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
9
0
~
2
3
9
8
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
9
0
-
2398
2390
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
M
ining
t
he
Web
Da
ta for Clas
sify
i
ng
and Pred
icting
Users’
Requests
G
iri
s
h
S
,
Ra
m
a
m
urt
hy
B
,
Senthilna
t
ha
n T
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter S
c
ien
c
e
,
Ch
rist
U
n
iv
e
rsity
,
Ho
su
r
Ro
a
d
,
Be
n
g
a
lu
ru
,
Ka
rn
a
tak
a
,
In
d
ia
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
J
an
31
,
2
0
1
8
R
ev
i
s
ed
A
p
r
2
,
2
0
1
8
A
cc
ep
ted
A
p
r
20
,
2
0
1
8
Co
n
su
m
e
rs are
th
e
m
o
st i
m
p
o
rtan
t
a
ss
e
t
o
f
a
n
y
o
rg
a
n
iza
ti
o
n
.
T
h
e
c
o
m
m
e
r
c
ial
a
c
ti
v
it
y
o
f
a
n
o
rg
a
n
iza
ti
o
n
b
o
o
m
s
w
it
h
th
e
p
re
se
n
c
e
o
f
a
lo
y
a
l
c
u
sto
m
e
r
w
h
o
is
v
isib
ly
c
o
n
ten
t
w
it
h
th
e
p
ro
d
u
c
t
a
n
d
se
rv
ice
s
b
e
in
g
o
f
f
e
r
e
d
.
In
a
d
y
n
a
m
ic
m
a
rk
e
t,
u
n
d
e
rsta
n
d
i
n
g
v
a
riatio
n
s
in
c
li
e
n
t‟s
b
e
h
a
v
io
r
c
a
n
h
e
lp
e
x
e
c
u
ti
v
e
s
e
sta
b
li
sh
o
p
e
ra
ti
v
e
p
ro
m
o
ti
o
n
a
l
c
a
m
p
a
ig
n
s.
A
g
o
o
d
n
u
m
b
e
r
o
f
n
e
w
c
o
n
su
m
e
rs
a
re
f
re
q
u
e
n
tl
y
p
ick
e
d
u
p
b
y
trad
e
rs
d
u
ri
n
g
p
r
o
m
o
ti
o
n
s.
T
h
o
u
g
h
,
se
v
e
r
a
l
o
f
th
e
se
e
n
g
ro
ss
e
d
c
o
n
su
m
e
rs
a
r
e
o
n
e
-
ti
m
e
d
e
a
l
s
e
e
k
e
rs,
th
e
p
ro
m
o
ti
o
n
s
u
n
d
e
n
iab
ly
lea
v
e
a
p
o
siti
v
e
im
p
a
c
t
o
n
sa
les
.
It
is
c
ru
c
ial
f
o
r
trad
e
rs
to
i
d
e
n
ti
f
y
w
h
o
c
a
n
b
e
c
o
n
v
e
rted
to
lo
y
a
l
c
o
n
su
m
e
r
a
n
d
th
e
n
h
a
v
e
th
e
m
p
a
tro
n
ize
p
ro
d
u
c
ts
a
n
d
se
rv
ice
s
to
re
d
u
c
e
th
e
p
ro
m
o
ti
o
n
c
o
st
a
n
d
in
c
re
a
se
th
e
re
tu
rn
o
n
in
v
e
stm
e
n
ts.
T
h
is
stu
d
y
in
teg
ra
tes
a
c
las
sif
ier
th
a
t
a
ll
o
w
s
p
re
d
ictio
n
o
f
th
e
ty
p
e
o
f
p
u
rc
h
a
se
th
a
t
a
c
u
st
o
m
e
r
w
o
u
ld
m
a
k
e
,
a
s
w
e
ll
a
s
th
e
n
u
m
b
e
r
o
f
v
is
it
s
th
a
t
h
e
/sh
e
w
o
u
ld
m
a
k
e
d
u
rin
g
a
y
e
a
r.
T
h
e
p
ro
p
o
se
d
m
o
d
e
l
a
lso
c
re
a
te
s
o
u
tl
in
e
s
o
f
u
se
rs
a
n
d
b
ra
n
d
s
o
r
it
e
m
s
u
se
d
b
y
th
e
m
.
T
h
e
se
o
u
tl
in
e
s
m
a
y
n
o
t
b
e
u
se
f
u
l
o
n
ly
f
o
r
th
is
p
a
rti
c
u
lar
p
re
d
ictio
n
tas
k
,
b
u
t
c
o
u
ld
a
lso
b
e
u
se
d
f
o
r
o
t
h
e
r
im
p
o
rtan
t
tas
k
s
in
e
-
c
o
m
m
e
rc
e
,
su
c
h
a
s
c
li
e
n
t
se
g
m
e
n
tatio
n
,
p
ro
d
u
c
t
re
c
o
m
m
e
n
d
a
ti
o
n
a
n
d
c
li
e
n
t
b
a
se
g
ro
w
th
f
o
r
b
ra
n
d
s
.
K
ey
w
o
r
d
:
C
u
s
to
m
er
Se
g
m
en
tatio
n
C
las
s
i
f
icatio
n
P
r
ed
ictio
n
T
r
en
d
E
-
co
m
m
er
ce
P
r
o
d
u
ct
R
ec
o
m
m
e
n
d
atio
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
:
Gir
is
h
S,
Dep
ar
t
m
en
t o
f
C
o
m
p
u
ter
Scie
n
ce
,
C
h
r
i
s
t U
n
iv
er
s
it
y
,
Ho
s
u
r
R
o
ad
,
B
en
g
al
u
r
u
,
Kar
n
atak
a,
I
n
d
ia.
E
m
ail:
g
ir
is
h
.
s
@
cs.c
h
r
is
t
u
n
iv
e
r
s
it
y
.
i
n
1.
I
NT
RO
D
UCT
I
O
N
E
-
co
m
m
er
ce
w
ith
t
h
e
h
elp
o
f
th
e
W
o
r
ld
W
id
e
W
eb
h
as
tak
en
w
o
r
ld
b
u
s
i
n
ess
e
s
to
th
e
n
ex
t
le
v
el,
w
h
er
e
t
h
e
tr
ad
itio
n
al
ap
p
r
o
ac
h
i
s
o
v
er
s
h
ad
o
w
ed
b
y
t
h
e
s
ea
m
les
s
,
q
u
ic
k
,
e
f
f
icie
n
t
an
d
m
in
d
-
b
lo
w
in
g
ass
is
tan
ce
o
f
tec
h
n
o
lo
g
y
.
I
t
h
as
p
av
ed
th
e
w
a
y
f
o
r
a
co
n
v
en
ie
n
t
f
o
r
m
o
f
co
n
d
u
ct
in
g
b
u
s
i
n
ess
.
No
w
,
th
a
t
E
-
co
m
m
er
ce
is
t
h
e
tr
en
d
i
n
b
u
s
in
e
s
s
o
r
g
a
n
izatio
n
s
an
d
it
i
s
h
er
e
to
s
tay
f
o
r
th
e
co
m
in
g
y
ea
r
s
,
f
ir
m
s
in
g
e
n
er
al
m
u
s
t
g
ea
r
to
a
u
to
m
atin
g
th
eir
co
n
s
u
m
er
d
ata
a
n
d
t
h
e
v
ar
ied
i
n
f
o
r
m
atio
n
o
n
t
h
eir
p
r
ef
er
e
n
ce
s
i
n
o
r
d
er
to
r
etain
th
eir
e
x
is
t
in
g
cl
ien
t
s
a
n
d
attr
ac
t
th
e
n
e
w
clie
n
ts
.
„
C
u
s
to
m
ized
Ser
v
ici
n
g
‟
is
p
r
ed
icted
to
b
e
th
e
o
n
l
y
w
a
y
f
o
r
w
ar
d
in
th
e
s
e
ti
m
es
o
f
d
e
m
an
d
in
g
co
n
s
u
m
er
s
,
w
h
o
to
d
a
y
n
o
t
o
n
l
y
lo
o
k
f
o
r
w
ar
d
to
th
e
b
est
s
er
v
ice
i
n
t
h
e
least
p
o
s
s
ib
le
ti
m
e
b
u
t
co
s
t
-
ef
f
ec
ti
v
e
s
er
v
ice
to
o
.
