I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
pu
t
er
Science
Vo
l.
3
9
,
No
.
2
,
A
u
g
u
s
t
2
0
2
5
,
p
p
.
1
399
~
1
4
1
0
I
SS
N:
2
5
0
2
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/ijeecs.v
3
9
.i
2
.
pp
1
3
9
9
-
1
4
1
0
1399
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
Enha
ncing
ma
rke
ting ef
ficiency
thr
o
ug
h data
-
driv
en
custo
mer
seg
menta
tion wi
t
h ma
chine l
ea
rni
ng
appro
a
ches
F
a
nin
dia
P
urna
m
a
s
a
ri,
Um
a
y
a
Ra
m
a
dh
a
ni P
utr
i N
a
s
utio
n,
M
a
rischa
E
lv
eny
D
e
p
a
r
t
me
n
t
o
f
I
n
f
o
r
m
a
t
i
o
n
Te
c
h
n
o
l
o
g
y
,
F
a
c
u
l
t
y
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
I
n
f
o
r
mat
i
o
n
T
e
c
h
n
o
l
o
g
y
,
U
n
i
v
e
r
si
t
a
s
S
u
ma
t
e
r
a
U
t
a
r
a
,
M
e
d
a
n
,
I
n
d
o
n
e
s
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Oct
4
,
2
0
2
4
R
ev
is
ed
Mar
28
,
2
0
2
5
Acc
ep
ted
J
u
l
2
,
2
0
2
5
Th
e
imp
o
rtan
c
e
o
f
u
n
d
e
rsta
n
d
i
n
g
c
o
n
su
m
e
r
b
e
h
a
v
i
o
r
in
tran
sa
c
ti
o
n
d
a
ta
h
a
s
b
e
c
o
m
e
a
k
e
y
t
o
im
p
ro
v
in
g
m
a
rk
e
ti
n
g
e
fficie
n
c
y
.
T
h
is
st
u
d
y
a
ims
to
e
x
p
lo
re
th
e
a
p
p
li
c
a
ti
o
n
o
f
m
a
c
h
in
e
le
a
rn
in
g
(
M
L)
tec
h
n
i
q
u
e
s
fo
r
d
a
ta
-
d
riv
e
n
c
o
n
su
m
e
r
se
g
m
e
n
tatio
n
,
fo
c
u
s
in
g
o
n
imp
r
o
v
i
n
g
p
ro
d
u
c
t
m
a
rk
e
ti
n
g
stra
teg
ies
.
Th
is
wo
r
k
a
d
d
re
ss
e
s
th
e
li
m
it
a
ti
o
n
s
in
th
e
e
x
ist
in
g
li
tera
tu
re
,
e
sp
e
c
ially
in
term
s
o
f
h
a
n
d
li
n
g
h
ig
h
-
d
ime
n
si
o
n
a
l
d
a
ta
t
h
a
t
c
a
n
re
d
u
c
e
se
g
m
e
n
tatio
n
q
u
a
li
ty
.
P
re
v
io
u
sl
y
,
v
a
rio
u
s
st
u
d
ies
h
a
v
e
u
se
d
c
lu
ste
rin
g
a
lg
o
rit
h
m
s
su
c
h
a
s
K
-
m
e
a
n
s
wi
t
h
o
u
t
c
o
n
si
d
e
rin
g
d
ime
n
sio
n
a
li
ty
re
d
u
c
ti
o
n
,
wh
ich
o
ften
lea
d
s
to
d
e
c
re
a
se
d
a
c
c
u
ra
c
y
a
n
d
lo
n
g
c
o
m
p
u
tatio
n
ti
m
e
.
In
th
is
stu
d
y
,
we
p
ro
p
o
se
a
n
e
w
a
p
p
r
o
a
c
h
th
a
t
c
o
m
b
in
e
s
p
ri
n
c
ip
a
l
c
o
m
p
o
n
e
n
t
a
n
a
ly
sis
(P
CA)
f
o
r
d
ime
n
sio
n
a
l
it
y
re
d
u
c
ti
o
n
a
n
d
K
-
m
e
a
n
s
c
lu
s
terin
g
fo
r
c
o
n
su
m
e
r
se
g
m
e
n
tatio
n
b
a
se
d
o
n
p
u
rc
h
a
sin
g
b
e
h
a
v
io
r.
E
x
p
e
rime
n
tal
re
su
lt
s
sh
o
w
th
a
t
u
si
n
g
P
CA
to
re
d
u
c
e
d
a
ta
d
ime
n
sio
n
a
li
ty
sig
n
ifi
c
a
n
tl
y
imp
ro
v
e
s
se
g
m
e
n
tatio
n
q
u
a
li
t
y
wit
h
a
n
i
n
e
rti
a
sc
o
re
o
f
1
,
4
5
5
,
6
5
0
a
n
d
a
silh
o
u
e
tt
e
sc
o
re
o
f
0
.
4
8
6
3
6
6
.
By
imp
lem
e
n
ti
n
g
t
h
is
m
e
th
o
d
,
we
c
a
n
g
ro
u
p
c
o
n
su
m
e
rs
in
to
t
h
re
e
se
g
m
e
n
ts
b
a
se
d
o
n
fr
e
q
u
e
n
tl
y
p
u
rc
h
a
se
d
p
ro
d
u
c
t
c
a
teg
o
ries
a
n
d
th
e
m
o
st co
m
m
o
n
p
a
y
m
e
n
t
m
e
th
o
d
s.
Th
e
se
fi
n
d
i
n
g
s
p
ro
v
id
e
a
sc
a
lab
le,
d
a
ta
-
d
riv
e
n
se
g
m
e
n
tatio
n
fra
m
e
wo
rk
th
a
t
c
a
n
b
e
a
p
p
li
e
d
to
imp
ro
v
e
m
a
rk
e
ti
n
g
e
ffe
c
ti
v
e
n
e
ss
b
y
p
ro
v
id
in
g
sp
e
c
i
a
l
d
isc
o
u
n
ts
o
n
v
a
rio
u
s
p
ro
d
u
c
ts
b
a
se
d
o
n
th
e
p
a
y
m
e
n
t
m
e
th
o
d
u
se
d
.
K
ey
w
o
r
d
s
:
C
lu
s
ter
in
g
C
u
s
to
m
er
b
eh
av
io
r
C
u
s
to
m
er
s
eg
m
en
tatio
n
Ma
ch
in
e
lear
n
in
g
Ma
r
k
etin
g
ef
f
icien
cy
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Fan
in
d
ia
Pu
r
n
am
asar
i
Dep
ar
tm
en
t o
f
I
n
f
o
r
m
atio
n
T
e
ch
n
o
lo
g
y
,
Facu
lty
o
f
C
o
m
p
u
te
r
Scien
ce
an
d
I
n
f
o
r
m
atio
n
T
ec
h
n
o
lo
g
y
Un
iv
er
s
itas
Su
m
ater
a
Utar
a
Me
d
a
n
,
I
n
d
o
n
esia
E
m
ail:
f
an
in
d
ia@
u
s
u
.
ac
.
id
1.
I
NT
RO
D
UCT
I
O
N
C
u
s
to
m
er
s
eg
m
en
tatio
n
h
elp
t
h
e
co
m
p
an
y
to
im
p
r
o
v
e
th
e
p
r
o
f
its
.
Acq
u
ir
in
g
n
ew
co
n
s
u
m
er
s
is
n
o
t
q
u
ite
as
im
p
o
r
tan
t
as
k
ee
p
in
g
o
ld
cu
s
to
m
er
Acc
o
r
d
in
g
t
o
th
e
Par
eto
p
r
in
cip
le,
a
c
o
m
p
an
y
'
s
cu
s
to
m
er
s
co
n
tr
ib
u
te
to
2
0
%
o
f
its
r
ev
e
n
u
e,
wh
ich
is
h
ig
h
er
th
an
th
a
t
o
f
o
th
er
c
u
s
to
m
er
s
[
1
]
.
T
h
e
co
m
p
an
y
ca
n
u
s
e
cu
s
to
m
er
s
eg
m
e
n
tatio
n
to
m
ar
k
etin
g
b
u
d
g
etin
g
cu
s
to
m
ize
m
ar
k
etin
g
s
tr
ateg
ies,
o
b
s
er
v
e
tr
en
d
s
,
o
r
g
an
ize
p
r
o
d
u
ct
d
ev
elo
p
m
en
t,
d
esig
n
ad
v
er
tis
in
g
ca
m
p
ai
g
n
s
,
an
d
p
r
o
v
id
e
ap
p
r
o
p
r
iat
e
p
r
o
d
u
cts
b
y
u
tili
zin
g
a
r
an
g
e
o
f
d
is
tin
ctiv
e
cu
s
to
m
er
attr
ib
u
tes
[
2
]
.
An
o
r
g
an
izatio
n
m
a
y
n
o
lo
n
g
e
r
h
av
e
to
p
r
o
m
o
te
p
r
o
d
u
cts
at
r
an
d
o
m
b
ec
au
s
e
th
er
e
is
a
d
ir
ec
t
ass
o
ciati
o
n
b
etwe
en
t
h
e
p
r
o
d
u
cts
b
ein
g
p
r
o
m
o
ted
a
n
d
i
n
ter
ested
clien
ts
ar
e
m
o
r
e
lik
ely
to
m
ak
e
f
r
e
q
u
en
t
p
u
r
c
h
ases
.
T
o
id
en
tify
th
e
r
e
g
io
n
s
wh
er
e
m
ar
k
ets,
cu
s
to
m
er
s
,
an
d
tr
an
s
ac
tio
n
s
ar
e
m
o
s
t
p
r
ev
alen
t,
f
ir
s
t
u
s
ed
lo
ca
tio
n
d
is
tr
ib
u
tio
n
h
ea
tm
a
p
s
[
3
]
.
