I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
1
5
,
No
.
5
,
Octo
b
er
2
0
2
6
,
p
p
.
4
0
7
1
~
4
0
8
0
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
15
.i
5
.
p
p
4
0
7
1
-
4
0
8
0
4071
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
An
int
ellig
en
t
a
r
ti
ficial int
ellig
ence
framewo
rk
for bu
sines
s
insig
ht
g
eneratio
n f
ro
m st
ream
ing
data
B
eny
m
o
l J
o
s
e
1
,
Sa
j
imo
n Abra
ha
m
2
,
Nik
i
t
ha
Sa
j
it
h
3
1
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
A
p
p
l
i
c
a
t
i
o
n
s,
M
a
r
i
a
n
C
o
l
l
e
g
e
K
u
t
t
i
k
k
a
n
a
m
A
u
t
o
n
o
m
o
u
s,
K
u
t
t
i
k
k
a
n
a
m
,
I
n
d
i
a
2
S
c
h
o
o
l
o
f
M
a
n
a
g
e
m
e
n
t
a
n
d
B
u
si
n
e
ss
S
t
u
d
i
e
s,
M
a
h
a
t
m
a
G
a
n
d
h
i
U
n
i
v
e
r
s
i
t
y
,
K
o
t
t
a
y
a
m,
I
n
d
i
a
3
LTM
L
i
mi
t
e
d
,
B
e
n
g
a
l
u
r
u
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Dec
1
5
,
2
0
2
5
R
ev
is
ed
Au
g
1
,
2
0
2
6
Acc
ep
ted
Au
g
1
2
,
2
0
2
6
Big
d
a
ta
a
n
a
l
y
ti
c
s
h
e
lp
s
o
rg
a
n
iz
a
ti
o
n
s
lev
e
ra
g
e
t
h
e
ir
d
a
ta
t
o
u
n
c
o
v
e
r
n
e
w
o
p
p
o
rt
u
n
it
ies
,
lea
d
in
g
to
sm
a
rter
d
e
c
isio
n
s,
m
o
re
e
fficie
n
t
o
p
e
ra
ti
o
n
s,
a
n
d
u
lt
ima
tely
h
ig
h
e
r
p
r
o
fit
s.
Eff
e
c
ti
v
e
u
se
o
f
b
i
g
d
a
ta
a
n
d
c
re
a
ti
n
g
imp
o
rtan
t
p
re
d
ictiv
e
m
o
d
e
ls
m
a
k
e
it
m
o
re
stra
ig
h
tf
o
rwa
rd
f
o
r
a
b
u
si
n
e
ss
to
re
a
li
z
e
wh
a
t
th
e
c
o
n
su
m
e
r
li
k
e
s
a
n
d
wh
a
t
h
e
o
r
s
h
e
d
o
e
s
n
o
t
.
T
h
is
will
e
n
a
b
le
t
h
e
b
u
sin
e
ss
t
o
p
r
o
p
o
se
b
e
tt
e
r
o
ffe
rs
a
n
d
a
d
m
i
n
istratio
n
s
wit
h
th
e
ir
it
e
m
s
wh
e
n
c
o
n
tras
ted
to
it
s
c
o
m
p
e
ti
to
rs.
In
t
h
is
p
a
p
e
r,
t
h
e
a
u
th
o
rs
su
g
g
e
st
th
e
u
se
o
f
a
b
ig
d
a
ta
a
n
a
ly
t
ics
p
latfo
rm
c
o
m
b
in
e
d
wit
h
n
a
tu
ra
l
lan
g
u
a
g
e
p
ro
c
e
ss
in
g
(NLP
)
tec
h
n
i
q
u
e
s
to
d
o
t
h
e
in
tell
ig
e
n
t
p
ro
c
e
ss
in
g
o
f
u
n
stru
c
t
u
re
d
stre
a
m
in
g
d
a
ta.
A
n
e
w
m
e
th
o
d
u
ti
li
z
i
n
g
Ap
a
c
h
e
S
p
a
rk
p
i
p
e
li
n
i
n
g
is
p
r
o
p
o
se
d
to
a
c
h
iev
e
p
re
d
icti
v
e
a
n
a
l
y
ti
c
s
f
o
r
u
n
stru
c
tu
re
d
d
a
ta
stre
a
m
s.
Wi
th
th
is
p
re
d
ictiv
e
m
o
d
e
l,
m
in
i
n
g
o
f
t
h
e
u
n
stru
c
tu
re
d
stre
a
m
in
g
d
a
ta
c
a
n
b
e
d
o
n
e
a
n
d
in
telli
g
e
n
t
p
re
d
ictio
n
s
a
re
m
a
d
e
p
o
ss
ib
le
i
n
re
a
l
ti
m
e
u
sin
g
a
rti
ficia
l
in
telli
g
e
n
c
e
tec
h
n
i
q
u
e
s.
K
ey
w
o
r
d
s
:
C
h
am
p
io
n
m
o
d
el
Gin
i c
o
ef
f
icien
t
Ma
ch
in
e
lear
n
in
g
p
ip
elin
e
Natu
r
al
lan
g
u
ag
e
p
r
o
ce
s
s
in
g
tech
n
iq
u
es
R
ec
ei
v
er
o
p
er
ati
n
g
c
h
a
r
a
cte
r
is
tic
cu
r
v
e
Stre
am
in
g
d
ata
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
:
B
en
y
m
o
l Jo
s
e
Dep
ar
tm
en
t o
f
C
o
m
p
u
ter
A
p
p
licatio
n
s
,
Ma
r
ian
C
o
lleg
e
Ku
ttik
k
an
am
Au
t
o
n
o
m
o
u
s
Ku
ttik
k
an
am
P.O,
Peer
m
ad
e
,
I
d
u
k
k
i D
is
tr
ict,
Ker
ala,
I
n
d
ia
E
m
ail:
b
en
y
m
o
l.j
o
s
e@
m
ar
ian
co
lleg
e.
o
r
g
1.