Data
-
m
in
i
n
g
tec
h
n
iq
u
es
an
al
y
s
e
lar
g
e
a
m
o
u
n
t
o
f
d
ata
w
it
h
o
u
t
an
y
p
r
e
-
d
e
f
in
ed
h
y
p
o
th
esi
s
to
ex
tr
ac
t
m
ea
n
i
n
g
f
u
l
i
n
f
o
r
m
at
io
n
,
r
u
les
an
d
c
o
n
s
tr
ain
t
s
.
Data
Min
in
g
is
a
p
r
o
ce
s
s
o
f
ex
tr
ac
ti
n
g
k
n
o
w
led
g
e
f
r
o
m
d
atab
ase
s
,
in
v
o
l
v
in
g
p
atter
n
ex
tr
ac
ti
o
n
a
lg
o
r
it
h
m
s
.
Data
-
m
i
n
i
n
g
h
as
a
co
llectio
n
o
f
v
ar
io
u
s
tech
n
iq
u
es
to
e
x
tr
ac
t
p
atter
n
s
a
n
d
to
b
u
ild
m
o
d
el
s
f
r
o
m
lar
g
e
d
ata
-
s
ets.
T
h
e
ch
al
len
g
i
n
g
p
ar
t
f
o
r
m
o
s
t
o
f
t
h
e
b
u
s
i
n
es
s
es
to
d
ay
i
s
to
u
n
d
er
s
ta
n
d
th
e
n
ee
d
s
o
f
th
e
ir
cu
s
to
m
er
s
i
n
a
d
y
n
a
m
ic
e
n
v
ir
o
n
m
e
n
t.
I
n
s
u
c
h
a
s
it
u
atio
n
,
c
h
a
n
g
e
m
i
n
in
g
i
s
u
s
ed
b
y
a
n
al
y
s
t
s
t
o
u
n
d
er
s
ta
n
d
cu
s
to
m
er
‟
s
n
ee
d
s
.
As
d
ec
is
io
n
-
tr
ee
alg
o
r
ith
m
i
s
a
clas
s
i
f
icatio
n
-
b
ased
alg
o
r
ith
m
,
it
d
o
es
n
‟
t
i
n
v
o
l
v
e
a
co
m
p
lete
ch
a
n
g
e
in
th
e
d
ata
-
s
et.
„
R
u
l
e
E
x
tr
ac
tio
n
‟
w
as
t
h
e
m
o
s
t
w
id
el
y
u
s
ed
tech
n
iq
u
es
i
n
u
n
d
er
s
tan
d
in
g
t
h
e
r
elatio
n
s
h
ip
b
et
wee
n
v
ar
io
u
s
p
r
o
d
u
ct
ite
m
s
p
u
r
c
h
ased
b
y
th
e
cu
s
to
m
er
s
p
r
o
p
o
s
ed
a
s
y
s
te
m
t
h
at
c
an
id
en
tify
c
h
a
n
g
e
s
i
n
c
u
s
to
m
er
s
‟
b
e
h
av
io
u
r
.
T
h
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
Min
in
g
th
e
W
eb
Da
ta
fo
r
C
la
s
s
ifyin
g
a
n
d
P
r
ed
ictin
g
User
s
’
R
eq
u
ests
(
Gi
r
is
h
S
)
2391
s
y
s
te
m
u
s
e
s
p
r
ev
io
u
s
p
u
r
ch
a
s
e
r
ec
o
r
d
s
to
d
eter
m
i
n
e
an
d
id
en
tify
p
r
o
d
u
cts th
at
a
c
u
s
to
m
er
m
a
y
lik
e.
C
u
s
to
m
er
b
eh
av
io
u
r
g
i
v
es
u
s
d
e
m
o
g
r
ap
h
ic
v
ar
iab
les
u
s
ed
to
an
al
y
s
e
p
atter
n
s
.
R
ec
en
c
y
,
Fre
q
u
e
n
c
y
a
n
d
m
o
n
etar
y
(
R
FM)
ca
n
b
e
u
s
ed
to
d
if
f
er
en
tiate
b
e
t
w
ee
n
co
n
s
u
m
er
s
.
T
h
is
s
tu
d
y
i
n
te
g
r
ates
a
class
i
f
ier
w
h
ic
h
ca
n
p
r
ed
ict
th
e
n
ee
d
s
o
f
th
e
cu
s
to
m
er
s
an
d
t
h
e
t
y
p
e
o
f
p
u
r
ch
ase
s
th
e
y
m
i
g
h
t
m
a
k
e.
I
n
th
i
s
p
ap
er
,
w
e
h
a
v
e
ca
teg
o
r
ized
b
o
th
p
r
o
d
u
cts
an
d
cu
s
to
m
er
s
b
ased
o
n
th
eir
p
r
ev
io
u
s
r
ec
o
r
d
s
an
d
th
en
s
i
m
ilar
u
s
er
s
w
er
e
g
r
o
u
p
ed
to
g
eth
er
.
T
h
e
ex
tr
ac
ted
in
f
o
r
m
a
t
io
n
ca
n
b
e
u
s
ed
f
o
r
v
ar
io
u
s
e
-
co
m
m
er
ce
ap
p
lic
atio
n
s
s
u
c
h
as
p
r
o
d
u
ct
r
ec
o
m
m
e
n
d
atio
n
,
C
u
s
to
m
er
I
n
s
i
g
h
ts
,
C
u
s
to
m
er
s
eg
m
e
n
tatio
n
a
n
d
User
-
b
ased
b
r
an
d
id
en
tif
icatio
n
.
T
h
e
g
e
n
er
ated
u
s
er
p
r
o
f
iles
o
r
f
ea
tu
r
es
ca
n
b
e
u
s
ed
b
y
m
er
ch
a
n
ts
to
k
n
o
w
t
h
eir
cu
s
to
m
er
s
b
etter
.
2.
RE
L
AT
E
D
WO
RK
Min
i
n
g
is
t
h
e
p
r
o
ce
s
s
o
f
e
x
tr
ac
tin
g
k
n
o
w
le
d
g
e
o
r
i
n
f
o
r
m
a
t
io
n
f
r
o
m
th
e
w
eb
[
1
]
.
W
eb
Min
i
n
g
a
ls
o
h
as
its
o
w
n
t
y
p
e
s
o
r
w
a
y
s
i
n
w
h
ic
h
it
is
tr
ea
ted
ac
co
r
d
in
g
t
o
w
h
at
k
in
d
o
f
d
ata
it
co
n
tain
s
[
2
]
.
C
o
n
ten
t
Min
in
g
is
k
n
o
w
led
g
e
o
r
in
f
o
r
m
atio
n
g
ain
ed
f
r
o
m
t
h
e
co
n
te
n
t
o
f
t
h
e
s
ite
[
3
]
.
Stru
ctu
r
e
Mi
n
in
g
i
s
th
e
to
p
o
lo
g
y
o
f
t
h
e
s
ite
o
r
a
w
a
y
in
w
h
ic
h
t
h
e
r
ef
er
en
ce
s
o
r
lin
k
s
ar
e
p
u
t
at
t
h
e
s
ite;
Usa
g
e
Mi
n
i
n
g
i
s
ex
tr
ac
t
io
n
o
f
I
n
f
o
r
m
atio
n
f
r
o
m
t
h
e
u
s
er
lo
g
i
n
-
i
n
cr
ed
en
tials
an
d
s
to
r
ed
as
u
s
er
d
etail
s
ac
co
r
d
in
g
l
y
.
A
l
s
o
th
e
co
n
c
ep
t
o
f
w
eb
m
i
n
i
n
g
f
r
o
m
s
er
v
er
lo
g
d
etai
l
s
w
as
th
r
o
w
n
i
n
to
li
g
h
t
[
4
]
.
A
n
atte
m
p
t
w
a
s
m
ad
e
to
clas
s
if
y
u
s
er
s
b
ased
o
n
th
e
s
ite
‟
s
v
is
i
to
r
s
b
u
t
it
lac
k
ed
ac
cu
r
ac
y
b
ec
au
s
e
co
n
te
n
t
w
as
n
‟
t
ta
k
en
in
to
co
n
s
id
er
atio
n
.
C
las
s
i
f
ica
tio
n
atte
m
p
t
s
w
er
e
also
m
ad
e
o
n
te
x
t
co
n
te
n
ts
o
f
th
e
u
s
er
s
v
is
ited
s
ite
s
w
it
h
th
e
h
elp
o
f
t
h
e
lo
ca
l
ca
c
h
e
an
d
co
o
k
ies
[
5
]
.