I
t
d
eliv
er
ed
to
th
e
r
i
g
h
t
lo
ca
tio
n
in
th
e
s
h
o
r
test
tim
e
an
d
at
th
e
least
co
s
t
t
o
co
n
s
u
m
er
s
.
T
h
e
m
ar
k
et
d
i
s
tr
ib
u
tio
n
s
tr
ateg
y
p
er
m
ea
tes
all
asp
ec
t
s
o
f
th
e
o
r
g
an
izatio
n
’
s
ac
tio
n
s
an
d
en
co
m
p
ass
es
it
s
s
u
p
p
ly
,
p
r
o
d
u
cti
o
n
,
p
r
o
m
o
tio
n
,
an
d
d
is
tr
ib
u
tio
n
en
v
ir
o
n
m
en
ts
[
4
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
3
9
,
No
.
2
,
Au
g
u
s
t
20
25
:
1
3
9
9
-
1
410
1400
I
n
th
is
er
a,
m
ac
h
in
e
lear
n
in
g
(
ML
)
h
as
b
ec
o
m
e
an
ef
f
ec
tiv
e
ap
p
r
o
ac
h
f
o
r
im
p
r
o
v
in
g
cu
s
to
m
er
s
eg
m
en
tatio
n
,
p
ar
ticu
lar
ly
wh
e
n
it
co
m
es
to
u
n
s
u
p
er
v
is
ed
lear
n
in
g
m
eth
o
d
s
lik
e
clu
s
ter
in
g
alg
o
r
ith
m
s
.
ML
alg
o
r
ith
m
s
h
a
v
e
th
e
ab
ilit
y
to
id
en
tify
u
n
d
is
co
v
er
ed
tr
en
d
s
an
d
cu
s
to
m
er
s
b
a
s
ed
o
n
ac
tu
al
b
eh
av
io
r
al
ten
d
e
n
cies
b
y
ev
alu
atin
g
v
ast
am
o
u
n
ts
o
f
c
u
s
to
m
er
d
ata
[
5
]
,
[
6
]
.
B
u
s
in
ess
es
ca
n
ad
ap
t
th
eir
m
ar
k
eti
n
g
ca
m
p
a
ig
n
s
to
th
e
u
n
i
q
u
e
r
eq
u
ir
em
e
n
ts
an
d
p
r
ef
er
e
n
ce
s
o
f
v
ar
io
u
s
co
n
s
u
m
e
r
s
eg
m
en
ts
as
a
r
esu
lt
o
f
th
ese
d
ata
-
d
r
iv
e
n
in
s
ig
h
ts
.
C
u
s
to
m
er
s
eg
m
en
tatio
n
o
r
cu
s
to
m
er
clu
s
ter
in
g
is
s
tr
ateg
ic
d
ec
is
io
n
m
ak
in
g
f
o
r
b
u
s
in
ess
to
s
ee
k
s
u
s
tain
ab
le
g
r
o
wth
a
n
d
cu
s
to
m
er
s
atis
f
ac
tio
n
.
C
lu
s
ter
in
g
is
u
s
ed
to
id
en
tify
p
atter
n
s
,
s
u
ch
as
t
h
e
to
p
s
ellin
g
p
r
o
d
u
ct
an
d
th
e
p
r
ef
er
r
ed
p
a
y
m
en
t
m
et
h
o
d
b
ased
o
n
c
u
s
to
m
er
tr
an
s
ac
tio
n
[
7
]
.
K
-
m
ea
n
s
is
o
n
e
o
f
th
e
m
o
s
t
f
r
eq
u
e
n
tly
u
s
ed
in
cu
s
to
m
er
clu
s
ter
in
g
[
8
]
,
[
9
]
.
Ho
wev
e
r
,
s
o
m
e
clu
s
ter
in
g
alg
o
r
ith
m
s
s
u
ch
as
K
-
m
ea
n
s
o
f
ten
en
c
o
u
n
te
r
p
r
o
b
lem
s
wh
en
a
p
p
lied
to
d
at
a
with
h
ig
h
d
im
en
s
io
n
s
o
r
f
ea
t
u
r
es.
So
m
e
p
r
o
b
lem
s
ar
e
d
ec
r
ea
s
ed
class
if
icatio
n
ac
cu
r
ac
y
,
p
o
o
r
q
u
ality
o
f
clu
s
ter
,
an
d
lo
n
g
co
m
p
u
tin
g
tim
es.
Dim
en
s
io
n
r
ed
u
ctio
n
is
o
n
e
s
tr
ateg
y
th
at
ca
n
b
e
us
ed
to
p
r
eser
v
e
o
p
tim
al
alg
o
r
ith
m
p
er
f
o
r
m
an
ce
.
T
h
er
e
a
r
e
two
ap
p
r
o
ac
h
es
to
d
im
en
s
io
n
ality
r
ed
u
ctio
n
:
f
ea
tu
r
e
ex
tr
ac
tio
n
an
d
f
ea
tu
r
e
s
elec
tio
n
[
1
0
]
–
[
1
3
]
.
Acc
o
r
d
in
g
to
C
h
r
is
ty
et
a
l
.
[
1
4
]
,
s
eg
m
en
tatio
n
h
el
p
s
o
r
g
a
n
izatio
n
s
b
etter
u
n
d
er
s
tan
d
to
cu
s
to
m
er
n
ee
d
s
an
d
id
en
tify
f
u
tu
r
e
cu
s
to
m
er
s
ch
ar
ac
ter
is
tics
.
T
h
e
r
esear
ch
er
s
u
s
ed
r
ec
en
cy
,
f
r
e
q
u
en
cy
,
m
o
n
etar
y
(
R
F
M
)
a
n
a
l
y
s
i
s
f
o
r
s
e
g
m
e
n
t
a
tio
n
a
n
d
a
d
a
p
t
e
d
i
t
t
o
o
t
h
e
r
a
l
g
o
r
i
t
h
m
s
li
k
e
K
-
m
e
a
n
s
a
n
d
R
F
M
K
-
m
e
a
n
s
[
1
5
]
–
[
1
7
]
.
I
n
e
-
c
o
m
m
er
ce
,
th
e
u
s
er
b
e
h
av
io
r
will
b
e
o
b
s
er
v
ed
in
th
eir
ac
tiv
ities
u
s
in
g
web
s
ite
[
1
8
]
–
[
2
0
]
.
So
m
e
co
m
p
an
ies
o
f
ten
s
eg
m
e
n
t
cu
s
to
m
er
b
ased
o
n
e
-
c
o
m
m
er
ce
ch
ec
k
o
u
t
ab
an
d
o
n
m
en
t
r
ates.
C
o
m
p
an
ies
m
ig
h
t
o
f
f
er
d
is
co
u
n
ts
to
en
co
u
r
ag
e
c
u
s
to
m
er
s
to
m
a
k
e
a
p
u
r
c
h
ase
r
ath
er
th
an
f
illi
n
g
t
h
eir
ca
r
t
with
o
u
t
c
h
ec
k
o
u
t,
an
d
also
co
m
p
an
ies
p
r
o
v
id
e
s
u
p
p
o
r
t
to
cu
s
to
m
er
s
with
r
e
v
iew
q
u
esti
o
n
s
ab
o
u
t
p
r
o
d
u
cts,
p
a
y
m
en
t,
an
d
q
u
ali
ty
.
T
h
e
d
ata
wh
ich
co
n
tain
s
p
u
r
c
h
ased
p
r
o
d
u
cts
ca
n
b
e
f
u
r
th
er
an
aly
ze
d
to
d
e
v
elo
p
p
r
o
d
u
ct
p
r
o
m
o
tio
n
,
p
r
o
d
u
ct
o
p
er
atio
n
s
tr
ateg
ies
f
o
r
c
u
s
to
m
er
s
in
ea
c
h
cl
u
s
ter
[
2
1
]
,
[
2
2
]
.
C
lu
s
ter
in
g
is
t
h
e
p
r
o
ce
d
u
r
e
o
f
g
r
o
u
p
in
g
a
s
et
o
f
da
ta
in
to
g
r
o
u
p
s
th
at
e
x
h
ib
it
s
im
ilar
ch
ar
ac
ter
is
tics
[
2
3
]
.
A
clu
s
ter
is
a
co
llectio
n
o
f
o
b
s
er
v
at
io
n
s
th
at
a
r
e
s
im
ilar
with
in
th
e
s
am
e
clu
s
te
r
b
u
t d
if
f
er
f
r
o
m
o
b
s
er
v
atio
n
s
in
o
th
er
clu
s
ter
s
.
K
-
m
ea
n
s
clu
s
ter
in
g
is
o
n
e
o
f
th
e
clu
s
ter
in
g
m
eth
o
d
s
wh
ich
s
im
p
le
an
d
p
o
p
u
lar
way
to
s
eg
m
e
n
t
a
d
ataset
in
to
K
d
if
f
er
en
t
clu
s
ter
s
.
T
o
p
er
f
o
r
m
K
-
m
ea
n
s
clu
s
ter
in
g
,
th
e
n
u
m
b
er
o
f
K
(
clu
s
ter
)
s
h
o
u
ld
b
e
d
e
ter
m
in
ed
.
Dete
r
m
in
in
g
th
e
n
u
m
b
er
o
f
clu
s
ter
s
is
cr
u
cial
p
ar
t
f
o
r
m
an
a
g
in
g
th
e
s
e
g
r
o
u
p
s
.
T
h
e
co
n
tr
ib
u
tio
n
i
s
u
s
in
g
p
r
in
cip
al
co
m
p
o
n
e
n
t
an
aly
s
is
(
PC
A
)
as
d
im
en
s
io
n
ality
r
ed
u
ctio
n
in
h
i
g
h
d
im
e
n
s
io
n
ality
d
ata
as e
x
p
l
ain
ed
in
th
e
h
ig
h
lig
h
ted
s
en
ten
ce
I
n
th
is
s
tu
d
y
,
we
ar
e
m
o
r
e
f
o
cu
s
es
o
n
th
e
ap
p
licatio
n
o
f
K
-
m
ea
n
s
f
o
r
cu
s
to
m
e
r
s
eg
m
e
n
tatio
n
b
y
u
s
in
g
PC
A
as
a
d
im
en
s
io
n
al
ity
r
ed
u
ctio
n
s
tep
,
to
h
an
d
le
th
e
p
r
o
b
lem
o
f
h
ig
h
d
im
en
s
io
n
al
d
ata
th
at
ca
n
d
eg
r
ad
e
th
e
q
u
ality
o
f
s
eg
m
e
n
tatio
n
.