I
NT
RO
D
UCT
I
O
N
Stre
am
co
m
p
u
tin
g
r
ep
r
esen
ts
a
m
o
d
e
r
n
p
ar
a
d
ig
m
d
r
iv
e
n
b
y
em
er
g
in
g
d
ata
-
g
en
er
atio
n
s
ce
n
ar
io
s
s
u
ch
as
th
e
in
ter
n
et
o
f
th
in
g
s
(
I
o
T
)
,
s
en
s
o
r
n
etwo
r
k
s
,
f
in
a
n
cial
m
ar
k
et
f
ee
d
s
,
m
o
b
ile
d
ev
ices,
g
eo
s
p
atial
s
er
v
ices,
an
d
s
o
cial
m
ed
ia
p
latf
o
r
m
s
[
1
]
.
T
h
is
ap
p
r
o
ac
h
is
b
ased
o
n
t
h
e
k
ey
p
r
em
is
e
th
at
t
h
e
v
alu
e
o
f
d
ata
is
h
ig
h
est
wh
en
it
is
m
o
s
t
r
ec
en
t
[
2
]
.
Da
ta
p
r
o
d
u
ce
d
co
n
tin
u
o
u
s
ly
is
k
n
o
wn
as
s
tr
ea
m
in
g
d
ata,
an
d
it
m
u
s
t
b
e
an
aly
ze
d
im
m
ed
iately
as
it
ar
r
iv
es
to
g
en
er
ate
tim
ely
in
s
ig
h
ts
.
T
o
h
an
d
le
s
u
ch
p
r
o
ce
s
s
in
g
d
em
a
n
d
s
,
r
o
b
u
s
t
p
ar
allel
ar
ch
itectu
r
es
,
an
d
s
ca
lab
le
co
m
p
u
tin
g
s
y
s
tem
s
ar
e
ess
e
n
tial
[
3
]
.
Stre
am
co
m
p
u
tin
g
en
ab
les
o
n
g
o
in
g
ev
alu
atio
n
o
f
s
tatic
q
u
er
ies
ag
ain
s
t
d
ata
t
h
at
is
co
n
s
tan
tly
in
m
o
tio
n
[
4
]
.
Alth
o
u
g
h
v
ast
am
o
u
n
ts
o
f
u
n
s
tr
u
ctu
r
ed
d
ata
ar
e
p
r
o
d
u
ce
d
,
ex
t
r
ac
tin
g
r
elev
an
t
an
d
m
ea
n
in
g
f
u
l
in
f
o
r
m
atio
n
f
r
o
m
th
ese
s
tr
ea
m
s
is
cr
itical
f
o
r
m
ak
i
n
g
p
r
o
m
p
t d
ec
is
io
n
s
.
Su
ch
tim
ely
d
ec
is
io
n
-
m
a
k
in
g
i
s
p
o
s
s
ib
le
o
n
ly
wh
en
in
co
m
in
g
d
ata
is
p
r
o
ce
s
s
ed
,
an
aly
ze
d
,
a
n
d
v
is
u
alize
d
i
n
r
e
al
tim
e
[
5
]
.
Dem
ir
ez
en
an
d
Nav
r
u
z
[
6
]
s
u
g
g
est
em
p
lo
y
in
g
th
e
lam
b
d
a
(
λ
)
ar
ch
itectu
r
e,
b
u
ilt
o
n
to
p
o
f
h
ad
o
o
p
,
to
s
u
p
p
o
r
t
n
ea
r
r
ea
l
-
tim
e
d
ata
p
r
o
ce
s
s
in
g
th
at
ac
c
o
m
m
o
d
ates
b
o
th
s
tr
ea
m
in
g
a
n
d
b
a
tch
an
aly
tics
.
T
h
is
ar
ch
itectu
r
e
co
m
p
r
is
es
th
r
ee
p
r
im
ar
y
lay
er
s
k
n
o
wn
as
th
e
b
a
tch
lay
er
,
th
e
s
p
ee
d
lay
er
,
a
n
d
th
e
s
er
v
in
g
lay
er
.
T
h
e
b
atch
lay
er
g
en
er
ates
p
r
e
co
m
p
u
ted
r
esu
lts
u
s
in
g
a
d
is
tr
ib
u
ted
p
r
o
ce
s
s
in
g
f
r
am
ewo
r
k
an
d
en
s
u
r
es
h
ig
h
ac
cu
r
ac
y
b
y
h
a
n
d
lin
g
th
e
f
u
l
l
d
ataset
[
7
]
.
T
h
e
s
p
ee
d
lay
er
m
an
ag
es
in
co
m
in
g
d
ata
s
tr
ea
m
s
in
r
ea
l
tim
e,
co
m
p
en
s
atin
g
f
o
r
th
e
laten
cy
in
h
er
en
t
i
n
th
e
b
atch
lay
er
.
B
ec
au
s
e
it
f
o
c
u
s
es
o
n
r
ap
id
o
u
t
p
u
t,
its
r
esu
lts
m
ay
b
e
less
p
r
ec
is
e
o
r
in
co
m
p
lete
co
m
p
ar
e
d
to
th
o
s
e
p
r
o
d
u
ce
d
b
y
th
e
b
atch
lay
e
r
,
an
d
th
es
e
f
ast
o
u
tp
u
ts
ar
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
5
,
Octo
b
er
2
0
2
6
:
4
0
7
1
-
4
0
8
0
4072
f
r
eq
u
e
n
tly
s
to
r
ed
in
n
o
t
o
n
l
y
s
tr
u
ctu
r
ed
q
u
e
r
y
lan
g
u
ag
e
(
No
SQL
)
s
y
s
tem
s
[
8
]
,
s
u
ch
as
Mo
n
g
o
DB
,
Vo
ltDB
,
an
d
C
ass
an
d
r
a.
