i
c
a
t
i
o
n
d
o
r
man
c
y
,
mu
l
t
i
p
l
e
st
r
e
a
m
st
r
u
c
t
u
r
e
s
a
r
e
i
n
a
d
e
q
u
a
t
e
in
t
h
e
d
a
t
a
s
t
r
e
a
m
d
a
t
a
b
a
se
s
d
u
e
to
c
e
r
t
a
i
n
p
r
o
b
l
e
ms
I
n
d
e
p
e
n
d
e
n
t
a
n
d
se
l
f
-
d
i
a
g
n
o
si
s
K
n
o
w
l
e
d
g
e
d
i
sco
v
e
r
y
in
d
a
t
a
b
a
se
s
n
e
e
d
t
h
e
s
k
i
l
l
s
f
o
r
p
r
o
g
n
o
st
i
c
se
l
f
-
d
i
a
g
n
o
si
s.
A
me
a
n
i
n
g
f
u
l
a
n
d
b
e
n
e
f
i
c
i
a
l
i
n
t
e
l
l
e
c
t
u
a
l
f
e
a
t
u
r
e
is
d
i
a
g
n
o
st
i
c
s
in
si
t
u
a
t
i
o
n
of
f
a
i
l
u
r
e
o
c
c
u
r
r
e
n
c
e
a
n
d
a
l
so
p
r
o
g
n
o
st
i
c
a
n
d
c
o
n
s
u
l
t
a
t
i
v
e
.
T
h
e
e
v
o
l
u
t
i
o
n
of
t
h
e
se
t
y
p
e
s
of
se
l
f
-
o
r
g
a
n
i
z
i
n
g
,
se
l
f
-
e
n
h
a
n
c
i
n
g
,
a
n
d
se
l
f
-
r
e
st
o
r
i
n
g
sy
st
e
ms
is
f
o
r
e
mo
st
c
h
a
l
l
e
n
g
e
.
M
e
r
g
i
n
g
o
f
f
l
i
n
e
a
n
d
o
n
l
i
n
e
mo
d
e
l
s
R
e
a
l
-
t
i
me
a
n
d
b
a
t
c
h
l
e
a
r
n
i
n
g
a
r
e
f
r
e
q
u
e
n
t
l
y
c
o
n
si
d
e
r
e
d
as
se
p
a
r
a
t
e
d
i
d
e
n
t
i
t
i
e
s
a
c
c
o
r
d
i
n
g
to
t
h
e
i
r
a
c
t
i
o
n
,
but
t
h
e
i
r
g
r
o
u
p
i
n
g
mi
g
h
t
b
o
o
st
t
h
e
d
a
t
a
v
a
l
u
e
.
In
l
a
mb
d
a
f
r
a
me
w
o
r
k
t
h
e
t
w
o
mo
d
e
l
s
can
be
c
o
mb
i
n
e
d
f
o
r
p
l
a
n
n
i
n
g
b
i
g
d
a
t
a
mo
d
e
l
s.
3.
H
ADO
O
P
Had
o
o
p
is
an
o
p
en
-
s
o
u
r
ce
f
r
a
m
e
w
o
r
k
w
h
ic
h
p
er
m
its
to
ac
c
u
m
u
late
a
n
d
r
u
n
b
ig
d
ata
in
a
d
is
tr
ib
u
ted
ar
r
an
g
e
m
en
t
in
t
h
e
n
et
w
o
r
k
of
co
m
p
u
ter
s
co
n
s
u
m
in
g
m
o
d
est
p
r
o
g
r
am
m
i
n
g
m
o
d
els.
T
h
is
w
h
o
le
p
r
o
ce
s
s
s
ca
les
up
f
r
o
m
s
o
litar
y
s
er
v
e
r
s
to
th
o
u
s
a
n
d
s
of
m
ac
h
i
n
es,
co
llectiv
el
y
p
u
t
f
o
r
w
ar
d
lo
ca
l
m
an
ip
u
latio
n
an
d
s
to
r
in
g
.
Had
o
o
p
ex
ec
u
tes
th
e
ap
p
licatio
n
s
v
ia
Ma
p
R
ed
u
ce
alg
o
r
ith
m
,
w
h
er
e
on
d
i
v
er
s
e
C
P
U
n
o
d
es;
i
n
f
o
is
s
o
r
t
o
u
t
in
p
ar
allel.
In
a
n
u
ts
h
ell,
Had
o
o
p
f
r
a
m
e
w
o
r
k
is
p
r
o
f
icien
t
to
en
co
u
r
a
g
e
ap
p
licatio
n
s
t
h
at
ar
e
q
u
alif
ied
of
ex
ec
u
tin
g
on
t
h
e
g
r
o
u
p
of
m
ac
h
i
n
es
an
d
all
co
u
ld
d
eliv
er
f
u
ll
y
s
tati
s
tical
i
n
ter
p
r
etatio
n
f
o
r
i
m
m
e
n
s
e
v
o
lu
m
es
of
d
ata
[
1
6
]
.
T
h
e
ap
p
licatio
n
w
h
ich
is
d
ep
en
d
en
t
on
Had
o
o
p
f
r
a
m
e
w
o
r
k
r
u
n
s
in
an
e
n
v
ir
o
n
m
e
n
t
wh
ich
g
iv
e
s
d
is
tr
ib
u
ted
s
to
r
ag
e
an
d
co
m
p
u
tatio
n
s
on
t
h
e
g
r
o
u
p
of
m
ac
h
i
n
es
in
t
h
e
n
et
w
o
r
k
.
E
x
ten
s
io
n
of
Had
o
o
p
co
u
ld
be
n
u
m
er
o
u
s
s
er
v
er
s
,
each
g
iv
i
n
g
th
e
n
ati
v
e
co
m
p
u
tatio
n
an
d
s
t
o
r
ag
e
s
er
v
ice.
In
F
ig
u
r
e
3,
Had
o
o
p
A
r
ch
itect
u
r
e
is
d
ef
i
n
ed
w
h
ic
h
p
r
i
m
ar
il
y
in
clu
d
e
s
s
u
b
s
eq
u
e
n
t
m
o
d
u
les
[
1
7
]
:
i.
Had
o
o
p
C
o
m
m
o
n
:
Had
o
o
p
co
m
m
o
n
co
n
s
is
t
of
lib
r
ar
ies
of
J
av
a
a
n
d
s
er
v
ices
n
ee
d
ed
by
o
t
h
er
Had
o
o
p
elem
e
n
ts
.
T
h
ese
lib
r
ar
ies
o
f
f
er
OS
lev
e
l
ab
s
tr
ac
tio
n
s
,
f
ile
s
s
y
s
te
m
an
d
co
m
p
r
i
s
e
s
es
s
en
t
ial
J
av
a
lib
r
ar
ies
an
d
s
cr
ip
ts
r
eq
u
i
r
ed
to
in
itialize
Had
o
o
p
.
ii.
Had
o
o
p
YA
R
N:
It
is
k
in
d
of
s
tr
u
ctu
r
e
f
o
r
s
ch
ed
u
lin
g
of
j
o
b
an
d
clu
s
ter
r
eso
u
r
ce
m
a
n
ag
in
g
.
iii.
Had
o
o
p
Dis
tr
ib
u
ted
File
S
y
s
te
m
(
HD
FS
)
:
It
is
f
i
le
ar
ch
i
tectu
r
e
th
at
o
f
f
er
s
r
i
g
h
t
to
u
s
e
th
e
a
p
p
licatio
n
d
ata.
iv
.
Had
o
o
p
Ma
p
R
ed
u
ce
:
T
h
is
is
a
s
y
s
te
m
b
ased
on
Y
AR
N
f
o
r
p
ar
allel
p
r
o
ce
s
s
in
g
of
b
ig
s
et
s
of
d
ata.