T
h
er
ef
o
r
e,
th
e
m
ain
co
n
tr
ib
u
tio
n
o
f
t
h
is
r
esear
ch
i
s
th
e
ap
p
licatio
n
o
f
d
im
en
s
io
n
ality
r
e
d
u
ctio
n
tec
h
n
iq
u
e
(
PC
A)
an
d
th
e
u
s
e
o
f
K
-
m
ea
n
s
f
o
r
cu
s
to
m
er
s
eg
m
en
tatio
n
b
ased
o
n
p
u
r
ch
asin
g
b
eh
av
i
o
r
.
W
h
ile
th
e
p
r
ev
io
u
s
s
tu
d
ies
h
av
e
ex
p
l
o
r
ed
th
e
im
p
ac
t
o
f
K
-
m
ea
n
s
clu
s
ter
in
g
b
y
u
s
in
g
f
ix
ed
k
-
clu
s
ter
ed
.
T
h
e
p
ar
a
m
eter
in
cu
s
to
m
er
d
em
o
g
r
ap
h
y
an
d
b
e
h
av
io
r
s
u
ch
as
ag
e,
an
n
u
al
in
co
m
e,
s
p
en
d
in
g
s
co
r
e
,
p
u
r
ch
ase,
h
is
to
r
y
,
a
n
d
q
u
an
tity
[
2
4
]
.
2.
M
E
T
H
O
D
T
h
is
s
ec
tio
n
p
r
esen
ts
th
e
m
eth
o
d
o
lo
g
y
u
s
ed
in
th
is
s
tu
d
y
.
T
h
e
aim
is
to
s
eg
m
en
t
cu
s
to
m
er
s
d
ep
en
d
i
n
g
o
n
th
eir
b
eh
a
v
io
r
i
n
p
u
r
c
h
asin
g
tr
an
s
ac
tio
n
.
W
e
p
er
f
o
r
m
ed
th
r
ee
s
tag
es
n
am
e
ly
d
ata
an
aly
s
is
to
id
en
tify
th
e
o
u
tlier
f
r
o
m
d
at
aset
an
d
to
ch
ec
k
m
is
s
in
g
v
alu
e.
Seco
n
d
s
tag
e
was
PC
A
im
p
l
em
en
tatio
n
t
o
r
ed
u
ce
d
ata,
a
n
d
th
e
th
ir
d
K
-
m
ea
n
s
alg
o
r
ith
m
im
p
lem
e
n
tatio
n
th
en
an
aly
ze
d
th
e
in
ter
p
r
e
tatio
n
o
f
clu
s
ter
in
g
r
esu
lt
[
2
4
]
.
T
h
e
r
esear
c
h
m
eth
o
d
as sh
o
wn
as Fig
u
r
e
1
.
2
.
1
.
Da
t
a
a
na
ly
s
is
I
n
th
is
s
tu
d
y
we
u
s
ed
d
ata
co
n
tain
in
f
o
r
m
atio
n
ab
o
u
t
2
,
5
0
0
in
s
tan
ce
s
co
n
s
u
m
er
y
ea
r
ly
tr
an
s
ac
tio
n
en
co
m
p
ass
es
9
f
ea
tu
r
es
o
f
y
ea
r
ly
s
eg
m
en
tatio
n
s
u
c
h
as
ag
e,
an
n
u
al
in
co
m
e,
s
p
en
d
in
g
s
co
r
e,
p
u
r
c
h
ase
h
is
to
r
y
,
p
r
o
d
u
ct
ca
teg
o
r
y
,
q
u
an
tity
,
u
n
it
p
r
ice,
to
tal
p
r
ice,
an
d
p
a
y
m
en
t
m
eth
o
d
.
B
u
t
in
th
is
s
tu
d
y
,
we
co
n
s
id
er
ed
6
f
ea
tu
r
es
in
clu
d
in
g
a
g
e,
an
n
u
al
in
co
m
e,
s
p
en
d
in
g
s
co
r
e
,
p
u
r
ch
ase
h
is
to
r
y
,
p
r
o
d
u
ct
ca
te
g
o
r
y
,
a
n
d
p
ay
m
en
t
m
eth
o
d
.
I
n
d
ata
an
aly
s
is
,
f
ir
s
t
we
id
en
tifie
d
th
e
o
u
tlier
f
r
o
m
f
o
u
r
co
l
u
m
n
in
d
ata
d
is
tr
ib
u
tio
n
,
n
am
ely
a
g
e,
an
n
u
al
in
co
m
e,
s
p
en
d
i
n
g
s
c
o
r
e
an
d
p
u
r
c
h
ase
h
is
to
r
y
.
O
u
tlier
s
ar
e
o
b
s
er
v
atio
n
s
s
tatis
tically
s
ig
n
if
ican
t
d
if
f
er
en
t
f
r
o
m
th
e
b
u
l
k
o
f
th
e
d
ata.
T
h
e
p
r
o
ce
s
s
to
id
en
tify
th
e
o
u
tlier
s
u
s
in
g
b
o
x
p
lo
t
wi
th
th
e
i
n
ter
q
u
a
r
tile
r
an
g
e
(
I
QR
)
[
2
5
]
. T
h
e
id
e
n
tify
in
g
o
u
tlier
p
r
o
ce
s
s
as e
x
p
lain
e
d
b
elo
w:
C
alcu
late
th
e
f
ir
s
t q
u
ar
tile (
Q1
)
an
d
th
i
r
d
q
u
ar
tile (
Q3
)
:
a)
Q
1 i
s t
he 2
5
th
per
cent
i
l
e of
t
h
e dat
a (
t
he val
ue
bel
o
w
w
hi
c
h 25%
of
t
he
dat
a f
al
l
s)
.
b)
Q
3 i
s t
he 7
5
th
per
cent
i
l
e of
t
h
e dat
a (
t
he val
ue
bel
o
w
w
hi
c
h
75%
of
t
he
dat
a f
al
l
s)
.
i)
C
alcu
late
th
e
I
QR
:
=
3
−
1
(
1
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
E
n
h
a
n
cin
g
ma
r
ke
tin
g
efficien
cy
th
r
o
u
g
h
d
a
ta
-
d
r
iven
cu
s
to
mer seg
men
ta
tio
n
…
(
F
a
n
i
n
d
i
a
P
u
r
n
a
ma
s
a
r
i
)
1401
ii)
Dete
r
m
in
e
th
e
lo
wer
b
o
u
n
d
an
d
u
p
p
e
r
b
o
u
n
d
:
=
1
−
1
.
5
×
=
3
+
1
.
5
×
(
2
)
iii)
I
d
en
tify
o
u
tlier
s
iv
)
Po
s
itio
n
d
ata
p
o
in
t le
s
s
th
an
th
e
lo
wer
b
o
u
n
d
o
r
g
r
ea
ter
th
a
n
th
e
u
p
p
e
r
b
o
u
n
d
is
co
n
s
id
er
ed
an
o
u
tlier
.
B
ased
o
n
th
e
b
o
x
p
lo
t
v
is
u
ali
za
tio
n
th
e
o
u
tlier
was
n
o
t
f
o
u
n
d
as
Fig
u
r
e
2
.
T
h
e
n
we
p
er
f
o
r
m
ed
d
ata
p
r
e
-
p
r
o
ce
s
s
in
g
is
u
s
ed
f
o
r
f
u
r
th
er
an
aly
s
is
o
n
PC
A
an
d
K
-
m
ea
n
s
clu
s
ter
in
g
.
T
h
e
d
ata
wer
e
f
ir
s
t
n
o
r
m
aliz
ed
u
s
in
g
m
in
–
m
a
x
n
o
r
m
aliza
tio
n
in
to
th
e
r
a
n
g
e
[
0
,
1
]
.
No
r
m
al
ized
d
ata
m
ea
n
s
ea
ch
f
ea
tu
r
e
o
f
d
ata
h
as
eq
u
al
weig
h
t,
wh
ich
p
r
ev
en
ts
lar
g
e
r
-
s
ca
le
f
ea
tu
r
es f
r
o
m
d
o
m
in
atin
g
th
e
an
aly
s
is
r
esu
lts
.
Fig
u
r
e
1
.
R
esear
ch
m
eth
o
d
Fig
u
r
e
2
.
Data
d
is
tr
ib
u
tio
n
s
to
id
en
tify
th
e
o
u
tlier
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
3
9
,
No
.
2
,
Au
g
u
s
t
20
25
:
1
3
9
9
-
1
410
1402
2
.
2
.
P
rincipa
l
co
m
po
nent
a
na
ly
s
is
PC
A
is
u
s
ed
to
r
ed
u
ce
th
e
co
m
p
lex
ity
o
f
d
ata.
Af
ter
id
e
n
tif
y
in
g
th
e
o
u
tlier
,
we
p
er
f
o
r
m
e
d
PC
A
d
u
e
to
K
-
m
ea
n
s
cl
u
s
ter
in
g
is
s
en
s
itiv
ity
with
h
ig
h
d
im
en
s
io
n
al
d
ata.
First,
we
ca
lcu
late
th
e
co
v
ar
ian
ce
to
id
en
tif
y
am
o
n
g
f
ea
t
u
r
es
in
d
ata.
T
h
e
n
p
er
f
o
r
m
eig
e
n
d
ec
o
m
p
o
s
itio
n
wh
ich
a
tech
n
iq
u
e
in
lin
e
ar
alg
eb
r
a
u
s
ed
t
o
d
ec
o
m
p
o
s
e
m
atr
ices in
to
eig
e
n
v
ec
to
r
s
an
d
eig
en
v
al
u
es.