T
h
e
r
esu
lts
f
r
o
m
th
e
b
atch
a
n
d
s
p
ee
d
lay
e
r
s
ar
e
k
ep
t
in
th
e
s
er
v
in
g
lay
er
,
wh
ich
an
s
wer
s
q
u
er
ies
b
y
r
etu
r
n
in
g
p
r
e
-
co
m
p
u
ted
in
te
r
p
r
etatio
n
s
f
r
o
m
th
e
p
r
o
ce
s
s
ed
d
ata
[
9
]
,
a
n
d
r
ep
r
o
ce
s
s
in
g
ev
er
y
b
atch
cy
cle
is
n
o
t b
en
e
f
icial
in
m
an
y
s
ce
n
ar
io
s
[
1
0
]
.
Ap
ac
h
e
Sp
ar
k
is
a
r
o
b
u
s
t
o
p
e
n
-
s
o
u
r
ce
en
g
in
e
ca
p
a
b
le
o
f
h
a
n
d
lin
g
b
o
th
b
atc
h
an
d
s
tr
ea
m
p
r
o
ce
s
s
in
g
task
s
[
1
0
]
.
I
n
th
eir
r
ec
e
n
t
s
tu
d
y
,
Kam
al
et
a
l.
[
1
1
]
em
p
l
o
y
Ap
ac
h
e
S
p
ar
k
f
o
r
r
ea
l
-
tim
e
d
ata
p
r
o
ce
s
s
in
g
co
m
b
in
ed
with
te
x
t
m
in
in
g
an
d
n
atu
r
al
lan
g
u
a
g
e
p
r
o
ce
s
s
in
g
(
NL
P)
.
T
h
eir
wo
r
k
aim
s
to
illu
s
tr
ate
h
o
w
b
u
s
in
ess
in
tellig
en
ce
ca
n
b
e
en
h
an
ce
d
b
y
g
ath
er
i
n
g
an
d
an
aly
zin
g
web
d
ata
in
r
ea
l
tim
e.
An
o
th
er
in
v
esti
g
atio
n
b
y
Sv
y
atk
o
v
s
k
i
y
et
a
l.
[
1
2
]
ass
ess
es
th
e
Ap
ac
h
e
S
p
ar
k
f
r
am
ewo
r
k
f
o
r
d
ata
-
in
ten
s
iv
e
m
ac
h
in
e
lear
n
in
g
(
ML
)
wo
r
k
lo
ad
s
.
T
h
ey
p
r
o
p
o
s
e
a
tex
t
-
p
r
o
ce
s
s
in
g
p
ip
elin
e
th
at
in
co
r
p
o
r
ates
th
e
Av
r
o
s
er
ializatio
n
f
r
am
ewo
r
k
,
Sp
ar
k
ML
,
Gr
ap
h
Fra
m
es,
an
d
th
e
h
is
to
g
r
am
m
a
r
s
u
ite
o
f
d
ata
ag
g
r
eg
atio
n
to
o
ls
,
ap
p
ly
in
g
it
to
a
p
o
licy
d
if
f
u
s
io
n
p
r
o
b
lem
.
I
n
a
s
u
r
v
ey
c
o
n
d
u
cted
b
y
Ah
m
e
d
e
t
a
l.
[
1
3
]
,
t
h
e
a
u
th
o
r
s
d
is
cu
s
s
th
e
id
ea
o
f
b
i
g
d
ata
an
d
th
e
n
u
m
er
o
u
s
f
o
r
m
s
o
f
a
n
aly
tics
th
at
r
esear
ch
er
s
ca
n
p
e
r
f
o
r
m
u
s
in
g
lar
g
e
-
s
ca
le
d
ata.
T
h
eir
wo
r
k
o
u
tlin
es
s
tr
ateg
ies
f
o
r
im
p
r
o
v
i
n
g
th
e
p
er
f
o
r
m
an
ce
o
f
b
ig
d
ata
p
r
o
c
ess
i
n
g
an
d
e
x
p
lain
s
d
if
f
er
en
t
b
ig
d
ata
a
n
aly
tics
to
o
ls
.
I
n
an
o
th
er
ap
p
r
o
ac
h
,
Azh
ir
et
a
l.
[
1
4
]
in
v
esti
g
ate
b
ig
d
ata
an
aly
s
is
an
d
h
ig
h
lig
h
t
th
e
im
p
o
r
tan
ce
o
f
Ap
ac
h
e
S
p
ar
k
i
n
r
ea
l
-
tim
e
p
r
o
ce
s
s
in
g
b
y
ex
tr
ac
tin
g
m
ea
n
in
g
f
u
l in
s
ig
h
ts
f
r
o
m
X
(
f
o
r
m
er
ly
T
witter
)
d
ata.
T
h
e
wo
r
k
p
r
esen
ted
b
y
Nair
et
a
l.
[
1
5
]
ce
n
ter
s
o
n
d
ev
el
o
p
in
g
a
r
ea
l
-
tim
e
s
y
s
tem
f
o
r
f
o
r
ec
asti
n
g
r
em
o
te
h
ea
lth
co
n
d
itio
n
s
,
wh
ich
is
im
p
lem
en
ted
u
s
in
g
th
e
o
p
en
-
s
o
u
r
ce
b
ig
d
ata
p
r
o
ce
s
s
in
g
f
r
am
ewo
r
k
Ap
ac
h
e
S
p
ar
k
a
n
d
h
o
s
ted
w
ith
in
a
cl
o
u
d
-
b
ased
in
f
r
astru
ctu
r
e.