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
I
C
T
I
SS
N:
2252
-
8776
R
ec
en
t
tr
en
d
s
in
b
ig
d
a
ta
u
s
in
g
h
a
d
o
o
p
(
C
h
etn
a
K
a
u
s
h
a
l)
43
Sin
ce
2012,
th
e
co
n
ce
p
t
"
Had
o
o
p
"
r
e
p
ea
ted
ly
m
e
n
tio
n
s
to
be
th
e
b
ase
u
n
it
s
a
n
d
to
th
e
v
ar
iet
y
of
o
th
er
s
o
f
t
w
ar
e
s
et
s
t
h
at
can
be
m
o
u
n
ted
b
esid
e
Had
o
o
p
,
lik
e
A
p
ac
h
e
Hiv
e,
A
p
ac
h
e
P
ig
etc.
Fig
u
r
e
3.
A
r
ch
itectu
r
e
of
Had
o
o
p
[
1
6
]
3
.
1
.
H
a
do
o
p
Dis
t
ribute
d
F
il
e
Sy
s
t
e
m
(
H
DF
S)
T
h
e
Had
o
o
p
Dis
tr
ib
u
ted
File
S
y
s
te
m
(
HDF
S)
is
ce
n
tr
ed
on
th
e
Go
o
g
le
File
S
y
s
te
m
(
G
FS
)
an
d
o
f
f
er
s
f
ile
s
y
s
te
m
w
h
ich
is
d
is
tr
ib
u
ted
in
n
at
u
r
e
t
h
at
is
in
te
n
d
ed
to
ex
ec
u
te
on
lar
g
e
g
r
o
u
p
of
co
m
p
u
ter
in
th
e
n
et
w
o
r
k
in
a
co
n
s
i
s
ten
t
a
n
d
er
r
o
r
-
r
ec
ep
tiv
e
m
a
n
n
er
[
1
8
]
.
In
co
n
tr
a
s
t
to
t
h
e
ad
d
itio
n
al
d
is
tr
ib
u
ted
s
y
s
te
m
s
,
HDFS
is
ex
ce
ed
i
n
g
l
y
f
a
u
lt
i
n
d
u
lg
e
n
t
a
n
d
d
esi
g
n
ed
w
it
h
lo
w
co
s
t
h
ar
d
w
ar
e.
HDF
S
g
r
asp
s
v
er
y
h
u
g
e
a
m
o
u
n
t
of
d
ata
a
n
d
o
f
f
er
s
co
m
f
o
r
tab
le
ac
ce
s
s
.
T
h
e
f
ile
s
ar
e
s
to
r
ed
ac
r
o
s
s
s
e
v
er
al
m
ac
h
in
e
s
in
o
r
d
er
to
p
r
o
tect
s
u
c
h
a
h
u
g
e
d
ata.
T
h
ese
f
ile
s
ar
e
k
ep
t
in
a
r
ep
etitiv
e
m
an
n
er
to
r
ec
o
v
er
d
ata
lo
s
s
es
in
th
e
s
y
s
te
m
in
ev
e
n
ts
of
f
ai
lu
r
e.
HDFS
p
r
i
m
ar
il
y
ad
ap
ts
t
h
e
m
aster
/
s
la
v
e
d
esi
g
n
.
In
th
i
s
d
esig
n
t
h
e
m
aster
co
m
p
r
is
es
a
s
i
n
g
le
Na
m
eNo
d
e
th
at
m
a
in
ta
in
s
th
e
m
e
tad
ata
an
d
s
lav
e
co
m
p
r
is
e
s
m
u
ltip
le
Data
No
d
es
th
at
p
r
eser
v
e
th
e
o
r
ig
i
n
al
d
ata.
In
t
h
e
F
ig
u
r
e
4,
ar
ch
i
tectu
r
e
of
H
DFS
is
s
h
o
w
n
an
d
is
d
i
v
id
ed
in
to
d
ata
n
o
d
es.
A
f
i
le
in
r
ef
er
r
ed
as
HDFS
n
a
m
esp
ac
e
is
d
i
v
id
ed
in
to
a
n
u
m
b
er
of
b
lo
ck
s
.
T
h
ese
in
d
iv
id
u
al
b
lo
ck
s
ar
e
k
ep
t
in
a
clas
s
of
Data
No
d
es.
[
1
9
]
.
T
h
e
Data
No
d
es
ar
e
r
esp
o
n
s
ib
le
f
o
r
t
h
e
r
ea
d
in
g
a
n
d
w
r
iti
n
g
p
r
o
ce
d
u
r
e
of
th
e
f
ile
s
y
s
te
m
.
T
h
ey
f
u
r
t
h
er
ar
e
r
esp
o
n
s
ib
le
f
o
r
th
e
b
lo
ck
f
o
r
m
u
lat
io
n
,
ter
m
i
n
atio
n
,
a
n
d
d
u
p
licatio
n
as
p
er
th
e
in
s
tr
u
ctio
n
s
p
r
o
v
id
ed
by
Na
m
e
No
d
e.
T
h
e
HDFS
r
en
d
er
s
a
s
h
el
l
s
i
m
ilar
l
y
to
m
a
n
y
o
t
h
er
f
ile
s
y
s
te
m
(
m
eta
d
ata)
an
d
a
lis
t
of
in
s
tr
u
ctio
n
s
ar
e
p
r
ep
ar
ed
to
co
m
m
u
n
icate
to
th
e
f
ile
s
y
s
te
m
.
Fig
u
r
e
4.
A
r
ch
itectu
r
e
of
H
DF
S
[
1
8
]
4.
C
L
US
T
E
RIN
G
A
L
G
O
RI
T
H
M
C
lu
s
ter
i
n
g
is
th
e
j
o
b
of
d
iv
is
i
o
n
of
th
e
p
o
p
u
latio
n
/d
ata
p
o
in
ts
in
g
r
o
u
p
n
u
m
b
er
lik
e
d
at
a
p
o
in
ts
in
s
i
m
ilar
g
r
o
u
p
t
h
at
ar
e
s
a
m
e
as
an
o
th
er
d
ata
p
o
in
t
s
in
s
i
m
ilar
g
r
o
u
p
as
co
m
p
ar
ed
to
th
e
a
n
o
th
er
g
r
o
u
p
s
.
It
ca
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8776
IJ
-
I
C
T
Vo
l.
8
,
No
.
1
,
A
p
r
il 2
0
1
9
:
39
–
49
44
also
be
s
aid
th
at
t
h
e
o
b
j
ec
tiv
e
of
cl
u
s
ter
i
n
g
is
to
s
ep
ar
ate
th
e
g
r
o
u
p
in
s
a
m
e
tr
aits
an
d
allo
ca
te
t
h
e
m
in
cl
u
s
ter
f
o
r
m
.
C
l
u
s
ter
i
n
g
ca
n
be
cla
s
s
i
f
ied
in
to
t
w
o
al
g
o
r
i
th
m
s
,
n
a
m
el
y
,
K
-
m
ea
n
a
n
d
K
-
m
ed
o
id
s
cl
u
s
ter
i
n
g
.
T
h
e
ex
p
lan
atio
n
of
t
h
e
s
a
m
e
is
g
iv
e
n
.
4
.
1
.
K
-
m
ea
ns
a
lg
o
rit
h
m
T
h
e
m
o
s
t
g
e
n
er
al
alg
o
r
it
h
m
u
tili
ze
s
an
iter
at
iv
e
r
e
f
in
e
m
e
n
t
m
eth
o
d
.
B
ec
au
s
e
of
i
ts
u
b
iq
u
it
y
,
it
f
r
eq
u
en
tl
y
ter
m
ed
as
K
-
m
ea
n
s
alg
o
r
ith
m
or
as
L
lo
y
d
’
s
al
g
o
r
ith
m
.
C
o
n
s
id
er
K
-
m
ea
n
in
i
tial
s
et/
C
en
tr
o
id
s
,
t
h
e
alg
o
r
ith
m
h
as
b
ee
n
d
iv
id
ed
in
t
o
t
w
o
s
tep
s
[
2
0
]
:
Ass
i
g
n
m
e
n
t
Step
:
Ass
i
g
n
in
g
ev
er
y
o
b
s
er
v
atio
n
to
th
e
cl
u
s
ter
h
av
i
n
g
th
e
clo
s
e
s
t
m
ea
n
[
th
at
is
t
h
e
p
ar
titi
o
n
of
th
e
o
b
s
er
v
at
io
n
as
p
er
Vo
r
o
n
o
i
d
iag
r
am
p
r
o
d
u
ce
d
by
t
h
e
m
ea
n
s
.