Ho
w
e
v
er
,
as
th
i
s
w
as
o
n
l
y
b
a
s
ed
o
n
th
e
r
ec
en
t
v
is
i
ts
an
d
cli
ck
s
a
n
d
u
s
er
‟
s
i
n
ten
tio
n
s
ar
e
s
u
b
j
ec
t
to
ch
an
g
e
at
an
y
ti
m
e,
it
w
as
n
‟
t
ea
s
y
to
g
iv
e
ex
ac
t
p
r
ed
ictio
n
ea
ch
ti
m
e
a
s
u
s
er
s
tas
te
o
r
in
ter
est
s
te
n
d
to
ch
an
g
e
r
ap
id
l
y
.
I
t
b
ec
am
e
d
i
f
f
icu
l
t
to
ca
ch
e
p
ag
es
an
d
p
r
ed
ict
th
e
m
.
C
l
u
s
ter
i
n
g
B
ased
o
n
p
ag
es‟
ac
ce
s
s
an
d
p
ag
e
s
eq
u
en
ce
w
a
s
also
an
atte
m
p
t
m
ad
e
w
h
er
e
r
esu
lt
s
w
er
e
d
r
a
w
n
b
ased
o
n
th
e
s
e
s
s
io
n
ti
m
in
g
s
[
6
]
.
T
h
er
e
ar
e
s
o
m
e
alr
ea
d
y
ex
is
t
in
g
s
y
s
te
m
s
w
h
ic
h
h
el
p
th
e
w
eb
d
esi
g
n
er
s
i
n
o
r
g
an
izi
n
g
th
eir
w
eb
s
i
tes
ac
co
r
d
in
g
l
y
i
n
b
o
th
r
ec
o
m
m
e
n
d
at
io
n
s
m
eth
o
d
a
n
d
o
f
f
li
n
e
m
et
h
o
d
s
[
7
]
.
R
ec
o
m
m
en
d
atio
n
s
ar
e
g
e
n
er
all
y
b
ased
o
n
a
p
r
ev
io
u
s
u
s
er
‟
s
i
n
ter
est
an
d
i
f
t
h
e
p
atter
n
-
m
a
tch
o
cc
u
r
s
,
a
r
ec
o
m
m
e
n
d
atio
n
is
p
u
t
f
o
r
w
ar
d
to
t
h
e
u
s
er
.
A
cc
o
r
d
in
g
to
[
8
]
th
er
e
ar
e
f
e
w
w
a
y
s
i
n
w
h
ic
h
„
C
o
n
te
n
t
Mi
n
i
n
g
‟
ca
n
h
ap
p
en
;
P
r
e
-
m
in
in
g
,
w
h
er
e
th
e
s
es
s
io
n
s
o
n
l
y
i
n
v
o
l
v
e
co
n
ten
t
s
f
r
o
m
t
h
e
s
ite
an
d
P
o
s
t
-
m
i
n
i
n
g
,
w
h
er
e
th
e
co
n
te
n
t
an
d
th
e
r
esu
l
ts
ar
e
in
d
ep
en
d
en
t
[
9
]
.
Min
in
g
is
th
e
p
r
o
ce
s
s
o
f
ex
tr
ac
tin
g
k
n
o
w
led
g
e
o
r
in
f
o
r
m
atio
n
f
r
o
m
th
e
w
e
b
.
W
e
b
Min
in
g
also
h
a
s
its
o
w
n
t
y
p
es
o
r
w
a
y
s
i
n
w
h
ic
h
it
is
tr
ea
ted
ac
co
r
d
in
g
t
o
w
h
at
k
i
n
d
o
f
d
a
ta
it
co
n
tain
s
.
C
o
n
ten
t
Mi
n
in
g
,
i
s
n
o
th
i
n
g
b
u
t
k
n
o
w
led
g
e
o
r
in
f
o
r
m
atio
n
g
a
in
e
d
f
r
o
m
th
e
c
o
n
ten
t
o
f
th
e
s
ite
[
1
0
]
.
A
ls
o
th
e
co
n
ce
p
t
o
f
w
eb
m
in
in
g
f
r
o
m
s
er
v
er
lo
g
d
etail
s
w
a
s
th
r
o
w
n
i
n
to
lig
h
t
[
1
1
]
.
W
h
ich
is
a
co
m
b
i
n
atio
n
o
f
d
i
f
f
e
r
en
t
s
y
s
te
m
s
p
u
t
to
g
e
th
er
f
o
r
b
etter
r
esu
lts
[
12]
,
[
1
3
]
.
W
eb
m
in
i
n
g
h
elp
s
i
n
i
m
p
r
o
v
in
g
t
h
e
s
ca
lab
ilit
y
an
d
ef
f
ec
t
iv
e
n
es
s
o
f
a
s
ite.
An
ap
p
r
o
ac
h
o
f
u
s
in
g
s
e
m
a
n
ti
c
d
ata
g
at
h
er
ed
f
r
o
m
w
eb
m
in
i
n
g
an
d
s
h
o
w
h
o
w
s
e
m
an
tic
d
ata
ca
n
b
e
u
s
ed
to
p
er
s
o
n
alize
o
n
e‟
s
w
eb
s
i
te.
A
l
s
o
s
h
o
w
s
h
o
w
to
u
s
e
s
e
m
a
n
tic
d
ata
to
i
m
p
r
o
v
e
th
e
tr
af
f
ic
attr
ac
ted
to
w
ar
d
s
s
it
e
[
1
4
]
,
[
1
5
]
.
A
r
u
le
b
ased
p
ag
e
class
i
f
icatio
n
w
a
s
p
r
o
p
o
s
ed
[
1
6
]
.
A
m
o
d
el
w
h
er
e
u
s
er
n
a
v
i
g
atio
n
p
r
o
f
ile
s
ar
e
g
en
er
ated
w
it
h
t
h
e
h
elp
o
f
w
eb
m
in
i
n
g
f
r
o
m
th
e
d
ata
ac
q
u
ir
ed
f
r
o
m
th
e
s
er
v
er
s
.
T
h
is
ap
p
r
o
ac
h
is
b
ased
o
n
b
y
te
-
lev
el
a
n
d
lan
g
u
a
g
e
is
in
d
ep
en
d
e
n
t,
th
e
p
r
o
f
ile
s
ize
is
li
m
ited
an
d
t
h
e
ac
cu
r
ac
y
r
ate
i
s
b
ased
th
e
la
n
g
u
a
g
e
i
n
p
u
t
ted
[
1
7
]
.
T
h
e
r
ec
o
r
d
o
f
ev
en
ts
o
cc
u
r
r
in
g
o
v
er
a
p
er
io
d
o
f
ti
m
e
i
s
co
llected
f
r
o
m
t
h
e
s
er
v
er
d
o
m
ai
n
s
[
1
8
]
.
T
h
e
p
r
o
b
lem
s
i
n
r
ec
o
r
d
in
g
th
e
s
eq
u
en
tial
o
cc
u
r
r
en
ce
s
o
f
e
v
en
t
s
an
d
ea
ch
o
f
t
h
is
ac
ti
o
n
s
ar
e
s
p
lit
in
to
s
ess
io
n
s
th
e
y
s
h
o
w
th
a
t
o
n
e
s
u
ch
s
es
s
io
n
s
h
av
e
t
h
e
d
ata
wh
ich
w
e
ca
n
u
s
e
to
f
o
r
m
r
u
le
s
f
o
r
d
escr
ib
i
n
g
th
e
n
ex
t
o
cc
u
r
r
en
ce
o
f
an
ev
e
n
t.
T
h
ey
p
r
o
v
id
e
a
n
al
g
o
r
ith
m
wh
ich
is
u
s
ed
to
h
elp
r
ec
o
r
d
th
e
ev
en
ts
a
n
d
p
r
o
v
id
e
d
escr
ip
tio
n
f
o
r
it
[
1
9
]
,
[
2
0
]
.
3.
P
RO
P
O
SE
D
M
E
T
H
O
DO
L
O
G
Y
T
h
is
p
ap
er
ai
m
s
at
an
a
l
y
s
in
g
t
h
e
co
n
te
n
t o
f
a
n
E
-
co
m
m
er
ce
d
atab
ase.
B
ased
o
n
th
e
a
n
al
y
s
i
s
,
a
m
o
d
el
w
a
s
b
u
ilt to
p
r
ed
ict
th
e
p
u
r
ch
a
s
es o
f
a
n
e
w
c
u
s
to
m
er
b
ased
o
n
h
i
s
/
h
er
ea
r
lier
p
u
r
ch
asi
n
g
tr
a
ck
r
ec
o
r
d
.
3
.
1
.