I
n
t
h
is
s
ca
lab
le
f
r
am
e
wo
r
k
,
a
ML
m
o
d
e
l
d
eter
m
in
es
a
u
s
er
’
s
h
ea
lth
co
n
d
itio
n
b
ased
o
n
h
ea
lth
-
r
elate
d
twee
ts
th
at
th
e
ap
p
licatio
n
r
ec
eiv
es
in
s
tan
tly
.
Ko
lajo
et
a
l.
[
1
6
]
p
r
esen
t
a
s
y
s
tem
atic
r
ev
iew
o
f
b
ig
d
ata
s
tr
ea
m
an
aly
s
is
an
d
o
b
s
er
v
e
th
at,
alth
o
u
g
h
s
u
b
s
tan
tial
r
esear
ch
h
as
co
n
c
en
tr
ated
o
n
b
atc
h
p
r
o
ce
s
s
in
g
,
th
er
e
is
g
r
o
win
g
in
ter
est
in
ex
am
in
in
g
d
ata
in
m
o
tio
n
.
T
h
eir
r
ev
iew
p
r
o
v
id
e
s
a
co
m
p
r
e
h
en
s
iv
e
ass
ess
m
en
t
o
f
th
e
cu
r
r
en
t
s
tate
o
f
s
tr
ea
m
p
r
o
ce
s
s
in
g
an
d
r
ec
o
m
m
en
d
s
th
at
r
esear
ch
er
s
d
ev
o
te
g
r
ea
ter
atten
tio
n
to
em
p
ir
ical
ev
alu
atio
n
s
o
f
b
ig
d
ata
s
tr
ea
m
in
g
to
o
ls
an
d
tech
n
o
lo
g
ies.
I
n
an
o
th
er
co
n
tr
ib
u
tio
n
,
Ma
tteu
s
s
i
et
a
l.
[
1
7
]
s
u
cc
ess
f
u
lly
d
ev
elo
p
ed
a
n
d
d
em
o
n
s
tr
ated
a
s
ca
lab
le
s
y
s
tem
f
o
r
r
ea
l
-
tim
e
an
aly
s
is
an
d
f
ilter
in
g
o
f
jo
b
-
r
elate
d
twee
ts
f
r
o
m
am
o
n
g
m
i
llio
n
s
o
f
in
co
m
i
n
g
s
tr
ea
m
in
g
twee
ts
.
T
h
ese
twee
ts
ar
e
th
en
ca
teg
o
r
ized
in
t
o
v
ar
io
u
s
jo
b
class
if
icatio
n
s
,
th
er
eb
y
en
h
a
n
cin
g
th
e
ef
f
icien
cy
o
f
jo
b
s
ea
r
c
h
in
g
th
r
o
u
g
h
Sp
a
r
k
s
tr
ea
m
in
g
.
T
h
is
p
ap
er
s
h
o
wed
th
at
NL
P
m
eth
o
d
s
ca
n
b
e
a
p
p
lied
in
b
ig
d
ata
p
latf
o
r
m
s
s
u
ch
as
Sp
ar
k
to
ex
tr
ac
t
in
s
ig
h
ts
f
r
o
m
u
n
s
tr
u
ctu
r
e
d
tex
t
an
d
u
n
co
v
er
v
alu
a
b
le
an
d
m
ea
n
in
g
f
u
l
p
atter
n
s
f
o
r
b
u
s
in
ess
in
tellig
en
ce
.
I
n
th
is
wo
r
k
,
p
r
ed
ictiv
e
an
aly
tic
s
is
ap
p
lied
to
u
n
s
tr
u
ctu
r
ed
s
t
r
ea
m
in
g
d
ata
u
s
in
g
v
ar
i
o
u
s
ML
m
o
d
els
with
in
a
Sp
ar
k
ML
p
ip
elin
e.
T
h
e
m
o
d
els’
p
er
f
o
r
m
a
n
ce
s
ar
e
th
en
co
m
p
ar
ed
u
s
in
g
e
v
alu
atio
n
m
et
r
ics
s
u
ch
as
r
ec
eiv
er
o
p
er
atin
g
ch
ar
ac
ter
is
tic
(
R
OC
)
cu
r
v
es
an
d
Gin
i
co
ef
f
icien
t
s
.
T
h
e
b
est
-
p
e
r
f
o
r
m
in
g
m
o
d
el
is
s
elec
ted
as
th
e
ch
am
p
io
n
m
o
d
el
w
h
ich
ca
n
b
e
d
ep
lo
y
e
d
f
o
r
s
m
ar
ter
b
u
s
in
e
s
s
in
tellig
en
ce
b
ec
au
s
e
in
s
ig
h
ts
g
ain
ed
f
r
o
m
th
is
an
aly
s
is
o
f
f
er
o
r
g
an
izatio
n
s
a
clea
r
v
iew
o
f
v
ar
i
o
u
s
d
im
en
s
io
n
s
o
f
th
eir
o
p
er
atio
n
s
an
d
cu
s
to
m
er
b
eh
a
v
io
r
,
r
an
g
in
g
f
r
o
m
s
er
v
ice
c
o
n
s
u
m
p
tio
n
a
n
d
s
er
v
er
p
er
f
o
r
m
a
n
ce
to
web
s
ite
in
ter
ac
tio
n
s
,
d
ev
ice
lo
ca
tio
n
s
,
u
s
er
m
o
v
em
en
ts
,
an
d
th
e
f
l
o
w
o
f
p
h
y
s
ical
g
o
o
d
s
,
allo
win
g
th
em
to
r
ea
ct
q
u
ick
ly
to
n
ew
em
er
g
in
g
d
ev
elo
p
m
e
n
ts
.