R
i
(
q
)
=
{
a
j
:
‖
a
j
−
n
i
q
‖
≤
‖
a
j
−
n
i
∗
q
‖
for
all
i
∗
=
1
,
…
,
l
}
Up
d
ate
s
tep
:
C
o
m
p
u
te
t
h
e
n
o
v
el
m
ea
n
to
be
ce
n
tr
o
id
of
th
e
c
lu
s
ter
o
b
s
er
v
atio
n
n
i
(
q
+
1
)
=
1
|
R
i
q
|
∑
a
j
a
j
∈
R
i
(
q
)
T
h
e
K
-
m
ea
n
s
al
g
o
r
ith
m
is
b
eliev
ed
to
be
m
ee
t
w
h
e
n
th
e
a
s
s
i
g
n
m
e
n
t
d
o
esn
’
t
c
h
an
g
e
f
o
r
lo
n
g
[
2
1
]
.
4
.
2
.
K
-
m
edo
id
s
a
lg
o
rit
h
m
K
-
m
ed
o
id
alg
o
r
ith
m
is
ass
o
ci
ated
to
th
e
K
-
m
ea
n
al
g
o
r
ith
m
w
it
h
th
e
m
ed
o
id
s
h
i
f
t
alg
o
r
ith
m
.
T
h
e
K
-
m
ed
o
id
an
d
th
e
K
-
m
ea
n
al
g
o
r
ith
m
k
n
o
w
n
as
P
ar
titi
o
n
al
al
g
o
r
ith
m
s
.
K
-
m
ea
n
les
s
en
s
th
e
to
tal
s
q
u
ar
ed
er
r
o
r
an
d
th
e
K
-
m
ed
o
id
s
r
ed
u
ce
s
t
h
e
am
o
u
n
t
of
d
is
s
i
m
ilar
i
ties
a
m
o
n
g
p
o
in
ts
lab
eled
to
be
in
clu
s
ter
w
i
th
t
h
e
p
o
in
t
s
elec
ted
as
th
e
cl
u
s
ter
ce
n
tr
e
.
K
-
m
ed
o
id
s
s
e
lects
th
e
d
ata
p
o
in
ts
as
t
h
e
ce
n
tr
es
w
it
h
r
esp
ec
t
to
K
-
m
ea
n
alg
o
r
ith
m
.
It
is
a
p
ar
titi
o
n
i
n
g
m
et
h
o
d
f
o
r
cl
u
s
ter
i
n
g
t
h
e
d
at
a
s
ets
of
m
o
b
j
ec
ts
in
k
-
cl
u
s
t
er
s
by
K
ter
m
ed
as
P
r
io
r
i.
T
h
e
ef
f
ec
tiv
e
to
o
l
to
m
ea
s
u
r
e
is
S
ilh
o
u
ette
[
2
2
]
.
It
m
a
y
be
m
o
r
e
v
i
g
o
r
o
u
s
to
n
o
is
e
an
d
t
h
e
o
u
tlier
s
by
m
ea
n
s
of
k
-
m
ea
n
s
as
it
r
ed
u
ce
s
a
a
m
o
u
n
t
of
n
o
r
m
al
p
air
w
i
s
e
d
is
s
i
m
ilar
itie
s
th
a
n
s
q
u
ar
ed
E
u
clid
ea
n
d
is
ta
n
ce
s
u
m
.
T
h
e
m
ed
o
id
by
m
ea
n
s
of
f
in
ite
d
ataset
is
th
e
d
ata
p
o
in
t
f
o
r
m
t
h
e
s
et
h
av
i
n
g
a
v
er
ag
e
d
is
s
i
m
ilar
it
y
to
each
d
ata
p
o
in
t
is
less
m
ea
n
s
it
is
co
n
s
id
er
ed
as
lik
el
y
to
th
e
ce
n
tr
all
y
lo
ca
te
d
p
o
in
t
s
et.
Gen
er
al
r
ea
lizatio
n
of
k
-
m
ed
o
id
clu
s
ter
i
n
g
is
P
A
M
(
P
ar
titi
o
n
in
g
A
r
o
u
n
d
m
ed
o
id
)
alg
o
r
ith
m
a
n
d
is
d
ef
in
ed
b
elo
w
[
2
3
]
:
i.
I
n
itialize:
A
r
b
itra
r
il
y
s
elec
ted
K
of
m
d
ata
p
o
in
ts
as
m
ed
o
id
s
.
ii.
Ass
i
g
n
m
e
n
t
s
tep
:
C
o
n
n
ec
t
ev
e
r
y
d
ata
p
o
in
t
to
t
h
e
clo
s
est
m
e
d
o
id
.
iii.
Up
d
ate
s
tep
:
f
o
r
ev
er
y
m
ed
o
id
an
d
f
o
r
ev
er
y
d
ata
p
o
i
n
t
p,
li
n
k
ed
to
n
s
w
ap
n
an
d
p
an
d
ca
l
cu
late
th
e
to
tal
co
n
f
i
g
u
r
atio
n
co
s
t.
C
h
o
o
s
e
th
e
m
ed
o
id
p
w
ith
le
s
s
co
n
f
ig
u
r
at
io
n
co
s
t.
5.
CL
AS
SI
F
I
CAT
I
O
N
AL
G
O
RIT
H
M
C
las
s
i
f
icatio
n
h
a
s
w
id
e
r
a
n
g
e
of
m
et
h
o
d
s
to
ca
te
g
o
r
ize
th
e
d
ata
in
to
th
e
g
r
o
u
p
of
cl
u
s
te
r
s
.
T
h
er
e
is
u
tter
n
ee
d
of
t
h
e
clas
s
i
f
icatio
n
p
r
o
ce
s
s
as
t
h
e
h
u
g
e
v
o
lu
m
e
of
d
ata
is
ca
teg
o
r
ized
i
n
to
th
e
g
r
o
u
p
b
ased
on
t
h
e
r
elatio
n
b
et
w
ee
n
th
e
d
ata
o
b
j
ec
ts
.
Hen
ce
,
alg
o
r
ith
m
s
ar
e
r
eq
u
ir
ed
w
h
ich
h
as
tr
ain
i
n
g
d
ata
-
s
ets
i
n
b
u
il
t
ac
co
r
d
in
g
to
h
u
m
a
n
p
er
ce
p
tio
n
of
d
ata
c
lass
if
ica
tio
n
.
C
la
s
s
i
f
icatio
n
is
a
t
y
p
ical
d
ata
m
in
i
n
g
m
e
th
o
d
th
a
t
is
d
ep
en
d
en
t
on
m
ac
h
in
e
lear
n
i
n
g
[2
4
].
B
asicall
y
cla
s
s
i
f
icatio
n
is
n
ee
d
ed
to
class
i
f
y
ea
c
h
o
b
ject
in
to
a
p
ar
tic
u
lar
class
.
C
la
s
s
i
f
icat
io
n
is
f
u
r
th
er
d
iv
id
ed
in
to
Su
p
er
v
is
ed
an
d
U
n
s
u
p
er
v
is
ed
cla
s
s
i
f
icatio
n
.
Su
p
er
v
i
s
ed
lear
n
in
g
is
in
w
h
ich
th
e
tr
ai
n
in
g
s
et
of
p
r
ec
is
el
y
r
ec
o
g
n
ized
d
ataset
o
b
s
er
v
atio
n
s
ar
e
ac
ce
s
s
ib
le.
W
h
er
ea
s
,
in
t
h
e
u
n
s
u
p
er
v
is
ed
lear
n
in
g
ta
k
es
th
e
ch
an
ce
i
ts
el
f
by
g
r
o
u
p
in
g
d
ata
on
th
e
b
asi
s
of
s
i
m
ilar
m
ea
s
u
r
es
of
in
h
er
en
t
s
i
m
ilar
it
y
.