Da
t
a
p
re
pa
ra
t
io
n
T
h
e
d
ataset
w
a
s
s
elec
ted
f
r
o
m
a
n
E
-
co
m
m
er
ce
d
ataset
co
m
p
r
i
s
i
n
g
4
0
0
,
0
0
en
tr
ies.
Fi
g
u
r
e
1
s
h
o
ws
th
at
t
h
e
d
ata
co
n
tai
n
ed
4
3
7
2
u
s
er
s
a
n
d
t
h
e
y
h
ad
p
u
r
c
h
ase
d
ab
o
u
t
3
6
8
4
p
r
o
d
u
cts
an
d
th
e
to
tal
n
u
m
b
er
o
f
tr
an
s
ac
tio
n
s
ca
r
r
ied
o
u
t
w
er
e
2
2
0
0
0
.
T
h
e
n
ex
t
s
tep
w
a
s
to
ar
r
iv
e
a
t
th
e,
n
u
m
b
er
o
f
p
r
o
d
u
cts
b
o
u
g
h
t
p
er
tr
an
s
ac
t
io
n
a
n
d
a
f
ter
t
h
is
a
ll
t
h
e
n
u
l
l
v
al
u
es
a
n
d
tr
an
s
ac
tio
n
s
wh
er
e
o
r
d
er
s
h
ad
b
ee
n
ca
n
ce
lled
w
er
e
r
e
m
o
v
ed
f
r
o
m
t
h
e
d
ata
s
et.
A
v
ar
iab
le
w
a
s
th
en
cr
ea
ted
to
s
h
o
w
t
h
e
to
tal
p
r
ice
o
f
ea
ch
p
u
r
ch
ase
m
ad
e
b
y
th
e
c
u
s
to
m
er
.
Fi
g
u
r
e
2
s
h
o
w
s
a
s
a
m
p
le
o
f
h
o
w
b
ask
et
p
r
ice
is
ca
lcu
lated
f
o
r
ea
ch
tr
an
s
ac
tio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
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
9
0
–
2
3
9
8
2392
Fig
u
r
e
1
.
Su
m
m
ar
y
o
f
t
h
e
w
h
o
le
d
ata
s
et
Fig
u
r
e
2
.
A
s
a
m
p
le
o
f
b
as
k
et
p
r
ice
f
o
r
ea
ch
tr
an
s
ac
tio
n
3
.
2
.
Understa
nd
ing
pro
du
ct
ca
t
eg
o
ries
T
o
b
eg
in
w
it
h
,
w
e
f
ir
s
t
e
x
tr
ac
t
in
f
o
r
m
atio
n
r
eg
ar
d
in
g
p
r
o
d
u
cts
f
r
o
m
th
e
d
escr
ip
tio
n
Var
iab
le.
Fig
u
r
e
3
s
h
o
w
s
h
o
w
th
e
b
as
k
et
p
r
ice
o
f
ea
ch
tr
an
s
ac
tio
n
is
d
iv
id
ed
f
o
r
th
e
w
h
o
le
d
ata
s
et.
T
h
e
p
r
o
ce
s
s
o
f
ex
tr
ac
tio
n
o
f
i
n
f
o
r
m
atio
n
is
as
f
o
llo
w
s
:
a.
E
x
tr
ac
t th
e
ter
m
s
f
r
o
m
t
h
e
d
es
cr
ip
tio
n
b.
Fo
r
ea
ch
o
f
th
e
s
e
ter
m
s
,
tr
y
to
f
i
n
d
o
u
t th
e
r
o
o
t a
n
d
co
llectio
n
o
f
s
et
o
f
ter
m
s
r
elate
d
w
it
h
i
t
c.
C
alcu
late
th
e
o
cc
u
r
r
e
n
ce
o
f
ter
m
in
t
h
e
d
ata
s
et
Fig
u
r
e
3
.
A
p
ie
-
c
h
ar
t d
is
p
la
y
i
n
g
t
h
e
p
r
ice
ag
ai
n
s
t e
ac
h
in
v
o
i
c
e
Fig
u
r
e
4
.
C
o
u
n
t o
f
ea
ch
w
o
r
d
in
th
e
p
r
o
d
u
ct
d
escr
ip
tio
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
Min
in
g
th
e
W
eb
Da
ta
fo
r
C
la
s
s
ifyin
g
a
n
d
P
r
ed
ictin
g
User
s
’
R
eq
u
ests
(
Gi
r
is
h
S
)
2393
Fig
u
r
e
4
s
h
o
w
s
th
e
n
u
m
b
er
o
f
o
cc
u
r
r
en
ce
s
o
f
ea
ch
w
o
r
d
in
th
e
d
ataset.
W
ith
th
e
h
el
p
o
f
th
ese
k
e
y
w
o
r
d
s
,
w
e
co
u
ld
n
o
w
g
r
o
u
p
th
e
p
r
o
d
u
cts.
Firstl
y
,
th
e
p
r
o
d
u
cts
w
er
e
g
r
o
u
p
ed
in
to
an
m
*
n
m
a
tr
ix
:
w
h
er
e
th
e
a{
m
,
n
}
co
ef
f
ic
ien
t
is
1
i
f
t
h
e
d
escr
ip
tio
n
o
f
t
h
e
p
r
o
d
u
ct
m
c
o
n
tai
n
s
th
e
w
o
r
d
n
,
a
n
d
0
o
th
er
w
is
e.
Fig
u
r
e
5
is
a
m
atr
i
x
s
h
o
w
in
g
h
o
w
t
h
e
k
e
y
w
o
r
d
s
ar
e
m
ap
p
ed
w
it
h
th
e
p
r
o
d
u
cts.
Fig
u
r
e
5
.
A
m
atr
ix
d
ep
icti
n
g
h
o
w
k
e
y
w
o
r
d
s
ar
e
m
ap
p
ed
w
it
h
r
esp
ec
tiv
e
p
r
o
d
u
cts
3
.
3
.
Clus
t
er
p
ro
du
ct
s
I
n
t
h
is
s
ec
tio
n
,
p
r
o
d
u
cts
o
f
s
i
m
ilar
k
i
n
d
w
er
e
g
r
o
u
p
ed
i
n
to
r
esp
ec
tiv
e
cla
s
s
e
s
.
Fo
r
g
r
o
u
p
i
n
g
o
f
t
h
ese
p
r
o
d
u
cts,
w
e
u
s
ed
K
m
ea
n
s
te
ch
n
iq
u
e,
w
h
er
e
it
u
s
ed
E
u
clid
ea
n
Di
s
tan
ce
to
ca
lc
u
late
t
h
e
d
is
tan
ce
a
n
d
g
r
o
u
p
th
e
p
r
o
d
u
cts
ac
co
r
d
in
g
l
y
.
W
h
ile
C
l
u
s
ter
i
n
g
,
it
w
a
s
f
o
u
n
d
th
at
w
h
e
n
t
h
e
n
u
m
b
er
o
f
clu
s
ter
s
h
ad
g
o
n
e
b
ey
o
n
d
5
,
th
e
n
u
m
b
er
o
f
ele
m
en
ts
i
n
ea
c
h
cl
u
s
ter
b
ec
a
m
e
v
er
y
lo
w
.
T
h
er
ef
o
r
e,
it
w
as
d
ec
id
ed
to
s
eg
r
eg
ate
th
e
m
in
to
5
clu
s
ter
s
.
Fi
g
u
r
e
6
s
h
o
w
s
n
u
m
b
er
o
f
ele
m
e
n
ts
i
n
ea
ch
clu
s
ter
.
Fig
u
r
e
6
.
Nu
m
b
er
o
f
ele
m
e
n
ts
in
ea
ch
C
l
u
s
ter
W
e
n
o
ticed
t
h
at
ea
c
h
c
lu
s
ter
co
n
tai
n
ed
o
b
j
ec
ts
th
at
co
u
l
d
b
e
ass
o
ciate
d
w
i
th
ter
m
s
w
h
ic
h
h
ad
m
ea
n
in
g
.
Fi
g
u
r
e
7
s
h
o
w
s
w
o
r
d
clo
u
d
w
it
h
v
ar
io
u
s
ele
m
en
t
s
in
ea
ch
cl
u
s
ter
.
Fig
u
r
e
7
.
A
w
o
r
d
clo
u
d
o
f
v
ar
i
o
u
s
ele
m
en
t
s
in
d
i
f
f
er
en
t c
l
u
s
t
er
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
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
9
0
–
2
3
9
8
2
394
3
.
4
.