T
h
e
ex
tr
a
b
e
n
ef
its
o
f
s
tr
ea
m
-
p
r
o
ce
s
s
in
g
ca
p
ab
ilit
y
,
wh
ic
h
e
n
h
an
ce
s
t
h
e
v
elo
city
asp
ec
t
o
f
b
ig
d
ata
,
ar
e
m
o
r
e
ef
f
ec
tiv
ely
r
ea
lized
with
th
e
p
r
o
p
o
s
ed
m
o
d
el.
T
h
e
p
a
p
er
is
o
r
g
a
n
ized
as
f
o
ll
o
ws.
S
ec
tio
n
2
in
tr
o
d
u
ce
s
t
h
e
m
eth
o
d
o
l
o
g
y
,
th
e
p
r
o
ce
s
s
f
lo
w
d
iag
r
am
,
an
d
th
e
alg
o
r
ith
m
e
m
p
lo
y
e
d
.
T
h
e
d
ataset
u
tili
ze
d
,
alo
n
g
with
all
ex
p
er
im
en
tal
ass
ess
m
en
ts
ar
e
d
escr
ib
ed
in
s
ec
tio
n
3
.
L
astl
y
,
s
ec
t
io
n
4
p
r
o
v
id
es th
e
s
u
m
m
ar
y
an
d
t
h
e
co
n
clu
s
io
n
.
2.
M
E
T
H
O
D
2
.
1
.
I
nte
llig
ent
predict
io
ns
f
ro
m
un
s
t
ruct
ured
s
t
re
a
m
ing
da
t
a
us
ing
Sp
a
rk
m
a
chine l
ea
rning
pip
eli
nes
Sp
ar
k
s
u
p
p
o
r
ts
b
o
th
b
atch
a
n
d
s
tr
ea
m
p
r
o
ce
s
s
in
g
th
r
o
u
g
h
its
p
ip
elin
e
d
a
r
ch
itectu
r
e
,
wh
ich
is
a
f
u
n
d
am
e
n
tal
elem
en
t
o
f
c
o
m
p
u
ter
s
y
s
tem
d
esig
n
.
No
r
m
ally
,
p
ip
elin
es
ar
e
d
iv
id
ed
in
t
o
s
tag
es
,
an
d
ea
ch
s
tag
e
is
eith
er
a
‘
t
r
an
s
f
o
r
m
e
r
’
o
r
an
‘
esti
m
ato
r
’
th
at
is
r
elate
d
to
o
n
e
an
o
th
er
t
o
f
o
r
m
a
p
i
p
e
-
lik
e
s
tr
u
ctu
r
e
[
1
0
]
,
[
1
8
]
.
I
t
r
ef
er
s
to
a
s
eq
u
e
n
ce
o
f
d
ata
-
p
r
o
ce
s
s
in
g
c
o
m
p
o
n
en
ts
ar
r
an
g
ed
s
o
t
h
at
ea
ch
s
tag
e
p
ass
es
its
o
u
tp
u
t
d
ir
ec
tly
to
th
e
n
ex
t
s
tag
e
[
1
9
]
,
[
2
0
]
.
Data
o
r
in
s
tr
u
ctio
n
s
en
te
r
f
r
o
m
o
n
e
en
d
an
d
ex
it
f
r
o
m
th
e
o
th
er
,
th
er
eb
y
i
n
cr
ea
s
in
g
o
v
er
all
in
s
tr
u
ctio
n
th
r
o
u
g
h
p
u
t.
I
n
th
is
p
r
o
ce
s
s
,
a
tr
an
s
f
o
r
m
er
r
ef
er
s
to
an
alg
o
r
ith
m
th
at
co
n
v
er
ts
o
n
e
Data
Fra
m
e
in
to
an
o
t
h
er
,
wh
il
e
an
esti
m
ato
r
is
a
co
m
p
o
n
en
t
th
at
ca
n
b
e
f
itted
o
n
a
Data
Fra
m
e
to
p
r
o
d
u
ce
a
tr
an
s
f
o
r
m
er
.
I
n
th
is
co
n
tex
t,
a
lear
n
in
g
alg
o
r
ith
m
f
u
n
ctio
n
s
as
an
esti
m
ato
r
th
at
lear
n
s
p
atte
r
n
s
f
r
o
m
th
e
in
p
u
t
Data
Fra
m
e
an
d
g
en
er
ates a
tr
a
in
ed
m
o
d
el
as o
u
t
p
u
t.
Fo
r
esti
m
ato
r
s
tag
es,
th
e
f
it()
m
eth
o
d
is
ca
lled
to
p
r
o
d
u
ce
a
tr
an
s
f
o
r
m
er
,
a
n
d
th
at
tr
a
n
s
f
o
r
m
er
’
s
tr
an
s
f
o
r
m
(
)
m
eth
o
d
i
s
ca
lled
[
2
1
]
o
n
t
h
e
d
ata
f
r
a
m
e.
T
h
e
p
ip
elin
e
f
o
r
p
r
ed
ictiv
e
a
n
aly
tics
co
m
p
r
is
es
s
ev
er
al
s
eq
u
en
tial
s
tag
es,
i
n
clu
d
in
g
to
k
en
izatio
n
,
f
ea
tu
r
e
ex
tr
ac
tio
n
u
s
in
g
Hash
in
g
T
F
,
ter
m
weig
h
tin
g
th
r
o
u
g
h
i
n
v
er
s
e
d
o
cu
m
e
n
t
f
r
e
q
u
en
cy
(
I
DF
)
,
an
d
ap
p
licatio
n
o
f
ML
al
g
o
r
ith
m
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
A
n
in
tel
lig
en
t A
I
fr
a
mewo
r
k
fo
r
b
u
s
in
ess
in
s
ig
h
t g
en
era
tio
n
f
r
o
m
s
tr
ea
min
g
…
(
B
en
ymo
l
Jo
s
e
)
4073
f
o
r
m
o
d
el
t
r
ain
in
g
an
d
p
r
ed
ic
tio
n
.