T
h
er
e
ar
e
n
u
m
er
o
u
s
m
et
h
o
d
s
in
th
e
s
u
p
er
v
is
ed
lear
n
i
n
g
h
o
w
e
v
er
ac
co
r
d
in
g
to
th
e
p
r
ev
io
u
s
s
t
u
d
ies
KNN
is
t
h
e
b
est
m
e
th
o
d
f
o
r
cl
ass
i
f
icatio
n
in
t
h
e
ca
s
e
of
b
i
g
d
ata
an
d
g
iv
e
b
etter
r
es
u
lts
w
h
en
u
s
ed
.
Her
e
af
ter
,
KNN
alg
o
r
it
h
m
is
d
ef
i
n
ed
an
d
h
o
w
clas
s
i
f
icatio
n
is
d
o
n
e
w
it
h
t
h
e
h
elp
of
KN
N
alg
o
r
it
h
m
is
p
r
esen
ted
in
s
u
b
s
ec
tio
n
[
2
5
]
.
5
.
1
.
K
NN
Alg
o
rit
h
m
T
h
e
K
-
n
ea
r
est
n
ei
g
h
b
o
u
r
p
r
o
c
ed
u
r
e
(
KNN)
is
a
w
a
y
f
o
r
cla
s
s
i
f
icatio
n
of
e
n
titi
e
s
on
t
h
e
b
asis
of
th
e
ad
jo
in
in
g
tr
ain
i
n
g
s
p
ec
i
m
en
s
in
f
ea
t
u
r
e
s
p
ac
e
[2
6
].
T
h
e
p
r
i
m
e
in
ten
t
io
n
of
t
h
e
k
Nea
r
est
Neig
h
b
o
u
r
s
(
KNN)
p
r
o
ce
s
s
is
-
to
u
s
e
t
h
e
d
atab
ase
w
h
er
ein
t
h
e
d
ata
ar
e
d
iv
id
ed
i
n
to
a
n
u
m
b
er
of
is
o
lated
class
es
to
p
r
o
g
n
o
s
ticate
th
e
clas
s
i
f
icatio
n
of
a
n
e
w
s
a
m
p
le
p
o
in
t.
KNN
clas
s
if
icatio
n
d
is
tr
ib
u
tes
t
h
e
d
ata
i
n
to
test
s
et
an
d
tr
ai
n
i
n
g
s
ets.
T
h
en
th
e
K
n
ea
r
est
tr
ain
in
g
s
et
o
b
j
ec
ts
ar
e
o
r
ig
in
ated
f
o
r
ev
er
y
s
i
n
g
le
r
o
w
of
th
e
te
s
t
s
et
,
an
d
th
e
p
r
o
ce
s
s
or
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
I
C
T
I
SS
N:
2252
-
8776
R
ec
en
t
tr
en
d
s
in
b
ig
d
a
ta
u
s
in
g
h
a
d
o
o
p
(
C
h
etn
a
K
a
u
s
h
a
l)
45
task
of
c
lass
if
ica
tio
n
is
p
er
f
o
r
m
ed
by
p
r
ed
o
m
in
a
n
ce
v
o
t
e
w
it
h
co
n
n
ec
tio
n
s
w
h
ic
h
ca
n
be
b
r
o
k
en
at
a
n
y
m
o
m
e
n
t.
Fig
u
r
e
5
.
W
o
r
k
in
g
s
tep
s
of
K
NN
alg
o
r
ith
m
In
th
e
F
i
g
u
r
e
5,
T
h
e
K
-
n
ea
r
est
n
eig
h
b
o
u
r
alg
o
r
it
h
m
(
K
NN)
is
s
u
m
m
ar
is
ed
as:
i.
A
+
v
e
n
u
m
b
er
k
is
s
ta
ted
,
w
it
h
a
n
e
w
s
a
m
p
le
ii.
T
h
e
k
ite
m
s
ar
e
s
elec
ted
f
r
o
m
th
e
d
atab
ase
th
at
ar
e
n
e
x
t
to
n
e
w
s
a
m
p
le
iii.
T
h
e
u
t
m
o
s
t
m
u
tu
a
l
class
if
icati
o
n
of
s
elec
ted
en
tr
ie
s
is
d
eter
m
i
n
ed
.
iv
.
R
es
u
lted
C
la
s
s
i
f
icatio
n
is
o
f
f
e
r
ed
to
th
e
n
e
w
s
a
m
p
le.
In
KNN
clas
s
i
f
icatio
n
,
th
e
o
u
t
p
u
t
is
a
cla
s
s
m
e
m
b
er
s
h
ip
.
An
o
b
j
ec
t
is
class
i
f
ied
th
r
o
u
g
h
t
h
eb
u
lk
v
o
te
f
r
o
m
th
e
n
ea
r
b
y
n
ei
g
h
b
o
u
r
s
,
w
it
h
en
tit
y
b
ei
n
g
allo
ca
ted
to
class
m
o
s
t
m
u
t
u
al
a
m
o
n
g
t
h
e
en
titi
e
s
k
ad
j
o
in
in
g
n
eig
h
b
o
u
r
s
.
If
k
=
1,
th
e
o
b
ject
is
ass
ig
n
ed
to
class
of
th
at
s
o
le
n
ea
r
est
n
eig
h
b
o
u
r
.
A
p
ec
u
liar
it
y
of
KNN
alg
o
r
ith
m
is
t
h
at
it
s
s
e
n
s
i
tiv
i
t
y
to
lo
ca
l
s
tr
u
ctu
r
e
of
d
ata
[2
7
].
Ass
u
m
e,
tr
ain
i
n
g
s
et
D
i.
Ob
j
ec
t
to
be
test
ed
x
=
(
x
_
,
y
_
)
,
ii.
Af
ter
th
a
t
alg
o
r
it
h
m
ca
lc
u
lat
es
th
e
s
i
m
i
lar
it
y
b
et
w
ee
n
z
an
d
all
tr
ain
i
n
g
o
b
j
ec
ts
to
co
n
clu
d
e
its
n
ea
r
est
-
n
e
ig
h
b
o
u
r
lis
t
i.e
.
Dz.
T
r
ain
in
g
o
b
j
ec
ts
=(
x
,
y)
∈
D
iii.
x
=
d
ata
of
a
tr
ain
i
n
g
o
b
j
ec
t,
y=
is
it
s
clas
s
.
iv
.
Si
m
i
lar
l
y
,
x_
=
d
ata
of
th
e
test
o
b
j
ec
t
y
_
=
is
its
cla
s
s
T
h
e
class
if
icatio
n
of
test
o
b
j
ec
t
is
d
o
n
e
on
th
e
b
asis
of
m
aj
o
r
it
y
clas
s
of
its
n
ea
r
est
n
eig
h
b
o
u
r
s
w
h
ic
h
is
d
escr
ib
ed
in
th
e
eq
u
at
io
n
b
elo
w
:
Ma
jo
rit
y
V
oting
:
y
′
=
a
rgm
a
x
v
∑
I
(
v
=
y
i
)
,
(
x
i
,
y
i
)
ϵ
D
z
(
1
)
In
th
e
ab
o
v
e
eq
u
atio
n
;
v
=c
las
s
lab
el
y
i=c
las
s
lab
el
f
o
r
i
th
n
ea
r
est
n
eig
h
b
o
u
r
s
I
(·)=
in
d
icato
r
f
u
n
ctio
n
w
h
ic
h
r
etu
r
n
s
t
h
e
v
a
lu
e
1
if
its
ar
g
u
m
e
n
t
=
tr
u
e
a
n
d
o
th
er
w
i
s
e
0
is
r
etu
r
n
ed
as
a
v
al
u
e.
An
E
x
a
m
p
le
of
th
e
k
-
NN
clas
s
if
ica
tio
n
h
a
s
b
ee
n
ex
p
lai
n
ed
b
r
ief
l
y
alo
n
g
w
it
h
F
i
g
u
r
e
6.