T
o
ap
p
ly
PC
A
we
u
s
ed
f
o
r
m
u
las as
[
2
6
]
:
i)
Data
c
en
ter
in
g
=
−
(
3
)
W
h
er
e:
is
d
ata.
m
atr
ix
m
ea
n
o
f
d
ata
in
ea
c
h
f
ea
tu
r
e
ii)
C
o
v
ar
ian
ce
m
atr
ix
=
1
−
1
(
4
)
W
h
er
e:
is
am
o
u
n
t o
f
d
ata,
is
d
ata
th
at
h
as b
ee
n
r
e
d
u
ce
d
to
th
e
m
ea
n
o
f
d
ata
C
o
v
ar
ian
ce
is
a
m
ea
s
u
r
e
o
f
h
o
w
well
two
v
ar
iab
les
co
r
r
elate
with
o
n
e
a
n
o
th
er
.
I
ts
d
iag
o
n
a
l
co
n
tain
s
ev
er
y
v
ar
ia
n
ce
an
d
c
o
n
tain
s
a
ll
p
o
ten
tial
co
v
ar
ian
ce
p
air
s
b
etwe
en
th
e
m
v
ar
iab
les,
wh
il
e
s
m
aller
v
ar
ian
ce
v
alu
es
m
ay
in
d
icate
th
e
n
o
is
e
in
th
ese
d
ata
th
e
h
u
g
e
v
a
r
ian
c
e
v
alu
es
ar
e
s
ig
n
if
ican
t
s
in
ce
t
h
ey
c
o
r
r
elate
t
o
th
e
in
tr
ig
u
in
g
d
y
n
a
m
ics in
th
ese
d
ata.
iii)
E
ig
en
d
ec
o
m
p
o
s
itio
n
=
Λ
(
5
)
W
h
er
e:
is
eig
en
v
ec
to
r
m
atr
ix
(
p
r
in
ci
p
al
co
m
p
o
n
en
ts
)
,
Λ
is
a
d
iag
o
n
al
m
atr
i
x
co
n
tain
i
n
g
eig
en
v
al
u
es (
th
e
v
ar
ia
n
ce
ex
p
lain
ed
b
y
ea
ch
p
r
in
cip
al
c
o
m
p
o
n
en
t)
.
iv
)
Pro
jectio
n
in
to
p
r
in
cip
al
co
m
p
o
n
en
ts
=
(
6
)
W
h
er
e:
Z
is
d
ata
th
at
h
as b
ee
n
p
r
o
jecte
d
to
a
n
ew
s
p
ac
e.
I
n
th
is
s
tu
d
y
,
we
r
u
n
p
r
in
cip
al
co
m
p
o
n
en
ts
as
2
,
t
h
ey
a
r
e
th
e
two
p
r
im
ar
y
co
m
p
o
n
e
n
ts
u
s
ed
b
y
PC
A
f
o
r
r
ed
u
cin
g
th
e
d
im
en
s
io
n
o
f
th
e
d
ata.
R
ed
u
cin
g
th
e
m
ea
n
v
alu
es
f
r
o
m
th
e
in
itial
d
ataset
is
th
e
f
ir
s
t
s
tep
in
th
e
PC
A
an
aly
s
is
p
r
o
ce
s
s
,
in
o
r
d
er
to
p
r
o
v
id
e
th
e
o
r
i
g
in
al
d
ataset'
s
an
aly
s
i
s
p
h
ase
eq
u
al
weig
h
ts
an
d
ap
p
r
o
p
r
iate
n
o
r
m
aliza
tio
n
.
I
n
Fig
u
r
e
3
s
h
o
ws
th
at
th
e
P
C
1
an
d
PC
2
ax
es
r
ep
r
esen
t
th
e
two
o
f
PC
A
co
m
p
o
n
en
ts
,
in
d
icate
s
th
at
th
e
s
p
r
ea
d
an
d
d
is
tr
ib
u
tio
n
o
f
cu
s
to
m
er
s
in
a
s
im
p
ler
f
ea
tu
r
e
s
p
ac
e.
B
ased
o
n
s
u
itab
ilit
y
an
aly
s
is
,
th
e
d
ata
i
s
d
iv
id
ed
in
to
s
ev
er
al
clu
s
ter
s
h
av
e
a
s
tr
o
n
g
r
elatio
n
s
h
ip
with
o
th
er
cl
u
s
ter
.
I
n
co
n
tr
ast,
o
th
er
clu
s
ter
s
ap
p
ea
r
m
o
r
e
d
is
p
er
s
ed
in
d
icate
s
th
at
g
r
ea
ter
v
ar
iatio
n
am
o
n
g
cu
s
to
m
er
s
with
in
th
o
s
e
clu
s
ter
s
.
2
.
3
.
I
m
ple
m
ent
a
t
io
n o
f
K
-
mea
ns
clus
t
er
ing
T
h
e
d
ata
we
r
e
f
i
r
s
t
n
o
r
m
alize
d
u
s
in
g
m
in
–
m
ax
n
o
r
m
aliza
tio
n
in
to
th
e
r
an
g
e
[
0
,
1
]
.
T
h
e
d
et
er
m
in
atio
n
n
u
m
b
er
o
f
clu
s
ter
s
was
n
o
t
f
i
x
ed
in
ea
r
ly
s
tag
e,
b
u
t
u
s
in
g
t
h
e
s
ilh
o
u
tte
co
ef
f
icien
t.
T
h
e
f
i
r
s
t
ce
n
ter
is
ch
o
s
en
b
y
r
an
d
o
m
,
an
d
f
o
llo
win
g
p
o
i
n
ts
ar
e
ch
o
s
en
with
a
p
r
o
b
a
b
il
ity
p
r
o
p
o
r
tio
n
al
t
o
th
e
s
q
u
ar
e
d
d
is
tan
ce
f
r
o
m
th
e
n
ea
r
est
ce
n
ter
.
T
h
is
s
tu
d
y
u
s
ed
1
0
×
r
u
n
f
r
o
m
r
an
d
o
m
in
itial
p
o
s
itio
n
s
th
e
r
esu
lt
with
th
e
l
o
west
with
in
-
clu
s
ter
s
u
m
o
f
s
q
u
ar
es will b
e
u
s
ed
a
n
d
m
ax
iter
atio
n
3
0
0
i
n
o
r
d
er
t
o
th
e
alg
o
r
ith
m
r
ea
ch
co
n
v
er
g
en
ce
co
n
d
itio
n
.
T
h
e
iter
ativ
e
p
r
o
ce
s
s
aim
s
to
r
ed
u
ce
th
e
d
is
tan
ce
b
etwe
en
d
ata
p
o
in
ts
an
d
ce
n
tr
o
id
s
in
ea
c
h
clu
s
ter
,
m
o
v
in
g
th
e
ce
n
tr
o
id
s
to
b
etter
d
escr
ib
e
th
e
clu
s
ter
ce
n
ter
.
T
h
e
K
-
m
ea
n
s
s
tep
s
as e
x
p
lain
ed
b
elo
w
[
2
4
]
:
i)
C
en
tr
o
id
i
n
itializatio
n
Select
k
in
itial c
en
tr
o
id
s
r
an
d
o
m
ly
f
r
o
m
th
e
d
ataset,
wh
er
e
r
ep
r
esen
ts
th
e
r
eq
u
ir
e
d
n
u
m
b
er
o
f
clu
s
ter
s
.
ii)
C
alcu
late
E
u
clid
ea
n
d
is
tan
ce
Fo
r
ev
er
y
d
ata
p
o
in
t
,
co
m
p
u
te
th
e
E
u
clid
ea
n
d
is
tan
ce
to
ea
ch
ce
n
tr
o
id
.
T
h
e
eq
u
atio
n
f
o
r
E
u
clid
ea
n
d
is
tan
ce
is
:
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
E
n
h
a
n
cin
g
ma
r
ke
tin
g
efficien
cy
th
r
o
u
g
h
d
a
ta
-
d
r
iven
cu
s
to
mer seg
men
ta
tio
n
…
(
F
a
n
i
n
d
i
a
P
u
r
n
a
ma
s
a
r
i
)
1403
(
,
)
=
√
∑
(
−
)
2
=
1
(
7
)
W
h
er
e
:
is
th
e
d
ata
p
o
in
t to
-
,
is
th
e
ce
n
tr
o
id
to
-
,
an
d
is
th
e
v
alu
e
o
f
th
e
co
m
p
o
n
en
t
-
fr
o
m
d
ata
an
d
ce
n
tr
o
id
.
iii)
Ass
ig
n
d
ata
p
o
in
ts
to
n
ea
r
est c
lu
s
ter
s
E
ac
h
d
ata
p
o
i
n
t
will
b
e
all
o
ca
ted
to
th
e
clu
s
ter
wh
o
s
e
ce
n
tr
o
id
ex
h
ib
its
th
e
m
in
im
a
l
d
is
tan
ce
,
as
d
eter
m
in
ed
b
y
th
e
E
u
clid
ea
n
d
is
tan
ce
co
m
p
u
tatio
n
.
iv
)
Up
d
ate
C
en
tr
o
id
Po
s
itio
n
Af
ter
all
d
ata
h
as b
ee
n
all
o
ca
ted
to
clu
s
ter
s
,
r
ev
is
e
th
e
p
o
s
itio
n
o
f
ea
ch
ce
n
t
r
o
id
b
y
co
m
p
u
t
in
g
th
e
m
ea
n
o
f
all
d
ata
p
o
in
ts
co
n
tain
e
d
in
s
id
e
th
at
clu
s
ter
:
=
1
|
|
∑
∈
(
8
)
W
h
er
e
:
is
a
s
et
o
f
d
ata
p
o
in
ts
class
if
ied
in
to
clu
s
ter
s
,
is
th
e
n
ew
ce
n
tr
o
id
f
o
r
th
e
clu
s
ter
.
v)
R
ep
ea
t
p
r
o
ce
s
s
C
o
n
tin
u
e
ex
ec
u
tin
g
s
tep
s
2
to
4
u
n
til
th
e
ce
n
tr
o
i
d
s
ex
h
ib
it
n
eg
lig
ib
le
ch
an
g
es
o
r
a
p
r
e
d
ete
r
m
in
ed
iter
atio
n
lim
it
is
attain
ed
.