P
rincipa
l c
o
m
po
nent
a
na
ly
s
i
s
I
n
o
r
d
er
to
ch
ec
k
th
e
u
n
iq
u
e
n
ess
o
f
ele
m
e
n
ts
in
t
h
e
cl
u
s
ter
.
a
P
C
A
w
a
s
p
er
f
o
r
m
ed
to
u
n
d
er
s
tan
d
its
d
is
tin
ct
n
es
s
.
Fi
g
u
r
e
8
s
h
o
w
s
h
o
w
ele
m
e
n
t
s
ar
e
s
ca
tter
ed
w
it
h
i
n
t
h
eir
g
r
o
u
p
s
.
Fi
g
u
r
e
9
s
h
o
w
s
t
h
e
p
r
o
d
u
cts
clu
s
ter
ed
af
ter
P
C
A
.
Fig
u
r
e
8
.
Un
iq
u
e
n
es
s
o
f
t
h
e
el
e
m
en
ts
w
it
h
t
h
e
clu
s
ter
Fig
u
r
e
9
.
C
lu
s
ter
o
f
v
ar
io
u
s
p
r
o
d
u
cts
3
.
5
.
Cus
t
o
m
er
c
a
t
eg
o
ries
W
e
n
ee
d
ed
to
f
ir
s
t
g
r
o
u
p
th
e
v
ar
io
u
s
p
r
o
d
u
cts
i
n
to
5
cl
u
s
ter
s
.
Fo
r
a
n
al
y
s
is
,
t
h
is
in
f
o
r
m
atio
n
w
a
s
ad
d
ed
to
th
e
d
ataset
in
th
e
f
o
r
m
o
f
ca
t_
p
r
o
d
,
w
h
er
e
ea
c
h
c
lu
s
ter
w
as
d
e
n
o
ted
.
Nex
t,
C
at
_
N
v
ar
iab
les
w
er
e
cr
ea
ted
,
w
h
ich
in
d
icate
d
t
h
e
a
m
o
u
n
t
s
p
en
t
i
n
ea
c
h
p
r
o
d
u
ct
ca
teg
o
r
y
.
B
y
d
o
i
n
g
th
i
s
,
w
e
h
ad
all
th
e
d
ata
t
h
at
w
a
s
r
eq
u
ir
ed
in
o
n
e
Data
f
r
a
m
e
o
r
tab
le.
Fig
u
r
e
1
0
s
h
o
ws
h
o
w
ea
c
h
cu
s
to
m
er
‟
s
in
v
es
t
m
e
n
t
i
n
d
if
f
er
en
t
ca
teg
o
r
ies.
Fig
u
r
e
1
0
.
A
s
a
m
p
le
s
h
o
w
i
n
g
h
o
w
ea
c
h
cu
s
to
m
er
‟
s
in
v
est
m
en
t v
ar
ie
s
in
v
ar
io
u
s
ca
te
g
o
r
ie
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2088
-
8708
Min
in
g
th
e
W
eb
Da
ta
fo
r
C
la
s
s
ifyin
g
a
n
d
P
r
ed
ictin
g
User
s
’
R
eq
u
ests
(
Gi
r
is
h
S
)
2395
Hav
i
n
g
d
o
n
e
all
t
h
is
,
w
e
co
u
l
d
n
o
w
f
i
n
d
t
h
e
o
th
er
v
ar
iab
les
lik
e
m
a
x
,
m
in
,
m
ea
n
a
n
d
last
_
p
u
r
ch
ase
f
o
r
ev
er
y
u
s
er
w
h
ic
h
h
elp
ed
i
n
id
en
ti
f
y
i
n
g
a
n
i
n
d
iv
id
u
al
u
s
er
w
h
o
h
ad
m
ad
e
j
u
s
t
o
n
e
s
i
n
g
le
p
u
r
c
h
ase.
On
e
o
f
th
e
o
b
j
ec
tiv
e
w
as
to
tar
g
et
th
e
s
e
cu
s
to
m
er
s
an
d
tr
y
to
r
etain
th
e
m
.
F
ig
u
r
e
1
1
s
h
o
w
s
Un
iq
u
en
es
s
o
f
cu
s
to
m
er
s
af
ter
clu
s
ter
in
g
.
Fig
u
r
e
1
1
.
Un
iq
u
en
e
s
s
o
f
cu
s
t
o
m
er
s
w
h
e
n
cl
u
s
ter
ed
T
h
e
n
u
m
b
er
o
f
c
lu
s
ter
s
c
h
o
s
e
n
w
er
e
b
ased
o
n
t
h
e
s
ilh
o
u
ett
e
s
co
r
e
an
d
t
h
e
b
est
s
co
r
e
w
a
s
o
b
tain
ed
w
it
h
1
1
clu
s
ter
s
.
I
t
is
e
v
id
en
t
th
at
c
u
s
to
m
er
s
in
th
e
cl
u
s
ter
s
w
er
e
d
is
ti
n
ct
a
n
d
th
e
co
n
te
n
t
s
w
er
e
av
er
a
g
ed
b
y
s
elec
ti
n
g
v
ar
io
u
s
cl
ien
t
s
.
W
h
e
n
ch
ec
k
ed
af
ter
cl
u
s
ter
i
n
g
t
h
e
to
tal
n
u
m
b
er
o
f
cu
s
to
m
er
s
w
er
e
3
6
0
8
.
4.
E
VA
L
UA
T
I
O
N
AND
R
E
SU
L
T
S
4
.
1
.
Cla
s
s
if
ica
t
io
n o
f
c
us
t
o
m
er
s
Th
e
o
b
j
ec
tiv
e
h
er
e
w
as
to
b
u
ild
a
m
o
d
el
th
at
w
o
u
ld
cla
s
s
i
f
y
c
u
s
to
m
er
s
i
n
to
d
i
f
f
er
en
t
cu
s
to
m
er
ca
teg
o
r
ies
b
ased
o
n
th
e
r
ec
o
g
n
itio
n
o
f
th
eir
ea
r
lier
p
u
r
ch
a
s
i
n
g
p
atter
n
s
.
T
h
e
m
ai
n
ai
m
w
a
s
to
m
ak
e
s
u
r
e
th
is
class
i
f
icatio
n
h
ap
p
en
ed
at
th
e
v
er
y
f
ir
s
t
v
i
s
it
i
ts
el
f
.
Fo
r
th
is
,
w
e
to
o
k
th
e
h
elp
o
f
th
e
at
tr
ib
u
tes
o
f
ea
c
h
cla
s
s
an
d
class
if
ied
b
ased
o
n
th
e
s
e
attr
ib
u
tes.
As
t
h
e
ai
m
w
as
to
d
ef
in
e
a
cla
s
s
a
t
th
e
f
ir
s
t
v
is
it,
o
n
l
y
t
h
e
co
n
ten
t
o
f
th
e
ite
m
w
as c
o
n
s
id
er
ed
an
d
v
ar
iab
les s
u
c
h
as
f
r
eq
u
en
c
y
an
d
all
w
er
e
i
g
n
o
r
ed
.
1
)
Su
p
p
o
r
t
Vec
to
r
Ma
ch
in
e
class
if
ier
:
At
f
ir
s
t,
SVC
cla
s
s
i
f
ier
w
a
s
u
s
ed
to
cr
ea
te
an
I
n
s
ta
n
ce
o
f
C
L
ASS_
FIT
an
d
th
e
n
ca
ll g
r
id
_
s
ea
r
ch
(
)
.
P
r
o
v
id
ed
P
ar
am
eter
s
w
er
e:
a.
H
y
p
er
p
ar
am
e
ter
s
w
it
h
o
p
ti
m
al
v
alu
e
s
.
b.
Nu
m
b
er
o
f
f
o
ld
s
f
o
r
cr
o
s
s
v
a
li
d
atio
n
.
W
h
en
T
ested
th
e
m
o
d
el
g
a
v
e
a
p
r
ec
is
io
n
o
f
8
7
.
4
0
%.
2
)
C
o
n
f
u
s
io
n
Ma
tr
i
x
:
F
ig
u
r
e
12
,
is
a
co
n
f
u
s
io
n
m
atr
i
x
th
at
s
h
o
w
s
t
h
e
p
r
ed
icted
v
alu
e
s
ar
e
p
lo
tted
o
n
th
e
x
ax
is
,
t
h
e
tr
u
e
v
a
lu
e
s
ar
e
p
lo
tted
o
n
th
e
y
ax
is
an
d
t
h
e
ele
m
e
n
ts
p
lo
tted
ac
r
o
s
s
th
e
d
i
ag
o
n
al
ar
e
r
ig
h
tl
y
class
i
f
ied
.