T
h
e
F
ig
u
r
e
6
d
em
o
n
s
tr
ated
t
h
at
t
h
e
tes
t
m
o
d
el
(
i.e
.
g
r
ee
n
co
lo
u
r
ed
cir
cle)
can
be
class
i
f
ied
eit
h
er
to
f
i
r
s
t
class
of
t
h
e
b
lu
e
co
lo
u
r
ed
s
q
u
ar
es
or
to
th
e
o
th
er
class
of
r
ed
co
lo
u
r
ed
tr
ian
g
les.
If
k
=
3,
(
co
n
s
id
er
in
g
s
o
li
d
lin
e
cir
cle)
th
en
th
e
te
s
t
m
o
d
el
is
allo
ca
ted
to
t
h
e
s
ec
o
n
d
clas
s
as
t
h
er
e
ar
e2
t
r
ian
g
le
s
i
n
s
id
e
th
e
i
n
n
er
c
ir
cle
an
d
o
n
l
y
1
s
q
u
ar
e.
W
h
er
ea
s
,
if
k
=
5,
(
co
n
s
id
er
in
g
t
h
e
d
as
h
ed
li
n
e
cir
cle)
,
t
h
e
t
est
m
o
d
el
is
a
llo
tted
to
t
h
e
f
ir
s
t
clas
s
s
i
n
ce
t
h
er
e
ar
e
3
s
q
u
ar
es
in
s
id
e
th
e
o
u
ter
cir
cle
an
d
o
n
ly
2
tr
ian
g
le
s
.
T
h
e
allo
ca
tio
n
is
b
ased
on
th
e
m
aj
o
r
ity
v
o
te
of
its
n
eig
h
b
o
u
r
[
2
8
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8776
IJ
-
I
C
T
Vo
l.
8
,
No
.
1
,
A
p
r
il 2
0
1
9
:
39
–
49
46
Fig
u
r
e
6.
k
-
NN
class
if
ica
tio
n
E
u
clid
ea
n
Di
s
ta
n
ce
can
be
ca
l
cu
lated
by
u
s
i
n
g
:
D
(
x
,
y
)
=
√
∑
(
x
i
−
y
i
)
2
n
i
=
1
(
2
)
K
-
Nea
r
est
Ne
ig
h
b
o
u
r
can
be
p
r
ed
icted
by
e
m
p
lo
y
in
g
t
h
e
f
o
ll
o
w
i
n
g
eq
u
atio
n
:
y
=
1
k
∑
y
i
x
i
=
1
(
3
)
In
th
e
ab
o
v
e
eq
u
atio
n
,
y
i
=
ith
c
ase
of
test
m
o
d
el
;
y
=
o
u
tco
m
e
of
th
e
q
u
er
y
p
o
in
t.
I
n
clas
s
i
f
icatio
n
p
r
o
b
lem
s
,
o
n
a
v
o
ti
n
g
s
c
h
e
m
e
t
h
e
K
NN
p
r
ed
ictio
n
s
ar
e
b
ased
a
n
d
t
h
e
w
i
n
n
er
is
u
s
ed
to
lab
el
th
e
q
u
er
y
.
T
h
e
k
-
NN
al
g
o
r
ith
m
ac
c
u
r
a
c
y
can
be
s
tr
ictl
y
d
e
g
r
ad
ed
w
it
h
th
ee
x
i
s
ten
ce
of
n
o
is
y
f
ea
tu
r
es,
i
n
co
n
s
is
te
n
t
f
ea
t
u
r
e
s
ca
les
etc.
A
lo
t
of
r
ese
ar
ch
e
f
f
o
r
t
is
p
u
t
in
to
ch
o
o
s
i
n
g
or
s
ca
li
n
g
f
ea
t
u
r
es
to
i
m
p
r
o
v
e
th
e
p
er
f
o
r
m
a
n
ce
of
cla
s
s
i
f
icatio
n
.
T
h
e
ac
cu
r
ac
y
le
v
el
of
t
h
e
K
NN
alg
o
r
it
h
m
’
s
en
d
r
es
u
lt
ca
n
be
ca
lcu
lated
by
u
s
in
g
t
h
e
f
o
llo
w
in
g
eq
u
atio
n
.
A
ccu
ra
cy
=
(
No
.
of
co
r
r
e
ct
ly
cl
a
s
s
i
fi
e
d
e
x
a
mp
les
No
.
of
e
x
a
mp
les
)
×
100
(
4
)
P
s
eu
d
o
-
co
d
e
f
o
r
k
-
n
ea
r
est
n
ei
g
h
b
o
u
r
cla
s
s
i
f
icatio
n
alg
o
r
it
h
m
[2
9
]
k
n
u
m
b
er
of
n
ea
r
est
n
e
ig
h
b
o
r
s
f
o
r
each
o
b
j
ec
t
X
in
th
e
test
s
e
t
do
C
alcu
late
th
e
d
is
tan
ce
D(
X,
Y)
b
et
w
ee
n
X
a
n
d
ev
er
y
o
b
j
ec
t
Y
in
t
h
e
tr
ain
i
n
g
s
et
Neig
h
b
o
r
h
o
o
d
th
e
k
n
ei
g
h
b
o
u
r
s
in
th
e
tr
ai
n
i
n
g
s
e
t
clo
s
est
to
X
X.
class
Select
C
las
s
(
n
ei
g
h
b
o
r
h
o
o
d
)
en
d
f
o
r
So
m
eti
m
es
f
u
ll
d
escr
ip
tio
n
of
th
e
p
er
f
o
r
m
an
ce
of
cla
s
s
i
f
ica
tio
n
al
g
o
r
ith
m
is
r
eq
u
ir
ed
an
d
d
etailed
co
n
ce
p
tio
n
is
a
tab
le
en
t
itled
as
th
e
n
a
m
e
of
co
n
f
u
s
io
n
m
atr
i
x
.
T
h
e
r
o
w
s
d
en
o
te
t
h
e
r
ea
l
class
of
t
h
e
test
ca
s
es,
w
h
er
ea
s
,
co
lu
m
n
s
s
y
m
b
o
lis
e
s
th
e
p
r
ed
ictio
n
of
class
i
f
ier
s
.
T
h
e
title
co
n
f
u
s
io
n
m
atr
i
x
ar
is
e
s
f
r
o
m
o
b
s
er
v
atio
n
w
h
er
e
th
e
al
g
o
r
ith
m
g
e
ts
co
n
f
u
s
ed
.
Ass
u
m
e
th
e
d
atab
ase
c
o
n
tain
s
100
p
lay
er
s
f
r
o
m
th
e
w
o
m
e
n
g
y
m
n
ast
s
,
b
ask
etb
all
a
s
s
o
ciatio
n
a
n
d
m
a
r
ath
o
n
.
T
h
e
e
v
al
u
atio
n
of
clas
s
if
ier
is
d
o
n
e
w
i
th
10
-
f
o
ld
cr
o
s
s
v
alid
atio
n
.
T
h
e
r
esu
lt
s
of
t
h
is
te
s
t
ar
e
as
s
h
o
wn
in
T
ab
le
2
[
2
9
]
:
T
ab
le
2.
R
esu
lts
of
10
-
Fo
ld
C
r
o
s
s
-
Va
lid
atio
n
G
y
mn
a
st
s
B
a
s
k
e
t
b
a
l
l
P
l
a
y
e
r
s
M
a
r
a
t
h
o
n
e
r
s
G
y
mn
a
st
s
83
0
17
B
a
s
k
e
t
b
a
l
l
P
l
a
y
e
r
s
0
92
8
M
a
r
a
t
h
o
n
e
r
s
9
16
75
Evaluation Warning : The document was created with Spire.PDF for Python.
IJ
-
I
C
T
I
SS
N:
2252
-
8776
R
ec
en
t
tr
en
d
s
in
b
ig
d
a
ta
u
s
in
g
h
a
d
o
o
p
(
C
h
etn
a
K
a
u
s
h
a
l)
47
T
h
e
ac
tu
al
class
of
ea
c
h
ex
a
m
p
le
is
d
en
o
ted
by
r
o
w
s
;
t
h
e
cla
s
s
an
ticip
ated
by
o
u
r
class
i
f
ier
is
d
en
o
ted
by
co
lu
m
n
s
.