T
h
e
s
tag
es
ex
ec
u
te
s
eq
u
en
tially
,
an
d
th
e
in
p
u
t
Data
Fra
m
e
r
etr
iev
e
d
f
r
o
m
th
e
d
is
tr
ib
u
ted
d
ata
s
to
r
e
is
p
r
o
g
r
ess
iv
ely
tr
an
s
f
o
r
m
ed
as it f
lo
ws th
r
o
u
g
h
ea
ch
s
tag
e.
To
m
ak
e
in
tellig
en
t
p
r
ed
ictio
n
s
with
th
e
u
n
s
tr
u
ctu
r
ed
tex
tu
al
r
ev
iews,
to
d
ec
id
e
th
e
‘
s
t
ar
_
r
atin
g
’
ass
ig
n
ed
to
ea
ch
p
r
o
d
u
ct
[
2
2
]
,
th
e
Sp
ar
k
p
i
p
elin
in
g
a
n
d
ML
class
if
icatio
n
alg
o
r
ith
m
s
wer
e
u
s
ed
.
T
h
e
o
r
ig
in
al
p
r
o
b
lem
was
a
m
u
lticlas
s
class
if
icatio
n
with
f
iv
e
s
tar
-
r
at
in
g
v
al
u
es
1
,
2
,
3
,
4
,
an
d
5
,
wh
ich
m
ad
e
th
e
r
esu
lts
d
if
f
icu
lt
to
in
ter
p
r
et
an
d
co
m
p
licated
th
e
ev
alu
ati
o
n
o
f
th
e
m
o
d
els.
T
h
er
ef
o
r
e,
th
e
p
r
o
b
lem
was
r
ef
o
r
m
u
lated
as
a
b
in
a
r
y
cl
ass
if
icatio
n
task
to
f
ac
ilit
ate
a
m
o
r
e
s
tr
aig
h
tf
o
r
war
d
ev
a
lu
atio
n
.
Usi
n
g
th
e
p
r
o
p
o
s
ed
m
eth
o
d
,
th
e
attem
p
t
was
to
class
if
y
th
e
s
tar
_
r
at
in
g
in
to
‘
1
’
with
4
o
r
5
s
ta
r
_
r
atin
g
s
a
n
d
‘
0
’
,
co
n
s
is
tin
g
o
f
1
o
r
2
s
tar
_
r
atin
g
s
,
in
s
tead
o
f
tr
y
i
n
g
t
o
p
r
ed
ict
th
e
e
x
ac
t
‘
s
tar
_
r
atin
g
’
ass
o
cia
ted
with
th
e
r
ev
iew
body
.
All
r
ev
iews
with
a
s
tar
_
r
atin
g
o
f
3
wer
e
ex
clu
d
ed
s
in
ce
th
ey
r
e
f
lecte
d
n
e
u
tr
al
o
r
in
d
if
f
er
e
n
t
f
ee
d
b
ac
k
.
T
h
e
d
if
f
e
r
en
t
tex
t
p
r
e
-
p
r
o
ce
s
s
in
g
s
tep
s
wer
e
d
o
n
e
with
th
is
u
n
s
tr
u
ctu
r
ed
tex
t
u
al
d
ata
with
NL
P
an
d
tex
t
m
in
in
g
tech
n
iq
u
es b
ef
o
r
e
p
ass
in
g
it th
r
o
u
g
h
th
e
p
i
p
elin
e.
T
h
e
m
o
d
el
was tr
ain
ed
with
th
e
b
atch
d
ata,
wh
ich
is
s
to
r
ed
in
d
is
tr
ib
u
ted
f
ile
s
to
r
ag
e,
s
u
ch
as
a
d
o
cu
m
e
n
t
-
b
as
ed
No
SQL
an
d
th
e
d
ata
was
m
o
v
e
d
th
r
o
u
g
h
th
e
p
ip
elin
e
co
n
tin
u
o
u
s
ly
.
T
h
e
f
o
llo
win
g
s
tep
s
wer
e
ex
ec
u
ted
wh
ile
p
r
e
-
p
r
o
ce
s
s
in
g
th
e
u
n
s
tr
u
ctu
r
ed
tex
t
u
al
r
ev
ie
ws
s
u
ch
as
s
tr
ip
p
in
g
,
r
e
m
o
v
in
g
p
u
n
ctu
at
io
n
an
d
alp
h
a
n
u
m
er
ic
ch
ar
a
cter
s
,
r
em
o
v
in
g
all
s
to
p
w
o
r
d
s
,
tag
g
in
g
u
s
in
g
p
ar
t
-
of
-
s
p
ee
ch
(
POS
)
tag
g
in
g
,
an
d
lem
m
atiza
tio
n
.
L
e
m
m
at
izatio
n
in
v
o
lv
es
th
e
id
en
tific
a
tio
n
o
f
u
n
d
er
ly
i
n
g
lem
m
as
o
f
th
e
w
o
r
d
s
i
n
a
tex
t,
wh
er
ea
s
with
tag
g
in
g
,
a
p
a
r
t
o
f
s
p
ee
ch
was
ass
ig
n
ed
to
ea
ch
wo
r
d
i
n
a
lis
t
u
s
in
g
co
n
tex
t
clu
es.