T
h
e
K
-
m
ea
n
s
tech
n
iq
u
e
g
en
er
ates
clu
s
ter
s
,
with
ea
ch
clu
s
ter
in
clu
d
in
g
d
ata
p
o
in
ts
n
ea
r
est to
th
eir
co
r
r
esp
o
n
d
in
g
ce
n
tr
o
id
s
.
Fig
u
r
e
3
.
Op
tim
al
c
o
m
p
o
n
en
t
in
PC
A
an
aly
s
is
T
h
en
,
we
ev
alu
ate
th
e
clu
s
te
r
r
esu
lt
b
y
u
tili
zin
g
th
e
in
er
ti
a
an
d
s
ilh
o
u
ette
s
co
r
e.
T
h
e
i
n
er
tia
an
d
s
ilh
o
u
ette
s
co
r
e
ca
lcu
latio
n
s
ar
e
p
er
f
o
r
m
ed
,
wh
er
e
in
er
tia
m
ea
s
u
r
es
th
e
co
m
p
ac
tn
ess
o
f
th
e
clu
s
ter
.
T
h
e
lo
we
r
th
e
in
er
tia
v
al
u
e,
th
e
b
etter
t
h
e
clu
s
ter
in
g
in
ter
m
s
o
f
co
m
p
ac
tn
ess
(
d
ata
p
o
in
ts
ar
e
clo
s
er
to
th
eir
r
esp
ec
tiv
e
clu
s
ter
c
en
tr
o
id
s
)
.
T
h
e
h
i
g
h
er
th
e
Sil
h
o
u
ette
c
o
ef
f
icien
t
ap
p
r
o
ac
h
to
1
in
d
icate
s
th
e
b
etter
t
h
e
clu
s
ter
w
h
er
ea
s
s
co
r
e
clo
s
es to
0
in
d
icate
o
v
er
lap
p
in
g
clu
s
ter
s
,
ca
n
b
e
s
ee
n
i
n
T
ab
le
1
.
T
ab
le
1
.
C
lu
s
ter
ev
alu
atio
n
r
esu
lts
N
u
mb
e
r
o
f
c
l
u
s
t
e
r
s (
K
)
I
n
e
r
t
i
a
S
i
l
h
o
u
e
t
t
e
s
c
o
r
e
2
1
0
.
3
9
4
.
1
7
9
0
.
3
1
1
8
9
6
3
1
.
4
5
5
.
6
5
0
0
.
4
8
6
3
6
6
4
0
.
5
0
4
4
5
0
0
.
2
4
7
3
9
9
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
3
9
,
No
.
2
,
Au
g
u
s
t
20
25
:
1
3
9
9
-
1
410
1404
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
r
esu
lts
o
f
t
h
is
s
tu
d
y
is
h
i
g
h
lig
h
t
t
h
e
s
ig
n
if
ica
n
t
im
p
ac
t
th
at
ML
-
b
ased
c
u
s
to
m
er
s
eg
m
en
tatio
n
h
as
o
n
m
ar
k
etin
g
ef
f
icien
c
y
.
I
n
th
is
s
ec
tio
n
,
we
will
d
is
cu
s
s
th
e
k
ey
f
in
d
in
g
s
o
f
cl
u
s
ter
in
g
r
esu
lt
b
ased
o
n
an
aly
s
is
o
f
PC
A
,
d
i
s
tr
ib
u
tio
n
o
f
clu
s
ter
in
g
r
esu
lt,
in
ter
p
r
etat
io
n
o
f
clu
s
ter
in
g
r
esu
lts
an
d
b
u
s
in
ess
in
s
ig
h
t f
r
o
m
clu
s
ter
.
3
.
1
.
Da
t
a
a
na
ly
s
is
W
e
p
er
f
o
r
m
d
escr
ip
tiv
e
s
tatis
tic
to
h
elp
u
n
d
er
s
tan
d
an
d
ex
p
l
o
r
e
th
e
d
ata.
Descr
ip
tiv
e
s
tatis
tics
,
s
u
ch
as
m
ea
n
,
m
ed
ian
,
s
tan
d
ar
d
d
ev
iatio
n
,
m
i
n
im
u
m
,
an
d
m
ax
im
u
m
v
al
u
es,
p
r
o
v
id
e
a
n
o
v
er
v
iew
o
f
th
e
d
is
tr
ib
u
tio
n
,
an
d
ce
n
tr
al
ten
d
e
n
cy
.
T
ab
le
2
s
h
o
w
th
e
d
escr
ip
tiv
e
s
tatis
t
ics
o
f
th
e
d
ataset
u
s
ed
in
th
is
s
tu
d
y
.
T
h
is
d
ataset
co
n
s
is
ts
o
f
2
,
5
0
0
cu
s
to
m
er
d
ata
en
tr
ies
c
o
v
e
r
in
g
6
f
ea
tu
r
es,
in
cl
u
d
in
g
ag
e,
an
n
u
al
in
c
o
m
e,
s
p
en
d
in
g
s
co
r
e,
p
u
r
ch
ase
h
is
to
r
y
,
p
r
o
d
u
ct
ca
teg
o
r
y
,
an
d
p
ay
m
en
t
m
eth
o
d
.
T
h
is
d
at
a
is
r
ea
l
cu
s
to
m
er
tr
an
s
ac
tio
n
d
ata
co
llected
f
o
r
s
eg
m
en
tatio
n
a
n
aly
s
is
b
ased
o
n
p
u
r
ch
ase
b
eh
av
io
r
.
T
ab
le
2
.
Descr
ip
tiv
e
s
tatis
tics
f
o
r
d
ata
F
e
a
t
u
r
e
C
o
u
n
t
M
e
a
n
S
t
a
n
d
a
r
d
d
e
v
i
a
t
i
o
n
M
i
n
M
a
x
A
g
e
4
0
9
6
4
4
.
0
5
1
5
.
0
8
1
8
.
0
6
9
.
0
A
n
n
u
a
l
i
n
c
o
m
e
4
0
9
6
6
9
9
1
0
.
3
7
2
8
8
7
0
.
5
8
2
0
0
6
0
1
1
9
9
9
3
.
0
S
p
e
n
d
i
n
g
s
c
o
r
e
4
0
9
6
5
0
.
9
7
2
8
.
8
6
1
.
0
1
0
0
.
0
P
u
r
c
h
a
se
H
i
s
t
o
r
y
4
0
9
6
5
0
.
3
3
2
8
.
3
2
1
.
0
9
9
.
0
Q
u
a
n
t
i
t
y
4
0
9
6
5
.
0
3
1
.
9
9
1
.
0
9
.
0
3
.
2
.
P
rincipa
l
co
m
po
nent
a
na
ly
s
is
T
wo
co
m
p
o
n
e
n
ts
o
f
PC
A
(
P
C
1
an
d
PC
2
)
ex
p
lain
ab
o
u
t
4
2
.
5
7
%
o
f
th
e
to
tal
v
ar
ian
ce
in
th
e
d
ata.
T
h
is
m
ea
n
s
th
at
d
esp
ite
u
s
in
g
two
d
im
en
s
io
n
s
,
it
s
till
r
etain
s
q
u
ite
s
ig
n
if
ican
t
in
f
o
r
m
atio
n
f
r
o
m
th
e
o
r
ig
in
a
l
d
ata.
T
h
e
h
ig
h
v
ar
ia
n
ce
ex
p
lain
ed
b
y
th
ese
two
c
o
m
p
o
n
en
ts
in
d
icate
s
th
at
th
e
d
ata
ca
n
b
e
ef
f
ec
tiv
ely
r
ep
r
esen
ted
i
n
two
d
im
en
s
io
n
s
f
o
r
v
is
u
al
a
n
aly
s
is
an
d
in
ter
p
r
etatio
n
.
E
ac
h
v
alu
e
in
PC
1
an
d
PC
2
r
ep
r
esen
ts
a
p
r
o
jectio
n
o
f
th
e
cu
s
to
m
er
d
ata
in
to
th
e
n
ew
f
ea
tu
r
e
s
p
ac
e.
Po
s
itiv
e
o
r
n
eg
ativ
e
v
alu
e
s
in
PC
1
an
d
PC
2
in
d
icate
wh
er
e
th
e
d
ata
is
lo
ca
ted
r
elativ
e
to
th
e
m
ea
n
in
th
at
co
m
p
o
n
en
t.
Hig
h
n
eg
ativ
e
v
alu
es
in
PC
1
an
d
p
o
s
itiv
e
in
PC
2
(
f
ir
s
t
r
o
w,
P
C
1
=
-
2
.
2
6
4
7
3
8
,
PC
2
=0
.
8
2
9
7
7
7
)
in
d
icate
th
at
th
e
d
ata
h
as
ch
ar
ac
ter
is
tics
f
ar
b
elo
w
th
e
m
ea
n
f
o
r
th
e
PC
1
d
im
en
s
io
n
b
u
t
ab
o
v
e
th
e
m
ea
n
f
o
r
th
e
PC
2
d
im
en
s
io
n
.
C
o
n
v
e
r
s
ely
,
v
alu
es
clo
s
er
t
o
0
i
n
b
o
t
h
c
o
m
p
o
n
e
n
t
s
i
n
d
i
c
at
e
t
h
a
t
t
h
e
d
at
a
h
a
s
c
h
a
r
a
c
t
e
r
i
s
tic
s
c
l
o
s
e
t
o
t
h
e
m
ea
n
.
c
a
n
b
e
s
ee
n
i
n
t
h
e
T
a
b
l
e
3
.