So
ta
k
en
ex
a
m
p
le,
83
=
co
r
r
ec
tly
cla
s
s
i
f
ied
g
y
m
n
asts
17
=
m
is
c
lass
if
ied
as
m
ar
at
h
o
n
er
s
.
92=
co
r
r
ec
tly
cla
s
s
i
f
ied
b
ask
e
t
b
all
p
lay
er
s
8
=
m
i
s
clas
s
i
f
ied
as
m
ar
at
h
o
n
er
s
.
75=
co
r
r
ec
tly
cla
s
s
i
f
ied
m
ar
ath
o
n
er
s
9
=
m
i
s
clas
s
i
f
ied
as
g
y
m
n
a
s
ts
16
=m
i
s
clas
s
if
ied
as
b
ask
etb
al
l
p
lay
er
s
.
T
h
e
co
n
f
u
s
io
n
m
atr
ix
d
ia
g
o
n
a
l
r
ep
r
esen
ts
in
s
ta
n
ce
s
w
h
ic
h
wer
e
class
if
ied
co
r
r
ec
tly
.
In
th
i
s
ca
s
e
t
h
e
ac
cu
r
ac
y
of
th
e
alg
o
r
ith
m
is
:
83
+
92
+
75
300
=
250
300
=
83
.
33%
6.
RE
L
AT
E
D
WO
RK
W
u
,
X.
et
al,
p
r
esen
ted
a
c
o
m
p
r
eh
en
s
i
v
e
s
t
u
d
y
r
eg
ar
d
in
g
th
e
to
p
m
o
s
t
10
alg
o
r
ith
m
s
of
d
ata
m
in
i
n
g
[2
5
].
T
h
e
alg
o
r
ith
m
s
w
h
o
s
e
co
m
p
r
e
h
en
s
iv
e
ap
p
r
o
ac
h
w
as
m
e
n
tio
n
ed
w
er
e:
C
4
.
5
,
SV
M,
k
-
Me
a
n
s
,
EM,
A
p
r
io
r
i,
A
d
aB
o
o
s
t,
Naiv
e
B
ay
es,
C
AR
T
an
d
k
NN.
T
h
ese
alg
o
r
ith
m
s
i
n
cl
u
d
ed
all
clu
s
te
r
in
g
,
clas
s
i
f
icatio
n
,
ass
o
ciatio
n
an
a
l
y
s
is
;
s
ta
tis
tica
l
lear
n
i
n
g
a
n
d
las
t
li
n
k
i
n
g
th
a
t
w
er
e
tr
ea
ted
as
t
h
e
m
o
s
t
s
i
g
n
if
ica
n
t
to
p
ics
in
t
h
e
r
esear
ch
of
d
ata
m
i
n
i
n
g
.
T
h
e
i
m
p
ac
t
of
alg
o
r
it
h
m
s
h
as
b
ee
n
d
is
cu
s
s
ed
;
co
m
p
ar
is
o
n
w
a
s
d
o
n
e
on
th
e
b
asis
of
w
h
ic
h
f
u
tu
r
e
f
o
r
ec
ast
h
as
b
ee
n
d
eli
v
er
ed
.
In
t
h
e
later
y
ea
r
,
B
ak
s
h
i,
K.
,
et
al,
f
o
c
u
s
ed
on
an
a
l
y
s
is
of
u
n
s
tr
u
ct
u
r
ed
d
ata
w
h
ic
h
r
e
f
er
s
to
t
h
e
i
n
f
o
r
m
atio
n
w
h
ic
h
m
a
y
d
o
es
n
o
t
co
n
tai
n
p
r
ev
io
u
s
l
y
d
ef
in
ed
d
ata
m
o
d
el
or
w
as
n
o
t
s
u
i
tab
le
to
f
it
in
r
elatio
n
al
tab
les
[2
9
].
T
h
er
e
w
er
e
m
a
n
y
m
eth
o
d
s
to
tack
le
t
h
e
p
r
o
b
le
m
of
u
n
s
tr
u
ct
u
r
ed
d
ata.
T
h
e
m
et
h
o
d
s
s
h
ar
ed
m
u
t
u
al
f
ea
tu
r
e
s
of
elasticit
y
,
h
i
g
h
ac
ce
s
s
ib
ilit
y
a
n
d
s
ca
le
-
o
u
t.
Map
R
ed
u
ce
in
u
n
i
f
icatio
n
w
i
th
H
ad
o
o
p
f
ile
s
y
s
te
m
w
h
ich
is
m
ai
n
l
y
d
i
s
tr
ib
u
ted
an
d
H
-
B
as
e
d
atab
ase,
p
ar
t
of
A
p
ac
h
e
Had
o
o
p
p
lan
w
h
ic
h
h
elp
ed
in
a
n
al
y
s
i
n
g
t
h
e
u
n
s
tr
u
c
tu
r
ed
d
ata.
P
r
iy
ad
h
ar
s
i
n
i,
C
.
,
et
al,
p
r
esen
ted
an
ex
ten
s
i
v
e
s
tu
d
y
on
m
et
h
o
d
s
of
d
ata
m
in
in
g
a
n
d
also
s
u
m
m
ar
y
of
d
atab
ase
r
elate
d
to
k
n
o
w
led
g
e
d
i
s
co
v
er
y
[
1
1
]
.
T
h
e
m
ain
f
o
c
u
s
w
as
on
th
e
i
s
s
u
es
r
elate
d
to
th
e
d
ata
m
i
n
in
g
.
R
o
d
r
íg
u
ez
-
Ma
za
h
u
a
L.
et
a
l,
p
r
esen
ted
a
r
ev
ie
w
of
B
ig
Data
w
o
r
k
s
f
o
r
id
en
tif
icatio
n
of
t
h
e
ch
ie
f
p
r
o
b
lem
s
,
to
o
ls
,
ap
p
licatio
n
ar
ea
an
d
d
ev
elo
p
in
g
s
t
y
le
s
of
B
ig
Data
[
1
5
]
.
To
m
e
et
th
e
o
b
j
ec
tiv
e,
au
t
h
o
r
s
h
av
e
s
tu
d
ied
457
p
ap
er
s
to
clas
s
if
y
th
e
t
h
eo
r
ies
r
elate
d
to
B
ig
Data
.
T
h
is
a
n
al
y
ze
d
wo
r
k
o
f
f
er
ed
r
elate
d
m
ater
ial
to
r
esear
ch
er
s
r
e
g
ar
d
in
g
k
e
y
w
o
r
k
in
g
in
s
tu
d
y
a
n
d
B
ig
Data
ap
p
licatio
n
in
d
iv
er
s
e
p
r
ac
tical
ar
ea
s
.
L
ater
,
S
h
i
k
h
a
Si
n
g
h
,
D.
et
a
l,
d
is
c
u
s
s
ed
th
e
c
h
alle
n
g
e
s
t
h
at
ex
p
an
d
t
h
e
u
tili
t
y
of
lar
g
e
d
ata
th
o
u
g
h
a
tte
m
p
ti
n
g
to
g
r
a
s
p
th
e
ap
p
r
o
p
r
iate
s
tr
ateg
y
to
p
r
o
cu
r
e
p
r
ev
io
u
s
k
n
o
w
led
g
e
f
r
o
m
lar
g
e
d
ata
s
t
ac
k
[
2
]
.
T
h
er
e
w
as
y
et
a
d
is
p
u
te
co
n
ce
r
n
in
g
t
h
e
m
ec
h
a
n
is
m
s
a
n
d
estab
lis
h
ed
m
an
a
g
e
m
e
n
t
s
tr
u
c
tu
r
es
w
h
ic
h
w
er
e
in
e
f
f
icie
n
t
w
it
h
B
ig
Data
.