W
ith
a
v
iew
to
ap
p
ly
in
g
ML
class
if
icatio
n
with
th
e
s
tr
ea
m
in
g
p
i
p
elin
e,
th
e
clea
n
ed
d
ataset
was
s
p
lit
in
to
tr
ain
in
g
an
d
test
in
g
p
o
r
tio
n
s
u
s
in
g
a
7
0
:3
0
r
atio
,
a
n
d
a
r
an
g
e
o
f
m
eth
o
d
s
in
clu
d
in
g
lo
g
is
tic
r
eg
r
ess
io
n
(
L
R
)
,
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
es
(
SVM)
,
an
d
n
aïv
e
B
ay
es
(
NB
)
wer
e
em
p
lo
y
ed
o
n
th
e
d
ataset
to
esti
m
ate
th
e
p
r
o
b
ab
ilit
y
th
at
a
p
r
o
d
u
ct
wo
u
ld
b
e
ac
ce
p
ted
.
T
h
e
m
o
d
el
tr
ain
in
g
was
d
o
n
e
with
th
e
tr
ain
in
g
d
ataset
an
d
test
ed
wi
th
th
e
test
in
g
s
et.
T
h
e
‘
r
ev
ie
w_
b
o
d
y
’
an
d
‘
s
tar
_
r
atin
g
’
f
ie
ld
s
wer
e
s
elec
ted
as
th
e
‘
f
ea
tu
r
e
’
an
d
‘
lab
el’
f
iel
d
s
.
T
h
e
lem
m
atize
d
tex
t
u
al
d
ata
i
n
th
e
tr
ain
i
n
g
s
et
was
p
ass
ed
th
r
o
u
g
h
th
e
p
ip
elin
e
f
o
r
f
itti
n
g
with
th
e
d
if
f
er
en
t
ML
m
o
d
els,
wh
e
r
e
th
e
o
p
er
atio
n
s
o
f
to
k
en
iz
atio
n
,
h
ash
i
n
g
te
r
m
f
r
eq
u
en
cy
(
TF
)
an
d
I
DF
ca
lcu
latio
n
s
[
2
3
]
wer
e
p
e
r
f
o
r
m
ed
o
n
it.
W
ith
to
k
en
izatio
n
a
n
d
h
ash
in
g
,
th
e
d
ata
was c
o
n
v
er
ted
an
d
s
to
r
ed
as a
two
-
d
im
en
s
io
n
al
v
ec
to
r
.
W
ith
to
k
en
izatio
n
,
th
e
i
n
d
iv
id
u
al
r
ev
iew
tex
ts
ar
e
b
r
o
k
e
n
d
o
wn
i
n
to
s
ep
ar
ate
wo
r
d
s
k
n
o
wn
as
to
k
e
n
s
.
Usu
ally
,
a
s
im
p
le
to
k
e
n
izer
class
in
Sp
ar
k
p
r
o
v
id
es
th
is
f
u
n
ctio
n
ality
.
T
h
e
r
eg
u
lar
ex
p
r
ess
io
n
(
R
eg
ex
)
to
k
en
izer
u
s
ed
en
a
b
les
m
o
r
e
s
o
p
h
is
ticated
tex
t
s
p
litt
in
g
b
y
u
s
in
g
R
eg
ex
p
atter
n
s
to
id
en
tify
to
k
en
s
.
I
n
s
tead
o
f
t
h
e
T
F
-
I
DF
tech
n
iq
u
e,
th
e
h
ash
in
g
tr
ick
was
u
s
ed
h
er
e
[
2
4
]
.
Sin
ce
T
F
-
I
DF
ca
lcu
latio
n
s
ten
d
to
g
et
co
m
p
licated
in
a
g
en
er
al
s
ce
n
ar
io
,
it
will
b
e
f
ar
wo
r
s
e
in
a
d
is
tr
ib
u
ted
s
y
s
tem
s
etu
p
.
Hen
ce
,
to
r
ed
u
ce
th
e
m
em
o
r
y
an
d
co
m
p
u
tatio
n
al
o
v
er
h
ea
d
s
,
h
ash
in
g
was
im
p
lem
en
ted
with
Sp
ar
k
.
Hash
in
g
is
a
ty
p
e
o
f
tr
an
s
f
o
r
m
e
r
th
at
c
o
n
v
er
ts
s
ets
o
f
ter
m
s
in
to
f
ix
e
d
-
len
g
th
f
ea
tu
r
e
v
ec
to
r
s
.
I
n
te
x
t
p
r
o
ce
s
s
in
g
,
th
ese
s
ets
o
f
ter
m
s
co
u
ld
r
ep
r
esen
t
a
b
ag
o
f
wo
r
d
s
,
wh
ic
h
in
clu
d
es
all
wo
r
d
s
in
a
tex
t,
b
u
t
h
ash
in
g
ap
p
lies
th
e
h
ash
in
g
tr
ick
t
o
ac
h
iev
e
t
h
is
tr
an
s
f
o
r
m
atio
n
.
A
r
aw
attr
ib
u
te
was
co
n
v
er
ted
in
to
an
in
d
ex
b
y
ap
p
ly
i
n
g
a
h
ash
in
g
p
r
o
ce
d
u
r
e.
T
h
e
h
ash
in
g
m
ec
h
an
is
m
em
p
l
o
y
ed
was
Mu
r
m
u
r
Hash
3
[
1
2
]
,
af
ter
w
h
ich
ter
m
f
r
e
q
u
en
cies
wer
e
ca
lcu
lated
u
s
in
g
th
e
r
esu
ltin
g
in
d
ices.
T
h
is
ap
p
r
o
a
ch
r
em
o
v
es
th
e
r
eq
u
i
r
em
en
t
f
o
r
g
en
e
r
atin
g
a
g
lo
b
al
ter
m
-
to
-
in
d
ex
d
ictio
n
ar
y
wh
ich
ca
n
b
e
c
o
m
p
u
tatio
n
ally
co
s
tly
f
o
r
v
e
r
y
b
ig
d
atasets
.