T
ab
le
3
.
PC
A
a
n
aly
s
is
with
tw
o
co
m
p
o
n
en
ts
(
PC
1
an
d
PC
2
)
P
C
1
P
C
2
-
2
.
2
6
4
.
7
3
8
0
.
8
2
9
7
7
7
0
.
4
3
9
0
6
7
1
.
3
9
3
.
8
3
3
0
.
9
4
6
6
2
9
1
.
2
4
7
.
0
8
8
-
2
.
0
9
2
.
5
6
8
-
1
.
0
8
4
.
7
9
7
0
.
0
8
3
6
9
7
-
1
.
7
8
2
.
0
1
9
3
.
3
.
Appl
ica
t
io
n o
f
K
-
m
e
a
ns
a
lg
o
rit
hm
C
lu
s
ter
i
s
a
lab
el
to
id
en
tify
ea
ch
clu
s
ter
in
th
e
K
-
m
ea
n
s
an
aly
s
is
.
I
n
C
en
tr
o
id
X
an
d
C
en
t
r
o
id
Y,
th
e
co
o
r
d
in
ates
o
f
th
e
in
itial
ce
n
t
r
o
id
ar
e
in
itialized
r
a
n
d
o
m
ly
.
T
h
is
ce
n
tr
o
id
is
th
e
in
itial
c
en
tr
e
p
o
in
t
o
f
ea
ch
clu
s
ter
an
d
will
th
en
b
e
ad
ju
s
ted
th
r
o
u
g
h
th
e
iter
atio
n
o
f
th
e
K
-
m
ea
n
s
as
s
h
o
wn
as
T
a
b
le
4
.
T
h
e
clo
s
est
clu
s
ter
ce
n
tr
o
id
b
ased
o
n
E
u
cl
id
ea
n
d
is
t
an
ce
d
ata
is
ass
ig
n
e
d
to
th
is
clu
s
ter
in
th
e
in
itial
it
er
atio
n
as
s
h
o
wn
as
T
ab
le
5
.
T
ab
le
6
ex
p
lain
s
th
e
ce
n
tr
o
id
is
u
p
d
ated
af
ter
th
e
d
ata
is
ass
ig
n
ed
to
th
e
clo
s
est
clu
s
ter
;
th
e
ce
n
tr
o
id
p
o
s
itio
n
is
r
ec
alcu
lated
b
ase
d
o
n
th
e
a
v
er
ag
e
p
o
s
itio
n
o
f
all
d
ata
p
o
in
ts
in
t
h
e
clu
s
te
r
.
C
en
tr
o
id
X
an
d
C
en
tr
o
id
Y,
th
e
n
ew
ce
n
tr
o
id
p
o
s
itio
n
in
PC
A
co
o
r
d
in
ates (
PC
1
an
d
PC
2
)
.
W
e
f
o
u
n
d
th
at
th
e
ch
an
g
in
g
i
n
th
e
av
er
ag
e
d
is
tan
ce
o
f
ea
ch
ce
n
tr
o
id
f
r
o
m
th
e
p
r
ev
io
u
s
iter
atio
n
to
th
e
cu
r
r
e
n
t
iter
atio
n
in
th
e
co
n
v
er
g
in
g
p
r
o
ce
s
s
.
T
h
e
iter
atio
n
p
r
o
ce
s
s
en
d
s
wh
e
n
th
e
ce
n
tr
o
id
ch
an
g
e
b
ec
o
m
es
in
to
s
m
all
v
al
u
e.
T
h
e
r
esu
lt
a
s
s
h
o
wn
as
T
a
b
le
7
.
T
h
e
f
in
a
l
r
esu
lts
ar
e
s
h
o
wn
as
T
ab
le
8
,
wh
er
e
th
e
f
in
al
clu
s
ter
in
g
r
esu
lts
s
h
o
w
th
at
af
ter
th
e
co
n
v
er
g
e
n
ce
iter
atio
n
,
ea
ch
d
at
a
p
o
in
t
is
ass
ig
n
ed
to
a
clu
s
ter
b
ased
o
n
its
p
r
o
x
im
ity
to
th
e
n
ea
r
est ce
n
tr
o
id
.
Fig
u
r
e
4
s
h
o
ws
th
e
r
esu
lts
o
f
K
-
m
ea
n
s
clu
s
ter
in
g
o
n
PC
A
d
ata
with
two
p
r
in
cip
al
c
o
m
p
o
n
en
ts
(
PC
1
an
d
PC
2
)
.
T
h
e
co
lo
r
ed
d
o
ts
r
ep
r
esen
t
d
ata
in
d
if
f
e
r
en
t
clu
s
ter
s
(
C
lu
s
ter
s
0
,
1
,
an
d
2
)
.
Di
f
f
er
en
t
d
o
t
s
h
ap
es
(
r
o
u
n
d
,
s
q
u
ar
e,
a
n
d
d
iam
o
n
d
)
ar
e
u
s
ed
to
d
is
tin
g
u
is
h
ea
ch
clu
s
ter
.
T
h
e
clu
s
ter
ce
n
tr
o
id
s
ar
e
s
h
o
wn
with
a
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
E
n
h
a
n
cin
g
ma
r
ke
tin
g
efficien
cy
th
r
o
u
g
h
d
a
ta
-
d
r
iven
cu
s
to
mer seg
men
ta
tio
n
…
(
F
a
n
i
n
d
i
a
P
u
r
n
a
ma
s
a
r
i
)
1405
lar
g
e
r
ed
'
X'
,
in
d
icatin
g
th
e
ce
n
tr
e
o
f
ea
ch
clu
s
ter
.
T
h
is
v
is
u
aliza
tio
n
p
r
o
v
id
es
a
b
etter
id
e
a
o
f
h
o
w
th
e
d
ata
is
d
is
tr
ib
u
ted
in
two
-
d
im
en
s
io
n
al
s
p
ac
e,
as
wel
l
as
h
o
w
ea
ch
clu
s
ter
is
ce
n
tr
ed
ar
o
u
n
d
t
h
e
ce
n
tr
o
id
.
Fu
tu
r
e
s
tu
d
ies
m
ay
ex
p
lo
r
e
to
an
o
t
h
er
m
eth
o
d
to
en
s
u
r
e
t
h
e
in
itial
ce
n
tr
o
id
is
m
o
r
e
d
is
p
er
s
ed
,
th
u
s
ac
h
iev
in
g
co
n
v
er
g
en
ce
.
T
ab
le
4
.
I
n
itializatio
n
o
f
c
e
n
tr
o
id
Cl
u
st
e
r
C
e
n
t
r
o
i
d
X
C
e
n
t
r
o
i
d
Y
0
0
.
5
1
.
2
1
-
1
.
0
0
.
7
2
1
.
5
-
1
.
5
T
ab
le
5
.
R
esu
lt o
f
d
eter
m
in
in
g
n
ea
r
est ce
n
tr
o
id
i
n
in
itial c
lu
s
ter
D
a
t
a
P
o
i
n
t
P
C
1
P
C
2
I
n
i
t
i
a
l
c
l
u
st
e
r
N
e
a
r
e
st
c
e
n
t
r
o
i
d
1
-
1
.
8
6
8
-
1
.
2
8
4
0
1
2
-
0
.
5
3
2
-
0
.
9
4
3
2
1
3
0
.
5
4
8
1
.
3
1
1
3
0
4
-
0
.
3
1
3
1
.
8
2
9
3
0
5
3
.
0
8
2
-
0
.
8
6
5
1
2
T
ab
le
6
.
Up
d
ated
ce
n
tr
o
id
r
es
u
lts
af
ter
in
itial iter
atio
n
Clu
ste
r
Ce
n
tro
i
d
X
Ce
n
tro
i
d
Y
0
0
.
1
1
7
5
1
.
5
7
0
1
-
1
.
2
0
0
-
0
.
9
4
3
2
2
.
3
1
5
-
0
.
8
6
5
T
ab
le
7
.
T
h
e
r
esu
lt in
co
n
v
er
g
e
p
r
o
ce
s
s
I
t
e
r
a
t
i
o
n
C
e
n
t
r
o
i
d
0
(
X
,
Y
)
C
e
n
t
r
o
i
d
1
(
X
,
Y
)
C
e
n
t
r
o
i
d
2
(
X
,
Y
)
A
f
t
e
r
c
h
a
n
g
i
n
g
1
(
0
.
1
1
7
5
,
1
.
5
7
0
)
(
-
1
.
2
0
0
,
-
0
.
9
4
3
)
(
2
.
3
1
5
,
-
0
.
8
6
5
)
0
.
3
5
2
(
0
.
1
2
0
0
,
1
.
5
6
8
)
(
-
1
.
1
8
0
,
-
0
.
9
5
0
)
(
2
.
3
2
0
,
-
0
.
8
6
0
)
0
.
0
2
...
...
...
...
...
F
i
n
a
l
(
0
.
1
2
1
0
,
1
.
5
6
7
)
(
-
1
.
1
7
8
,
-
0
.
9
5
1
)
(
2
.
3
2
2
,
-
0
.
8
5
9
)
<
0
.
0
1
T
ab
le
8
.
T
h
e
r
esu
lt in
f
in
al
clu
s
ter
in
g
D
a
t
a
p
o
i
n
t
P
C
1
P
C
2
F
i
n
a
l
c
l
u
st
e
r
1
-
1
.
8
6
8
-
1
.
2
8
4
1
2
-
0
.
5
3
2
-
0
.
9
4
3
1
3
0
.
5
4
8
1
.
3
1
1
0
4
-
0
.
3
1
3
1
.
8
2
9
0
5
3
.
0
8
2
-
0
.
8
6
5
2
Fig
u
r
e
4
.
E
n
h
an
ce
d
K
-
m
ea
n
s
clu
s
ter
in
g
v
is
u
aliza
tio
n
with
PC
A
c
o
m
p
o
n
en
ts
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
3
9
,
No
.