It
h
ig
h
l
ig
h
ted
s
u
ch
d
o
cu
m
en
t
s
an
d
s
ev
er
al
n
e
w
tech
n
o
lo
g
ies
t
h
at
r
ev
ea
l
t
h
e
c
h
allen
g
es
b
ased
on
th
e
id
ea
of
B
ig
Data
.
A
la
m
,
A
.
,
et
al,
d
ef
in
ed
th
e
a
r
ch
itect
u
r
e
an
d
th
e
ch
alle
n
g
e
s
of
H
A
DOOP
[
1
7
]
.
T
h
e
m
ai
n
p
r
o
b
lem
ar
ea
w
h
ich
h
as
b
ee
n
m
e
n
tio
n
e
d
w
as
t
h
e
iter
ati
v
e
r
u
n
n
i
n
g
of
m
ap
-
r
ed
u
ce
p
r
o
ce
s
s
e
s
f
r
o
m
t
h
e
b
eg
in
n
i
n
g
e
v
en
in
little
m
i
n
o
r
alter
atio
n
in
i
n
p
u
t.
It
w
a
s
n
o
t
a
g
o
o
d
ap
p
r
o
ac
h
as
ev
er
y
ti
m
e
in
t
h
e
b
ig
d
ata
clo
u
d
th
e
en
tr
ies
ar
e
ad
d
ed
or
d
elete
d
in
th
e
b
u
l
k
a
m
o
u
n
t,
t
h
e
p
r
o
ce
s
s
in
g
s
p
ee
d
n
ee
d
s
to
be
at
its
u
t
m
o
s
t
le
v
el.
In
th
e
s
o
lu
t
io
n
,
ca
ch
in
g
s
ch
e
m
e
w
a
s
d
escr
ib
ed
at
s
m
a
ll
le
v
el
w
h
ich
h
elp
ed
in
m
an
a
g
i
n
g
th
e
ac
tiv
ities
v
er
y
w
ell
in
m
ap
r
ed
u
ce
f
u
n
ctio
n
s
.
Ke
s
a
v
ar
aj
,
G.
,
et
al,
s
p
ec
if
ied
t
h
e
ad
v
an
ta
g
e
s
an
d
d
r
a
w
b
ac
k
s
of
t
h
e
d
i
f
f
e
r
en
t
clas
s
i
f
icatio
n
alg
o
r
ith
m
s
a
n
d
th
e
b
est
alg
o
r
i
th
m
s
ac
co
r
d
in
g
to
p
r
ev
io
u
s
s
t
u
d
ies
w
as
KNN
[2
4
].
T
h
e
av
er
ag
e
ac
cu
r
ac
y
h
a
s
b
ee
n
ca
lcu
lated
a
n
d
th
e
g
e
n
eti
c
alg
o
r
ith
m
h
as
t
h
e
b
est
ac
c
u
r
ac
y
r
ate
w
it
h
4
6
.
6
7
%.
T
h
e
ef
f
icien
c
y
,
p
r
ec
is
io
n
,
ac
cu
r
ac
y
,
s
e
n
s
iti
v
it
y
of
t
h
e
class
i
f
icatio
n
alg
o
r
it
h
m
s
h
as
b
ee
n
co
m
p
ar
ed
an
d
th
e
n
eu
r
al
h
as
ac
h
ie
v
ed
th
e
s
ec
o
n
d
h
ig
h
es
t
6
2
.
8
af
ter
t
h
e
b
ac
k
-
p
r
o
p
ag
atio
n
al
g
o
r
ith
m
ac
co
r
d
in
g
to
p
r
ev
io
u
s
s
t
u
d
ies
.
So
k
o
lo
v
a,
M.
,
et
al,
p
r
esen
ted
t
h
e
a
n
al
y
s
is
of
t
h
e
m
ac
h
in
e
lear
n
in
g
cla
s
s
i
f
icatio
n
ta
s
k
s
w
h
ic
h
w
er
e
b
in
ar
y
,
m
u
lti
-
cla
s
s
,
h
ier
ar
ch
ical
an
d
m
u
lti
-
lab
ell
ed
[2
8
].
Dif
f
er
en
t
ch
a
n
g
es
in
t
h
e
co
n
f
u
s
io
n
m
atr
i
x
on
v
ar
io
u
s
w
ell
-
k
n
o
w
n
m
ea
s
u
r
es
h
av
e
b
ee
n
r
ev
ie
w
e
d
an
d
co
m
p
ar
ed
.
Gan
d
h
i
et
al.
h
av
e
i
m
p
le
m
e
n
ted
th
e
e
x
is
t
in
g
K
-
m
ea
n
,
K
-
m
ed
o
id
s
an
d
t
h
e
p
r
esen
ted
M
o
d
if
ied
K
-
m
ed
o
id
alg
o
r
ith
m
s
.
T
h
e
K
-
m
ed
o
id
is
b
ein
g
e
x
ec
u
ted
h
a
s
p
er
f
o
r
m
ed
b
etter
as
co
m
p
ar
ed
to
K
-
m
ea
n
an
d
ex
i
s
ti
n
g
K
-
me
d
o
id
s
on
h
u
g
e
d
ata
s
ets
f
o
r
ex
ec
u
tio
n
ti
m
e
an
d
clu
s
ter
i
n
g
q
u
alit
y
in
t
h
e
e
x
p
er
i
m
e
n
tal
o
u
tco
m
e
s
.
T
h
e
au
th
o
r
h
a
s
ca
lc
u
lated
Du
n
n
’
s
i
n
d
ex
,
to
tal
ti
m
e,
d
av
ies
b
o
u
ld
in
in
d
ex
,
Kr
za
n
o
w
s
k
i
an
d
L
ai,
C
ali
n
s
k
iHar
ab
asz
i
n
d
ex
f
o
r
th
e
v
er
i
f
icatio
n
of
t
h
e
m
o
d
i
f
i
ed
K
-
m
ed
o
id
s
w
it
h
ex
is
i
tin
g
K
-
m
ed
o
id
s
an
d
K
-
m
ea
n
p
er
f
o
r
m
a
n
ce
.
It
h
as
b
ee
n
co
n
cl
u
d
ed
f
r
o
m
t
h
e
r
es
u
lt
th
at
t
h
e
m
o
d
i
f
ied
k
-
m
ed
o
id
s
h
a
s
p
er
f
o
r
m
ed
b
etter
[
3
1
]
.
A
r
o
r
a
et
al.
h
as
o
b
tai
n
e
d
en
h
a
n
ce
d
r
es
u
lt
s
of
cl
u
s
ter
i
n
g
by
u
ti
lizin
g
t
w
o
clu
s
ter
i
n
g
al
g
o
r
ith
m
s
by
m
ea
n
s
of
v
ar
ied
cl
u
s
ter
s
b
ei
n
g
f
o
r
m
ed
a
n
d
by
m
ea
n
s
of
d
is
ta
n
c
e
m
etr
ic.
C
l
u
s
ter
in
g
alg
o
r
ith
m
s
,
li
k
e,
K
-
m
ea
n
an
d
K
-
m
ed
o
id
s
ha
b
ee
n
u
s
ed
on
t
h
e
d
ataset
tr
an
s
ac
tio
n
10k
of
K
E
E
L
.
T
h
e
in
p
u
t
h
a
s
b
ee
n
ar
b
itra
r
ily
d
i
s
p
en
d
ed
d
ata
p
o
in
ts
an
d
ac
co
r
d
in
g
l
y
,
s
i
m
i
lar
it
y
cl
u
s
ter
s
h
as
b
ee
n
p
r
o
d
u
ce
d
.
It
can
be
d
r
aw
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8776
IJ
-
I
C
T
Vo
l.
8
,
No
.
1
,
A
p
r
il 2
0
1
9
:
39
–
49
48
f
r
o
m
t
h
e
r
esear
ch
t
h
at
th
e
w
h
en
t
h
e
d
is
ta
n
ce
m
etr
ic
ch
an
g
es,
t
h
e
o
u
tco
m
e
of
clu
s
ter
i
n
g
alg
o
r
it
h
m
ch
an
g
es
[
3
1
-
34
].
7.