Fo
r
ex
a
m
p
le,
w
h
e
n
s
ee
n
i
n
th
e
ab
o
v
e
m
a
tr
ix
f
o
r
T
r
u
e
L
ab
el
„
2
‟
an
d
f
o
r
P
r
ed
icted
L
ab
el
„
2
‟
t
h
e
clas
s
i
f
ied
co
u
n
t o
f
ele
m
en
ts
is
2
7
2
.
Fig
u
r
e
1
2
.
A
co
n
f
u
s
io
n
m
atr
i
x
s
h
o
w
i
n
g
t
h
e
tr
u
e
v
al
u
es a
n
d
p
r
ed
icted
v
alu
es
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
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
9
0
–
2
3
9
8
2396
4
.
2
.
L
ea
rning
c
urv
e
Fig
u
r
e
13
is
a
lear
n
i
n
g
c
u
r
v
e
an
d
lear
n
i
n
g
c
u
r
v
e
i
s
a
t
y
p
ical
w
a
y
to
test
t
h
e
q
u
ali
t
y
o
f
a
f
it
.
I
n
p
ar
ticu
lar
,
th
e
s
e
t
y
p
e
o
f
c
u
r
v
es
allo
w
to
d
etec
t
p
o
s
s
ib
le
d
r
a
w
b
ac
k
s
i
n
a
m
o
d
el,
li
n
k
ed
f
o
r
ex
a
m
p
le
to
o
v
er
-
o
r
u
n
d
er
-
f
itti
n
g
.
T
h
is
al
s
o
s
h
o
w
s
to
w
h
ic
h
ex
te
n
t t
h
e
m
o
d
e
c
o
u
ld
b
en
ef
it
f
r
o
m
a
lar
g
er
d
ata
s
a
m
p
le.
Fig
u
r
e
1
3
.
A
s
et
lear
n
in
g
cu
r
v
es d
if
f
er
en
tiat
in
g
b
et
w
ee
n
v
ar
i
o
u
s
clas
s
i
f
ier
s
Fin
all
y
,
t
h
e
r
esu
lt
s
o
f
t
h
e
d
if
f
er
en
t
class
i
f
ier
s
p
r
esen
ted
in
t
h
e
p
r
ev
io
u
s
s
ec
tio
n
s
co
u
ld
b
e
co
m
b
i
n
ed
to
i
m
p
r
o
v
e
th
e
clas
s
if
icatio
n
m
o
d
el.
T
h
is
co
u
ld
b
e
ac
h
iev
e
d
b
y
s
elec
ti
n
g
th
e
c
u
s
to
m
er
ca
teg
o
r
y
as
t
h
e
o
n
e
in
d
icate
d
b
y
t
h
e
m
aj
o
r
ity
o
f
cl
ass
i
f
ier
s
.
T
o
d
o
th
is
,
th
e
„
Vo
ti
n
g
C
la
s
s
i
f
ier
‟
m
eth
o
d
o
f
t
h
e
s
k
lear
n
p
ac
k
ag
e
w
as
u
s
ed
.
A
s
a
f
ir
s
t
s
tep
,
th
e
p
a
r
a
m
eter
s
o
f
t
h
e
v
ar
io
u
s
clas
s
if
ier
s
u
s
i
n
g
th
e
b
est
p
r
ev
io
u
s
p
ar
am
e
ter
s
w
er
e
ad
j
u
s
ted
.
T
h
en
,
a
class
i
f
ier
w
a
s
d
ef
i
n
ed
th
at
m
er
g
ed
th
e
r
es
u
lts
o
f
t
h
e
v
ar
io
u
s
clas
s
i
f
ier
s
an
d
tr
ain
ed
th
e
m
.
A
f
e
w
c
lass
if
ier
s
w
er
e
tr
ai
n
e
d
to
ca
teg
o
r
ize
cu
s
to
m
er
s
.
U
n
til
t
h
at
p
o
i
n
t,
t
h
e
w
h
o
le
an
al
y
s
i
s
w
a
s
b
ased
o
n
th
e
d
ata
o
f
th
e
f
ir
s
t
1
0
m
o
n
t
h
s
.
I
n
th
i
s
s
ec
tio
n
,
th
e
m
o
d
el
f
o
r
th
e
last
t
w
o
m
o
n
t
h
s
o
f
th
e
d
ataset
t
h
at
h
ad
b
ee
n
s
to
r
ed
in
th
e
s
et_
test
d
ataf
r
a
m
e
w
a
s
test
ed
.
I
n
a
f
ir
s
t
s
tep
,
t
h
is
d
ata
w
as
r
eg
r
o
u
p
ed
an
d
r
e
f
o
r
m
atted
ac
co
r
d
in
g
to
t
h
e
s
a
m
e
p
r
o
ce
d
u
r
e
as
w
as
u
s
ed
o
n
t
h
e
tr
ai
n
i
n
g
s
et.
Ho
w
ev
er
,
to
tak
e
in
to
ac
co
u
n
t
th
e
d
if
f
er
en
ce
i
n
ti
m
e
b
et
w
ee
n
t
h
e
t
w
o
d
atasets
a
n
d
w
ei
g
h
th
e
v
ar
iab
les
co
u
n
t
an
d
s
u
m
t
o
o
b
tain
eq
u
iv
alen
ce
w
it
h
th
e
tr
ain
in
g
s
et,
th
e
d
ata
w
a
s
co
r
r
ec
ted
.
Fig
u
r
e
1
4
s
h
o
w
s
t
h
e
co
n
s
o
lid
ated
d
ataf
r
a
m
e
tak
e
n
f
o
r
test
i
n
g
.
Fig
u
r
e
1
4
.
A
s
a
m
p
le
d
ata
tak
e
n
f
o
r
test
i
n
g
T
h
en
,
th
e
d
ataf
r
a
m
e
w
a
s
co
n
v
er
ted
in
to
a
m
atr
i
x
an
d
o
n
l
y
v
ar
iab
les
t
h
at
d
ef
i
n
e
t
h
e
ca
teg
o
r
y
to
w
h
ic
h
co
n
s
u
m
er
s
b
elo
n
g
w
er
e
r
etain
ed
.
A
t
th
i
s
le
v
el,
th
e
m
et
h
o
d
o
f
n
o
r
m
al
izatio
n
u
s
ed
o
n
th
e
tr
ai
n
i
n
g
s
et
w
a
s
r
ec
alled
.
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
Min
in
g
th
e
W
eb
Da
ta
fo
r
C
la
s
s
ifyin
g
a
n
d
P
r
ed
ictin
g
User
s
’
R
eq
u
ests
(
Gi
r
is
h
S
)
2397
E
ac
h
lin
e
i
n
t
h
is
m
atr
ix
co
n
t
ain
ed
a
co
n
s
u
m
er
'
s
b
u
y
i
n
g
h
a
b
it
s
.
A
t
t
h
i
s
s
ta
g
e,
it
w
a
s
a
q
u
esti
o
n
o
f
u
s
i
n
g
t
h
ese
h
ab
it
s
i
n
o
r
d
er
t
o
d
ef
in
e
a
ca
te
g
o
r
y
to
w
h
ic
h
t
h
is
co
n
s
u
m
er
b
elo
n
g
s
.
Fi
g
u
r
e
1
5
s
h
o
w
s
t
h
e
p
r
ec
is
io
n
v
al
u
e
o
f
o
u
r
m
o
d
el
w
h
e
n
test
ed
w
it
h
v
ar
io
u
s
clas
s
if
ier
s
.
Fig
u
r
e
1
5
.
P
r
ec
is
io
n
v
alu
e
o
f
m
o
d
el
w
h
e
n
test
e
d
w
it
h
v
ar
io
u
s
cla
s
s
i
f
ier
s
5.
CO
NCLU
SI
O
N
T
h
e
w
o
r
k
d
escr
ib
ed
in
th
is
p
a
p
er
w
a
s
b
ased
o
n
a
d
atab
ase
p
r
o
v
id
in
g
d
etails
o
n
p
u
r
c
h
ase
s
m
ad
e
o
n
an
E
-
co
m
m
er
ce
p
latf
o
r
m
o
v
e
r
a
p
er
io
d
o
f
o
n
e
y
ea
r
.
E
ac
h
en
tr
y
i
n
t
h
e
d
ataset
d
escr
ib
ed
th
e
p
u
r
c
h
ase
o
f
a
p
r
o
d
u
ct,
b
y
a
p
ar
ticu
lar
cu
s
t
o
m
er
a
n
d
at
a
g
iv
e
n
d
ate.