2
,
Au
g
u
s
t
20
25
:
1
3
9
9
-
1
410
1406
3
.
4
.
Dis
t
ributio
n o
f
clus
t
er
ing
T
h
en
,
we
o
b
tain
ed
f
o
u
r
clu
s
te
r
b
ased
o
n
th
e
d
is
tr
ib
u
tio
n
o
f
t
h
e
d
ata
f
ea
tu
r
e.
C
lu
s
ter
0
,
h
ig
h
s
p
en
d
in
g
s
co
r
e
(
0
.
5
1
7
6
)
,
b
u
t
a
g
e
an
d
a
n
n
u
al
in
co
m
e
ar
e
clo
s
e
to
th
e
av
er
ag
e.
C
u
s
to
m
er
s
in
th
is
clu
s
ter
ten
d
to
h
av
e
h
ig
h
s
p
en
d
in
g
with
ab
o
v
e
-
a
v
e
r
ag
e
p
r
o
d
u
ct
p
r
ices
(
u
n
it
p
r
ice
:
0
.
3
4
2
)
.
C
lu
s
ter
1
,
lo
w
s
p
en
d
in
g
s
co
r
e
(
-
0
.
5
4
0
9
)
with
ab
o
v
e
av
er
ag
e
in
c
o
m
e
an
d
ag
e.
C
u
s
to
m
er
s
in
th
is
c
lu
s
ter
ar
e
m
o
r
e
ca
r
ef
u
l
in
s
h
o
p
p
in
g
o
r
f
o
c
u
s
o
n
lo
wer
-
p
r
iced
p
r
o
d
u
cts.
C
lu
s
ter
2
,
v
er
y
h
ig
h
q
u
a
n
tity
(
1
.
1
9
5
8
)
an
d
h
ig
h
est
to
tal
p
r
ic
e
(
1
.
9
5
7
0
)
.
T
h
ese
cu
s
to
m
er
s
ten
d
to
b
u
y
in
lar
g
e
q
u
an
titi
es
an
d
f
o
cu
s
o
n
h
ig
h
er
p
r
ice
d
p
r
o
d
u
cts.
C
lu
s
ter
3
,
v
er
y
l
o
w
q
u
a
n
tity
(
-
1
.
0
2
9
3
)
an
d
l
o
w
to
tal
p
r
ice
(
-
1
.
2
2
)
,
in
d
icatin
g
cu
s
to
m
er
s
w
ith
lo
w
p
u
r
ch
ase
f
r
eq
u
e
n
cy
an
d
a
tig
h
ter
b
u
d
g
et.
T
ab
le
9
h
elp
s
in
d
eter
m
i
n
in
g
s
p
ec
if
ic
m
ar
k
etin
g
s
tr
ateg
ies
f
o
r
ea
ch
cu
s
to
m
er
s
eg
m
en
t.
PC
A
clu
s
ter
d
is
tr
i
b
u
tio
n
h
el
p
s
to
r
e
d
u
ce
th
e
co
m
p
lex
ity
o
f
th
e
o
r
ig
in
al
d
ata
in
to
two
d
im
en
s
io
n
s
.
T
h
e
r
esu
lts
ca
n
b
e
s
ee
n
in
T
ab
le
1
0
.
T
ab
le
9
.
Av
e
r
ag
e
p
r
o
f
ile
f
o
r
e
ac
h
clu
s
ter
b
ased
o
n
f
ea
tu
r
es
C
l
u
st
e
r
A
g
e
A
n
n
u
a
l
i
n
c
o
m
e
S
p
e
n
d
i
n
g
sc
o
r
e
P
u
r
c
h
a
se
h
i
st
o
r
y
Q
u
a
n
t
i
t
y
U
n
i
t
p
r
i
c
e
To
t
a
l
p
r
i
c
e
0
-
0
.
5
2
-
0
.
0
4
0
.
5
1
7
6
-
0
.
0
9
1
0
-
0
.
3
0
7
0
0
.
3
4
2
-
0
.
0
8
1
0
.
5
1
0
.
0
3
-
0
.
5
4
0
9
0
.
1
0
3
9
0
.
3
4
4
3
-
0
.
2
5
-
0
.
0
8
2
0
.
0
1
-
0
.
1
0
0
.
0
2
1
7
0
.
0
2
6
6
1
.
1
9
5
8
1
2
.
2
6
1
9
.
5
7
3
0
.
0
0
0
.
1
0
0
.
0
2
7
8
-
0
.
0
4
6
7
-
1
.
0
2
9
3
-
1
1
.
5
-
1
.
2
2
T
ab
le
1
0
.
K
-
m
ea
n
s
clu
s
ter
in
g
r
esu
lts
o
n
PC
A
d
ata
with
f
o
u
r
clu
s
ter
s
P
C
1
P
C
2
C
l
u
st
e
r
-
1
.
8
6
8
.
1
6
9
-
1
.
2
8
3
.
9
7
6
0
-
0
.
5
3
1
6
3
8
-
0
.
9
4
3
2
6
5
2
0
.
5
4
8
0
0
7
1
.
3
1
1
.
3
9
5
3
-
0
.
3
1
2
8
2
9
1
.
8
2
8
.
9
4
9
3
3
.
0
8
1
.
8
3
6
-
0
.
8
6
5
2
9
6
1
W
e
o
b
tain
ed
d
ata
d
is
tr
ib
u
tio
n
b
y
cl
u
s
ter
in
g
.
Fig
u
r
e
5
s
h
o
w
s
th
e
n
u
m
b
er
o
f
p
r
o
d
u
cts
p
er
ca
teg
o
r
y
i
n
ea
ch
clu
s
ter
,
p
r
o
v
id
in
g
an
o
v
er
v
iew
o
f
th
e
p
r
o
d
u
ct
ca
teg
o
r
y
p
r
ef
er
e
n
ce
s
s
u
ch
as
elec
tr
o
n
ics,
f
ash
io
n
,
an
d
g
r
o
ce
r
ies
am
o
n
g
th
e
clu
s
ter
s
.
Fig
u
r
e
6
illu
s
tr
ates
th
e
m
o
s
t
c
o
m
m
o
n
l
y
u
s
ed
p
a
y
m
en
t
m
eth
o
d
s
in
ea
ch
clu
s
te
r
,
s
u
ch
as
cr
ed
it
ca
r
d
,
d
e
b
it
ca
r
d
,
an
d
Pay
Pal.
T
h
is
h
elp
s
u
n
d
er
s
tan
d
th
e
d
o
m
in
an
t
p
a
y
m
e
n
t
m
eth
o
d
s
am
o
n
g
cu
s
to
m
er
s
in
a
p
ar
ticu
lar
clu
s
ter
.
Fig
u
r
e
7
s
h
o
ws
th
e
av
er
a
g
e
to
tal
p
r
ice
p
er
p
r
o
d
u
ct
ca
teg
o
r
y
in
ea
ch
cl
u
s
ter
.
T
h
is
v
is
u
aliza
tio
n
p
r
o
v
id
es
i
n
s
ig
h
t
in
t
o
th
e
p
r
ice
tr
en
d
s
an
d
p
r
o
d
u
ct
p
r
ef
er
e
n
ce
s
b
y
ca
teg
o
r
y
i
n
ea
ch
clu
s
ter
.
Fig
u
r
e
8
illu
s
tr
ates
th
e
d
is
tr
ib
u
tio
n
o
f
tr
an
s
ac
tio
n
I
Ds
ag
ain
s
t
tr
an
s
ac
tio
n
d
ates
in
ea
ch
clu
s
ter
.
T
h
is
v
is
u
aliza
tio
n
p
r
o
v
i
d
es
in
f
o
r
m
atio
n
ab
o
u
t
tr
a
n
s
ac
tio
n
d
ate
an
d
f
r
e
q
u
en
c
y
o
f
tr
a
n
s
ac
tio
n
s
o
cc
u
r
am
o
n
g
th
e
clu
s
ter
s
.
Fig
u
r
e
5
.
Dis
tr
ib
u
tio
n
o
f
p
r
o
d
u
ct
ca
teg
o
r
y
b
y
cl
u
s
ter
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
E
n
h
a
n
cin
g
ma
r
ke
tin
g
efficien
cy
th
r
o
u
g
h
d
a
ta
-
d
r
iven
cu
s
to
mer seg
men
ta
tio
n
…
(
F
a
n
i
n
d
i
a
P
u
r
n
a
ma
s
a
r
i
)
1407
Fig
u
r
e
6
.
Dis
tr
ib
u
tio
n
o
f
p
ay
m
en
t m
eth
o
d
b
y
cl
u
s
ter
in
g
Fig
u
r
e
7
.
Dis
tr
ib
u
tio
n
o
f
p
r
o
d
u
ct
ca
teg
o
r
y
an
d
a
v
er
ag
e
to
tal
p
r
ice
b
y
cl
u
s
ter
in
g
Fig
u
r
e
8
.
T
r
an
s
ac
tio
n
d
ate
a
n
d
tr
an
s
ac
tio
n
I
D
clu
s
ter
in
g
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
,
Vo
l.
3
9
,
No
.
2
,
Au
g
u
s
t
20
25
:
1
3
9
9
-
1
410
1408
3
.
5
.
I
nte
rpre
t
a
t
io
n
o
f
clus
t
er
ing
re
s
ult
Un
d
er
s
tan
d
in
g
th
e
ch
a
r
ac
ter
i
s
tics
th
at
d
escr
ib
e
ea
ch
clu
s
ter
is
cr
u
cial
f
o
r
in
ter
p
r
etin
g
r
esu
lts
o
f
clu
s
ter
in
g
an
aly
s
es,
s
u
ch
as
t
h
o
s
e
p
r
o
d
u
ce
d
b
y
K
-
m
ea
n
s
.