CO
M
P
ARIS
O
N
OF
E
XIS
T
I
NG
WO
RK
T
h
is
s
ec
tio
n
d
ep
icts
t
h
e
co
m
p
ar
is
o
n
of
th
e
w
o
r
k
of
[
3
0
]
an
d
[
3
1
]
.
Fro
m
th
e
liter
at
u
r
e
s
t
u
d
y
,
t
h
e
co
m
p
ar
is
o
n
h
as
b
ee
n
m
ad
e
on
t
h
e
b
asis
of
ex
ec
u
tio
n
ti
m
e
on
b
ig
d
ata
cl
u
s
ter
i
n
g
f
o
r
d
if
f
er
e
n
t
cl
u
s
ter
i
n
g
ap
p
r
o
ac
h
es.
T
h
e
r
esu
lts
h
a
v
e
b
ee
n
s
h
o
w
n
b
elo
w
in
t
h
e
f
o
r
m
of
T
ab
le
3
an
d
Fig
u
r
e
7
.
T
ab
le
3
.
C
o
m
p
ar
is
o
n
of
E
x
ec
u
tio
n
T
im
e
o
f
B
ig
Data
C
lu
s
ter
i
n
g
K
-
m
e
a
n
(
Ex
e
c
u
t
i
o
n
t
i
me
(
se
c
s)
)
K
-
m
e
d
o
i
d
(
Ex
e
c
u
t
i
o
n
t
i
me
(
se
c
s)
)
G
o
p
i
G
a
n
d
h
i
a
n
d
R
o
h
i
t
S
r
i
v
a
st
v
a
[
3
1
]
P
r
e
e
t
i
A
r
o
r
a
et
a
l
.
[
3
0
]
G
o
p
i
G
a
n
d
h
i
a
n
d
R
o
h
i
t
S
r
i
v
a
st
v
a
[
3
1
]
P
r
e
e
t
i
A
r
o
r
a
et
a
l
.
[
3
0
]
0
.
2
0
1
4
0
.
0
3
5
8
0
.
2
2
2
3
0
.
0
3
8
4
Fig
u
r
e
7.
C
o
m
p
ar
is
o
n
of
cl
u
s
t
er
in
g
ap
p
r
o
ac
h
es
of
[
3
0
]
an
d
[
3
1
]
A
b
o
v
e
tab
le
an
d
g
r
ap
h
d
ep
ict
s
th
e
co
m
p
ar
is
o
n
of
b
ig
d
ata
clu
s
ter
i
n
g
of
ex
ec
u
tio
n
ti
m
e
of
[
3
0
]
an
d
[
3
1
]
.
T
h
e
co
m
p
ar
is
o
n
h
as
b
ee
n
m
ad
e
on
th
e
b
asis
of
K
-
m
e
an
a
n
d
K
-
m
ed
o
id
ap
p
r
o
ac
h
es.
T
h
e
au
t
h
o
r
Go
p
i
Gan
d
i
an
d
R
o
h
it
Sriv
a
s
t
v
a
h
as
u
s
ed
Si
m
ilar
it
y
i
n
d
ex
w
it
h
K
-
m
ed
o
id
s
clu
s
ter
in
g
tec
h
n
i
q
u
e
to
en
h
an
ce
th
e
p
er
f
o
r
m
a
n
ce
of
cl
u
s
ter
i
n
g
.
So
,
th
e
ex
ec
u
t
io
n
ti
m
e
in
t
h
eir
w
o
r
k
is
less
as
co
m
p
ar
ed
to
th
e
w
o
r
k
of
P
r
ee
ti
A
r
o
r
a
et
al.
As
s
h
o
w
n
in
th
e
g
r
ap
h
an
d
tab
le,
t
h
e
v
al
u
e
of
e
x
ec
u
tio
n
ti
m
e
f
o
r
[
3
1
]
is
0
.
2
0
1
4
an
d
f
o
r
[
3
0
]
,
it
is
0
.
0
3
5
8
f
o
r
K
-
m
ea
n
ap
p
r
o
ac
h
.
Si
m
i
lar
l
y
,
in
ca
s
e
of
k
-
m
ed
o
id
s
,
th
e
v
al
u
e
in
ca
s
e
of
[
3
1
]
is
0
.
2
2
2
3
an
d
f
o
r
[
3
0
]
,
it
is
0
.
0
3
8
4
.
T
h
e
b
lu
e
b
ar
in
t
h
e
g
r
ap
h
is
d
ep
icti
n
g
t
h
e
w
o
r
k
of
Go
p
i
Ga
n
d
h
i
a
n
d
R
o
h
it
S
r
iv
ast
v
a
a
n
d
r
ed
b
a
r
is
d
ep
ictin
g
t
h
e
w
o
r
k
of
P
r
ee
ti
ar
o
r
a
et
al.
T
h
e
X
-
a
x
is
is
f
o
r
th
e
ap
p
r
o
ac
h
es
b
ei
n
g
u
t
ilized
f
o
r
th
e
co
m
p
ar
is
o
n
an
d
Y
-
a
x
is
is
s
h
o
w
in
g
t
h
e
v
al
u
es
of
t
h
e
ex
ec
u
tio
n
ti
m
e
in
s
e
co
n
d
s
.
8.
CO
NCLU
SI
O
N
An
o
v
er
v
ie
w
of
b
ig
d
ata
is
p
r
esen
ted
alo
n
g
w
it
h
b
ig
d
ata
u
s
a
g
es
a
n
d
s
ev
er
al
ch
a
llen
g
e
s
th
at
ar
e
ass
o
ciate
d
w
it
h
b
i
g
d
ata.
T
h
is
p
ap
er
co
v
er
s
th
e
s
tu
d
y
on
d
ata
m
i
n
i
n
g
an
d
k
n
o
w
led
g
e
d
is
c
o
v
er
y
in
d
atab
ases
(
KDD)
w
ith
all
t
h
e
s
tep
s
t
h
at
ar
e
in
v
o
l
v
ed
in
t
h
e
KDD
p
r
o
ce
s
s
.
T
h
e
is
s
u
e
s
r
elate
d
to
t
h
e
c
lu
s
ter
i
n
g
tec
h
n
iq
u
es
in
d
ata
m
i
n
i
n
g
ar
e
al
s
o
d
is
c
u
s
s
ed
b
r
ief
l
y
.
T
h
e
co
m
p
lete
ar
ch
itectu
r
e
of
Had
o
o
p
an
d
HDFS
is
al
s
o
s
t
u
d
ied
an
d
d
is
cu
s
s
ed
.
Fo
r
class
i
f
icatio
n
of
th
e
d
ata,
s
e
v
er
al
tr
ad
itio
n
al
m
et
h
o
d
s
s
u
c
h
as
r
u
le
b
ased
,
d
ec
is
io
n
tr
ee
,
r
an
d
o
m
f
o
r
ests
,
b
o
o
s
tin
g
,
q
u
ad
r
atic
class
i
f
ier
s
a
s
s
o
ciate
d
w
it
h
cl
ass
i
f
icatio
n
ar
e
b
r
ief
l
y
s
t
u
d
i
ed
an
d
th
en
KNN
class
i
f
icatio
n
al
g
o
r
ith
m
is
s
el
ec
ted
f
o
r
th
e
d
ata
m
in
i
n
g
a
n
d
d
escr
ib
ed
in
t
h
is
p
ap
er
.
A
n
e
x
a
m
p
le
is
tak
e
n
to
p
r
o
v
e
th
e
ac
cu
r
ac
y
of
KNN
alg
o
r
ith
m
w
h
ich
is
m
ea
s
u
r
ed
to
be
8
3
.
3
3
%.
A
co
m
p
ar
is
o
n
h
as
b
ee
n
m
ad
e
on
clu
s
ter
i
n
g
al
g
o
r
ith
m
s
,
n
a
m
el
y
,
K
-
m
ea
n
a
n
d
K
-
m
ed
o
id
f
o
r
ex
ec
u
tio
n
ti
m
e
of
th
e
ex
i
s
ti
n
g
w
o
r
k
of
[
3
0
]
an
d
[
3
1
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.