Giv
e
n
th
e
a
v
ailab
le
in
f
o
r
m
at
io
n
,
a
m
o
d
el
w
a
s
d
ev
elo
p
ed
th
at
allo
w
ed
th
e
p
r
ed
ictio
n
o
f
th
e
t
y
p
e
o
f
p
u
r
c
h
ase
t
h
at
a
cu
s
to
m
er
w
o
u
ld
m
ak
e
a
s
w
e
ll
as
th
e
n
u
m
b
er
o
f
v
i
s
it
s
th
at
h
e/
s
h
e
wo
u
ld
m
ak
e
d
u
r
in
g
a
y
ea
r
,
r
ig
h
t
f
r
o
m
i
ts
f
ir
s
t
v
i
s
it to
th
e
E
-
co
m
m
er
ce
s
ite.
Fin
all
y
,
t
h
e
q
u
alit
y
o
f
t
h
ese
p
r
ed
ictio
n
s
o
f
th
e
d
i
f
f
er
e
n
t
c
lass
i
f
ier
s
w
er
e
tes
ted
o
v
er
th
e
last
t
w
o
m
o
n
t
h
s
o
f
th
e
d
ataset.
T
h
e
d
ata
w
as
t
h
e
n
p
r
o
ce
s
s
ed
in
t
w
o
s
t
ep
s
:
f
ir
s
t,
all
t
h
e
d
ata
w
as
co
n
s
id
er
ed
(
o
v
er
th
e
2
m
o
n
t
h
s
)
to
d
ef
i
n
e
th
e
ca
teg
o
r
y
to
w
h
ich
ea
ch
clie
n
t
b
e
lo
n
g
s
,
a
n
d
th
e
n
,
t
h
e
clas
s
if
i
er
p
r
ed
ictio
n
s
w
er
e
co
m
p
ar
ed
w
it
h
t
h
is
ca
te
g
o
r
y
a
s
s
i
g
n
m
e
n
t.
I
t
w
as
f
o
u
n
d
t
h
at
7
5
% o
f
clie
n
ts
w
er
e
a
w
ar
d
ed
th
e
r
ig
h
t c
lass
e
s
.
T
h
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
c
lass
if
ier
t
h
er
ef
o
r
e
s
ee
m
ed
c
o
r
r
ec
t
g
iv
e
n
th
e
p
o
ten
tia
l
s
h
o
r
tco
m
i
n
g
s
o
f
th
e
c
u
r
r
en
t
m
o
d
el.
I
n
p
ar
ticu
lar
,
a
b
ias
th
at
h
ad
n
o
t
b
ee
n
d
ea
lt
w
it
h
,
w
er
e
t
h
e
co
n
ce
r
n
s
o
n
t
h
e
s
ea
s
o
n
al
it
y
o
f
p
u
r
ch
a
s
es
an
d
th
e
f
ac
t
t
h
at
p
u
r
ch
a
s
in
g
h
ab
it
s
wo
u
ld
p
o
ten
tiall
y
d
ep
en
d
o
n
t
h
e
ti
m
e
o
f
y
ea
r
(
f
o
r
ex
a
m
p
l
e,
f
esti
v
al
ti
m
es
lik
e
C
h
r
i
s
t
m
as,
Di
w
a
li
etc.
)
.
I
n
p
r
ac
tice,
th
i
s
s
ea
s
o
n
al
e
f
f
ec
t
m
a
y
ca
u
s
e
t
h
e
ca
te
g
o
r
ies
d
ef
i
n
ed
o
v
er
a
1
0
-
m
o
n
t
h
p
er
io
d
to
b
e
q
u
ite
d
if
f
er
e
n
t
f
r
o
m
t
h
o
s
e
ex
tr
ap
o
lated
f
r
o
m
th
e
last
t
w
o
m
o
n
t
h
s
.
I
n
o
r
d
er
to
co
r
r
ec
t
s
u
ch
b
ias,
it
w
o
u
ld
b
e
b
en
ef
icial
to
h
a
v
e
d
a
ta
th
at
w
o
u
ld
co
v
er
a
lo
n
g
er
p
er
io
d
o
f
tim
e.
RE
F
E
R
E
NC
E
S
[1
]
O.
Et
z
io
n
i,
“
T
h
e
w
o
rld
-
w
id
e
w
e
b
:
Qu
a
g
m
ire
o
r
g
o
ld
m
in
e
”
,
Co
mm
u
n
ica
ti
o
n
s
o
f
th
e
ACM
,
v
o
l
.
39
,
n
o
.
11
,
1
9
9
6
,
pp.
65
-
6
8
.
[2
]
M
.
Ei
rin
k
i,
M
.
V
a
z
irg
ian
n
is,
“
Web
m
in
in
g
f
o
r
w
e
b
p
e
rso
n
a
li
z
a
ti
o
n
”
,
ACM
T
ra
n
sa
c
ti
o
n
s
o
n
In
ter
n
e
t
T
e
c
h
n
o
lo
g
y
,
v
o
l.
3
,
n
o
.
1
,
2
0
0
3
,
p
p
.
1
-
2
7
.
[3
]
M
.
He
n
z
in
g
e
r,
“
L
in
k
a
n
a
l
y
s
is
in
w
e
b
in
f
o
rm
a
ti
o
n
re
tri
e
v
a
l
”
,
Bu
ll
e
ti
n
o
f
th
e
tec
h
n
ica
l
c
o
m
m
it
tee
o
n
d
a
ta
e
n
g
in
e
e
rin
g
”
,
IEE
E
Co
m
p
u
ter
S
o
c
iet
y
,
v
o
l.
23
,
2
0
0
0
,
p
p
.
3
-
9.
[4
]
D.
S
h
e
n
,
Y.
C
o
n
g
,
J.
-
T
.
S
u
n
,
Y.
-
C.
L
u
,
“
S
t
u
d
ies
o
n
C
h
i
n
e
se
w
e
b
p
a
g
e
c
las
si
f
ica
ti
o
n
”
,
in
:
Pro
c
e
e
d
i
n
g
s
o
f
t
h
e
2
0
0
3
In
ter
n
a
t
io
n
a
l
C
o
n
fer
e
n
c
e
o
n
M
a
c
h
in
e
L
e
a
rn
i
n
g
a
n
d
Cy
b
e
rn
e
ti
c
s
,
v
o
l.
1
,
2
0
0
3
,
p
p
.
2
3
-
27.
[5
]
P
.
V
a
h
d
a
n
i
A
m
o
li
a
n
d
O.
S
o
j
o
o
d
i
S
h
.
,
“
S
c
ien
ti
f
ic
Do
c
u
m
e
n
ts
Clu
ste
rin
g
Ba
se
d
o
n
T
e
x
t
S
u
m
m
a
ri
z
a
ti
o
n
”
,
In
t
e
rn
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
E
lec
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
5
,
n
o
.
4
,
p
p
.
7
8
2
-
7
8
7
,
2
0
1
5
.
[6
]
H.
M
a
n
n
il
a
,
H.
T
o
iv
o
n
e
n
,
A
.
I.
V
e
rk
a
m
o
,
“
Disc
o
v
e
rin
g
f
re
q
u
e
n
t
e
p
iso
d
e
s
in
se
q
u
e
n
c
e
s”
,
in
:
Pro
c
e
e
d
in
g
s
o
f
t
h
e
Fi
rs
t
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Kn
o
wle
d
g
e
a
n
d
Da
t
a
M
in
i
n
g
,
1
9
9
5
,
p
p
.
2
1
0
-
2
1
5
.
[7
]
Ra
v
i
k
u
m
a
r
V
.
,
a
n
d
K.
Ra
g
h
u
v
e
e
r,
“
L
e
g
a
l
Do
c
u
m
e
n
ts
Clu
ste
rin
g
a
n
d
S
u
m
m
a
riz
a
ti
o
n
u
sin
g
Hie
r
a
rc
h
ica
l
L
a
ten
t
Dirich
let
A
ll
o
c
a
ti
o
n
”
,
IA
ES
I
n
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Arti
fi
c
ia
l
I
n
te
ll
ig
e
n
c
e
(
IJ
-
AI)
,
v
o
l.
2
,
n
o
.
1
,
p
p
.
2
7
-
3
5
,
2
0
1
3
.
[8
]
R
.