T
ab
le
1
1
p
r
o
v
id
e
in
f
o
r
m
atio
n
to
u
n
d
er
s
tan
d
th
e
r
elatio
n
s
h
ip
b
etw
ee
n
am
o
n
g
av
er
ag
e
PC
1
an
d
PC
2
to
m
o
s
t
co
m
m
o
n
p
r
o
d
u
ct
ca
teg
o
r
y
an
d
m
o
s
t
co
m
m
o
n
p
ay
m
en
t.
T
a
b
le
s
1
2
an
d
1
3
p
r
o
v
id
e
in
f
o
r
m
atio
n
to
u
n
d
er
s
tan
d
clu
s
ter
ch
ar
ac
ter
is
tics
,
b
u
s
in
ess
in
s
ig
h
ts
,
an
d
r
elev
an
t
s
tr
ateg
y
r
ec
o
m
m
en
d
atio
n
s
.
T
h
e
tab
les
p
r
o
v
id
e
clu
s
ter
in
g
r
esu
lts
,
u
n
iq
u
e
ch
ar
ac
ter
is
tics
o
f
ea
c
h
clu
s
ter
,
an
d
h
o
w
th
ese
r
esu
lts
ca
n
b
e
u
s
ed
f
o
r
m
o
r
e
e
f
f
ec
tiv
e
m
ar
k
etin
g
s
tr
ateg
ies.
T
ab
le
1
1
.
Pr
o
f
ile
clu
s
ter
a
n
aly
s
is
C
l
u
st
e
r
A
v
e
r
a
g
e
P
C
1
A
v
e
r
a
g
e
P
C
2
M
o
s
t
c
o
m
mo
n
p
r
o
d
u
c
t
c
a
t
e
g
o
r
y
M
o
s
t
c
o
m
mo
n
p
a
y
m
e
n
t
m
e
t
h
o
d
0
0
.
1
2
0
0
1
.
5
6
8
El
e
c
t
r
o
n
i
c
s
C
r
e
d
i
t
c
a
r
d
1
-
1
.
1
8
0
-
0
.
9
5
0
F
a
sh
i
o
n
D
e
b
i
t
c
a
r
d
2
2
.
3
2
0
-
0
.
8
6
0
G
r
o
c
e
r
i
e
s
P
a
y
P
a
l
T
ab
le
1
2
.
B
u
s
in
ess
in
s
ig
h
t f
r
o
m
clu
s
ter
s
C
l
u
st
e
r
B
u
s
i
n
e
ss I
n
si
g
h
t
s
0
C
u
s
t
o
mers
t
e
n
d
t
o
b
u
y
e
l
e
c
t
r
o
n
i
c
p
r
o
d
u
c
t
s a
n
d
u
se
c
r
e
d
i
t
c
a
r
d
s
.
M
a
r
k
e
t
i
n
g
s
t
r
a
t
e
g
y
:
d
i
sc
o
u
n
t
o
f
f
e
r
s fo
r
e
l
e
c
t
r
o
n
i
c
s
a
n
d
c
r
e
d
i
t
c
a
r
d
p
r
o
m
o
t
i
o
n
s
.
1
C
u
s
t
o
mer
h
a
v
e
a
p
r
e
f
e
r
e
n
c
e
f
o
r
f
a
s
h
i
o
n
a
n
d
u
se
d
e
b
i
t
c
a
r
d
s.
S
t
r
a
t
e
g
y
:
c
a
m
p
a
i
g
n
e
x
c
l
u
s
i
v
e
o
f
f
e
r
s fo
r
f
a
sh
i
o
n
p
r
o
d
u
c
t
s
w
i
t
h
d
e
b
i
t
c
a
r
d
i
n
s
t
a
l
men
t
s.
2
C
u
s
t
o
mer
t
e
n
d
t
o
b
u
y
w
h
o
l
e
sal
e
p
r
o
d
u
c
t
p
u
r
c
h
a
ses.
S
t
r
a
t
e
g
y
:
o
f
f
e
r
s
u
b
scri
p
t
i
o
n
p
r
o
m
o
t
i
o
n
s
o
r
l
a
r
g
e
p
u
r
c
h
a
se
d
i
sc
o
u
n
t
s.
T
ab
le
1
3
.
Ma
r
k
etin
g
s
tr
ateg
ies
r
ec
o
m
m
en
d
atio
n
f
r
o
m
cl
u
s
ter
s
C
l
u
st
e
r
M
a
r
k
e
t
i
n
g
S
t
r
a
t
e
g
i
e
s
0
P
r
o
v
i
d
e
s
p
e
c
i
a
l
d
i
s
c
o
u
n
t
s
o
n
e
l
e
c
t
r
o
n
i
c
p
u
r
c
h
a
s
e
s w
i
t
h
c
r
e
d
i
t
c
a
r
d
s.
La
u
n
c
h
c
a
s
h
b
a
c
k
p
r
o
g
r
a
m
o
r
r
e
w
a
r
d
p
o
i
n
t
s
c
a
mp
a
i
g
n
s
f
o
r
c
r
e
d
i
t
c
a
r
d
u
sers.
1
P
r
o
v
i
d
e
f
r
e
e
i
n
st
a
l
m
e
n
t
s
o
r
a
d
d
i
t
i
o
n
a
l
d
i
s
c
o
u
n
t
s
o
n
f
a
s
h
i
o
n
p
u
r
c
h
a
ses
f
o
r
c
u
st
o
m
e
r
w
h
o
u
s
i
n
g
d
e
b
i
t
c
a
r
d
.
C
r
e
a
t
e
a
l
o
y
a
l
t
y
p
r
o
g
r
a
mm
e
f
o
r
f
a
sh
i
o
n
p
r
o
d
u
c
t
.
2
P
r
o
v
i
d
e
a
su
b
s
c
r
i
p
t
i
o
n
p
r
o
g
r
a
m
t
o
b
u
y
e
r
w
h
o
b
u
y
w
h
o
l
e
s
a
l
e
p
r
o
d
u
c
t
.
O
f
f
e
r
sp
e
c
i
a
l
d
i
sc
o
u
n
t
s fo
r
b
u
l
k
o
r
r
e
c
u
r
r
i
n
g
p
u
r
c
h
a
s
e
s.
4.
CO
NCLU
SI
O
N
I
n
th
is
s
tu
d
y
d
em
o
n
s
tr
ates
th
e
s
ig
n
if
ican
t
o
f
r
o
le
ML
m
o
d
els
m
ay
co
n
v
er
t
u
n
s
tr
u
ct
u
r
ed
co
n
s
u
m
er
d
ata
in
to
u
s
ef
u
l
in
s
ig
h
ts
.
T
h
is
m
o
d
el
th
at
h
elp
s
b
u
s
in
ess
es
s
tay
co
m
p
etitiv
e
in
a
m
ar
k
et.
W
e
im
p
lem
en
ted
PC
A
an
aly
s
i
s
in
d
im
en
s
io
n
r
ed
u
cin
g
,
u
s
in
g
k
-
m
ea
n
s
clu
s
ter
in
g
,
th
en
ev
alu
ate
b
y
u
s
in
g
in
er
tia
an
d
s
illeh
o
u
te
s
co
r
e
with
1
.
4
5
5
.
6
5
0
an
d
0
.
4
8
6
3
6
6
r
esp
ec
tiv
ely
.
K
-
m
ea
n
s
clu
s
ter
in
g
is
s
u
cc
es
s
to
d
is
tin
ct
cu
s
to
m
er
in
to
s
ev
er
al
g
r
o
u
p
s
b
ased
o
n
p
r
ef
er
en
ce
a
n
d
h
is
to
r
y
tr
an
s
ac
tio
n
,
wh
ich
en
ab
les
m
o
r
e
a
cc
u
r
ate
m
ar
k
etin
g
allo
ca
tio
n
,
wh
ich
en
h
a
n
ce
s
cu
s
to
m
er
en
g
ag
e
m
en
t,
an
d
m
a
x
im
izes
s
ales
p
er
f
o
r
m
an
ce
.
Fu
tu
r
e
s
tu
d
ies
m
ay
in
v
esti
g
ate
th
e
u
s
e
o
f
lar
g
e
r
an
d
m
o
r
e
d
iv
er
s
e
d
atasets
,
in
co
r
p
o
r
atin
g
a
d
d
itio
n
al
f
ac
to
r
s
lik
e
lo
n
g
-
ter
m
cu
s
to
m
er
b
eh
av
io
r
o
r
p
r
o
d
u
ct
p
r
ef
er
en
ce
s
.
I
n
a
d
d
itio
n
,
u
s
in
g
th
e
K
-
m
ea
n
s
++
ap
p
r
o
ac
h
is
ex
p
ec
ted
to
p
r
o
v
id
e
th
e
ad
v
an
ta
g
e
o
f
s
elec
tin
g
a
n
i
d
ea
l c
en
tr
o
id
at
t
h
e
in
itial it
er
a
tio
n
an
d
s
p
ee
d
in
g
u
p
th
e
c
o
n
v
er
g
en
ce
p
r
o
ce
s
s
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
T
h
is
r
esear
ch
was
f
in
an
cially
s
u
p
p
o
r
ted
b
y
U
n
iv
er
s
itas
Su
m
ater
a
Utar
a
th
r
o
u
g
h
a
r
esear
ch
g
r
a
n
t
u
n
d
er
t
h
e
Gr
an
t N
u
m
b
er
3
6
4
1
/UN5
.
2
.
1
4
.
D/PT.
0
1
.
0
3
/
2
0
2
4
.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
id
u
a
l
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
Fan
in
d
ia
Pu
r
n
am
asar
i
✓
✓
✓
✓
✓
✓
✓
✓
Um
ay
a
R
am
ad
h
an
i
Pu
tr
i N
asu
tio
n
✓
✓
✓
✓
✓
✓
✓
Ma
r
is
ch
a
E
lv
en
y
✓
✓
✓
✓
✓
✓
✓
Evaluation Warning : The document was created with Spire.PDF for Python.