Bu
rk
e
,
“
H
y
b
rid
re
c
o
m
m
e
n
d
e
r
s
y
ste
m
s: su
rv
e
y
a
n
d
e
x
p
e
ri
m
e
n
ts
”
,
Us
e
r
M
o
d
e
ll
in
g
a
n
d
Us
e
r
-
Ad
a
p
ted
In
ter
a
c
ti
o
n
,
v
o
l.
12
,
n
o
.
4
,
2
0
0
2
,
p
p
.
3
3
1
-
3
7
0
.
[9
]
H.
Da
i,
B.
M
o
b
a
sh
e
r,
“
A
ro
a
d
m
a
p
to
m
o
re
e
ff
e
c
ti
v
e
w
e
b
p
e
rso
n
a
li
z
a
ti
o
n
:
In
teg
ra
ti
n
g
d
o
m
a
in
k
n
o
w
l
e
d
g
e
w
it
h
w
e
b
u
sa
g
e
m
in
in
g
”
,
in
:
I
n
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
In
ter
n
e
t
Co
mp
u
ti
n
g
,
2
0
0
3
,
p
p
.
5
8
-
6
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
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
9
0
–
2
3
9
8
2398
[1
0
]
W
.
L
in
,
S
.
A
.
A
l
v
a
re
z
,
C.
Ru
iz,
“
Co
ll
a
b
o
ra
ti
v
e
re
c
o
m
m
e
n
d
a
ti
o
n
v
ia
a
d
a
p
t
iv
e
a
ss
o
c
iatio
n
r
u
l
e
m
in
in
g
”
,
in
:
W
EB
KDD
2
0
0
0
–
W
e
b
M
in
in
g
fo
r
E
-
Co
mm
e
rc
e
–
Ch
a
ll
e
n
g
e
s
a
n
d
Op
p
o
rtu
n
it
ies
,
S
e
c
o
n
d
In
ter
n
a
ti
o
n
a
l
W
o
rk
sh
o
p
,
Bo
sto
n
,
M
A
,
USA
,
2
0
0
0
.
[1
1
]
O.R.
Zaia
n
e
,
M
.
X
i
n
,
J.
Ha
n
,
“
Disc
o
v
e
rin
g
W
e
b
a
c
c
e
ss
p
a
tt
e
rn
s
a
n
d
tren
d
s
b
y
a
p
p
ly
in
g
OLA
P
a
n
d
d
a
ta
m
in
in
g
tec
h
n
o
l
o
g
y
o
n
W
e
b
lo
g
s”
,
in
:
A
d
v
a
n
c
e
s in
Dig
it
a
l
L
ib
ra
ries
,
S
a
n
ta
Ba
rb
a
ra
,
CA
,
US
A
,
1
9
9
8
,
p
p
.
1
9
-
29.
[1
2
]
B.
Zh
o
u
,
S
.
C
.
Hu
i,
K.
Ch
a
n
g
,
“
A
n
in
telli
g
e
n
t
re
c
o
m
m
e
n
d
e
r
s
y
st
e
m
u
sin
g
se
q
u
e
n
ti
a
l
w
e
b
a
c
c
e
ss
p
a
tt
e
rn
s”
,
i
n
:
Pro
c
e
e
d
in
g
s
o
f
t
h
e
2
0
0
4
IEE
E
Co
n
fer
e
n
c
e
o
n
Cy
b
e
rn
e
t
ics
a
n
d
I
n
tel
li
g
e
n
t
S
y
ste
ms
,
S
in
g
a
p
o
re
,
2
0
0
4
,
p
p
.
1
-
3.
[1
3
]
H.
Ish
ik
a
w
a
,
T
.
Na
k
a
ji
m
a
,
T
.
M
izu
h
a
ra
,
S
.
Y
o
k
o
y
a
m
a
,
J.
Na
k
a
y
a
m
a
,
M
.
Oh
ta,
K.
Ka
tay
a
m
a
,
“
A
n
in
telli
g
e
n
t
w
e
b
re
c
o
m
m
e
n
d
a
ti
o
n
sy
ste
m
:
A
w
e
b
u
sa
g
e
m
in
in
g
a
p
p
ro
a
c
h
”
,
in
:
IS
M
I
S
,
2
0
0
2
,
p
p
.
3
4
2
-
3
5
0
.
[1
4
]
R.
M
e
tere
n
,
M
.
S
o
m
e
re
n
,
“
Us
in
g
c
o
n
ten
t
-
b
a
se
d
f
il
terin
g
f
o
r
re
c
o
m
m
e
n
d
a
ti
o
n
”
,
i
n
:
Pro
c
e
e
d
in
g
s
o
f
M
L
n
e
t/
ECM
L
2
0
0
0
W
o
rk
sh
op
,
Ba
rc
e
lo
n
a
,
S
p
a
in
,
3
0
M
a
y
2
0
0
0
.
[1
5
]
J.
L
i,
O.R.
Zaı¨a
n
e
,
“
c
o
m
b
in
in
g
u
sa
g
e
,
c
o
n
ten
t,
a
n
d
stru
c
t
u
re
d
a
ta
to
im
p
ro
v
e
we
b
site
re
c
o
m
m
e
n
d
a
ti
o
n
”
,
in
:
EC
-
W
e
b
,
2
0
0
4
,
p
p
.
3
0
5
–
3
1
5
.
[1
6
]
W
.
Co
h
e
n
,
A
.
M
c
Ca
ll
u
m
,
D.
Qu
a
ss
,
“
L
e
a
rn
in
g
to
u
n
d
e
rsta
n
d
t
h
e
w
e
b
”
,
IEE
E
Da
ta
En
g
in
e
e
r
in
g
B
u
ll
e
ti
n
,
v
o
l.
23
,
2
0
0
0
,
p
p
.
17
-
2
4
.
[1
7
]
S
.
K.
M
a
d
ria,
S
.
S
.
B
h
o
w
m
ic
k
,
W.
K.
Ng
,
E.
P
.
L
im
,
“
Re
se
a
rc
h
issu
e
s
in
W
e
b
d
a
ta
m
in
in
g
”
,
in
:
Pr
o
c
e
e
d
in
g
s
o
f
t
h
e
Fi
rs
t
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Da
ta
W
a
re
h
o
u
sin
g
a
n
d
K
n
o
w
led
g
e
Disc
o
v
e
ry
(
Da
W
a
K’9
9
)
,
1
9
9
9
,
p
p
.
3
0
3
-
3
1
2
.
[1
8
]
V
.
Ke
sˇe
lj
,
F
.
P
e
n
g
,
N.
Ce
rc
o
n
e
,
C.
T
h
o
m
a
s,
“N
-
g
ra
m
-
b
a
se
d
a
u
th
o
r
p
ro
f
il
e
s
f
o
r
a
u
th
o
rs
h
ip
a
tt
rib
u
ti
o
n
”
,
in
:
Pro
c
e
e
d
in
g
s
o
f
t
h
e
Co
n
fer
e
n
c
e
P
a
c
if
ic A
ss
o
c
ia
t
io
n
f
o
r Co
mp
u
t
a
ti
o
n
a
l
L
in
g
u
isti
c
s
,
No
v
a
S
c
o
ti
a
,
Ca
n
a
d
a
,
2
0
0
3
.
[1
9
]
Da
s,
S
.
,
M
a
th
e
w
,
M
.
a
n
d
V
ij
a
y
a
r
a
g
h
a
v
a
n
,
P
.
(
2
0
1
7
).
“
A
n
Eff
icie
n
t
A
p
p
ro
a
c
h
f
o
r
F
in
d
in
g
n
e
a
r
Du
p
li
c
a
te W
e
b
p
a
g
e
s
u
sin
g
M
in
im
u
m
Weig
h
t
Ov
e
rla
p
p
i
n
g
M
e
th
o
d
”
.
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
C
o
mp
u
ter
En
g
in
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
1
,
p
p
.
1
8
7
-
1
9
4
,
2
0
1
1
.
[2
0
]
Zh
a
n
g
,
L
.
,
Ya
n
g
,
S
.
a
n
d
Zh
a
n
g
,
M
.
(2
0
1
8
).
“
E
-
c
o
m
m
e
rc
e
W
e
b
sit
e
Re
c
o
m
m
e
n
d
e
r
S
y
ste
m
Ba
s
e
d
o
n
Diss
im
il
a
rit
y
a
n
d
A
ss
o
c
iatio
n
Ru
le”
.
In
d
o
n
e
sia
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
Co
m
p
u
ter
En
g
i
n
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
1
2
,
p
p
.
3
5
3
-
3
6
0
,
2
0
1
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.