B
u
t
it
in
tr
o
d
u
ce
s
t
h
e
p
o
s
s
ib
ilit
y
o
f
h
ash
c
o
llis
io
n
s
,
wh
er
e
u
n
r
elate
d
i
n
p
u
t
f
ea
t
u
r
e
s
m
ay
b
e
ass
ig
n
ed
th
e
s
am
e
h
ash
ed
ter
m
an
d
to
m
in
im
ize
co
llis
io
n
r
is
k
s
,
th
e
n
u
m
b
er
o
f
b
u
ck
ets
in
th
e
h
as
h
tab
le
ca
n
b
e
i
n
cr
ea
s
ed
.
B
ec
au
s
e
a
s
im
p
le
m
o
d
u
lo
o
p
er
ati
o
n
was
u
s
ed
to
m
ap
th
e
h
ash
o
u
tp
u
t
to
a
co
l
u
m
n
i
n
d
ex
,
th
e
f
ea
tu
r
e
d
im
en
s
io
n
w
as
ch
o
s
en
as
a
p
o
wer
o
f
two
to
en
s
u
r
e
a
u
n
if
o
r
m
d
is
tr
ib
u
tio
n
ac
r
o
s
s
co
lu
m
n
s
.
T
h
e
d
e
f
au
lt
d
im
en
s
io
n
ality
is
s
et
to
2
18
=
2
6
2
,
1
4
4
,
an
d
a
co
n
f
ig
u
r
ab
le
b
in
ar
y
p
ar
am
eter
d
eter
m
in
es
wh
et
h
e
r
ter
m
f
r
eq
u
en
cies
ar
e
co
u
n
te
d
in
b
in
ar
y
f
o
r
m
.
No
r
m
all
y
,
o
f
a
ter
m
‘
t’
is
ca
lcu
lated
u
s
in
g
(
1
).
(
)
=
(
)
(
1
)
W
h
er
e
is
th
e
to
tal
n
u
m
b
er
o
f
ter
m
s
an
d
(
)
is
th
e
f
r
eq
u
en
cy
o
f
‘
’
in
th
e
d
o
cu
m
en
t
[
2
5
]
.
B
u
t
wh
en
a
wo
r
d
th
at
was
n
o
t
in
th
e
v
o
ca
b
u
lar
y
o
cc
u
r
s
,
th
e
co
r
r
esp
o
n
d
i
n
g
(
)
will
b
e
ze
r
o
,
w
h
ich
was
o
b
tain
ed
f
r
o
m
th
e
h
ash
in
g
tr
ick
.
As
d
iv
is
io
n
b
y
ze
r
o
was
n
o
t
p
o
s
s
ib
le,
th
e
v
alu
e
was
s
m
o
o
th
ed
b
y
ad
d
in
g
1
to
th
e
d
en
o
m
in
at
o
r
.
Hen
ce
,
(
2
)
was u
s
ed
to
ca
lcu
late
th
e
I
DF
,
wh
e
r
e
is
th
e
to
tal
n
u
m
b
er
o
f
ter
m
s
.
(
)
=
l
og
(
(
)
+
1
)
(
2
)
T
h
en
,
m
o
d
el
tr
ain
in
g
was
ca
r
r
ied
o
u
t
u
s
in
g
class
if
icatio
n
alg
o
r
ith
m
s
s
u
ch
as
LR
,
SVM
,
an
d
NB
,
wh
ich
ar
e
av
ailab
le
with
in
th
e
Sp
ar
k
ML
lib
r
ar
y
.
T
h
e
o
p
e
r
atio
n
is
a
n
esti
m
ato
r
an
d
an
ad
d
itio
n
al
c
o
lu
m
n
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
5
,
Octo
b
er
2
0
2
6
:
4
0
7
1
-
4
0
8
0
4074
‘
p
r
ed
ictio
n
’
was
g
ettin
g
ad
d
e
d
to
it.
T
h
en
,
th
e
p
ip
elin
ed
ML
m
o
d
el
is
test
ed
with
th
e
test
d
ata.
W
h
en
e
v
er
a
n
esti
m
ato
r
f
u
n
ctio
n
s
u
ch
as
a
ML
alg
o
r
ith
m
was
tr
ig
g
er
ed
in
a
p
ip
elin
e,
it’s
p
r
ec
ed
in
g
tr
an
s
f
o
r
m
atio
n
f
u
n
ctio
n
s
wh
ich
wer
e
p
e
n
d
in
g
to
b
e
ex
ec
u
ted
a
r
e
en
ab
led
,
b
ef
o
r
e
it
esti
m
ated
th
e
p
r
e
d
ictiv
e
lab
els.
I
t
is
tr
an
s
f
o
r
m
in
g
o
u
r
r
esu
ltan
t
d
at
a
f
r
am
e
with
an
ad
d
itio
n
al
f
ea
tu
r
e
ca
lled
‘
p
r
ed
ictio
n
’
,
ap
ar
t
f
r
o
m
a
lab
el
wh
ich
r
ep
r
esen
ts
th
e
ex
p
ec
ted
p
r
ed
i
ctio
n
s
.
T
h
e
ac
cu
r
ac
y
o
f
th
ese
p
r
ed
ictio
n
s
m
ad
e
b
y
th
e
d
if
f
er
en
t
ML
m
o
d
els
u
s
in
g
th
e
ML
p
ip
elin
e
was
e
v
alu
ated
u
s
in
g
th
e
m
etr
ics:
th
e
R
OC
-
ar
ea
u
n
d
er
t
h
e
c
u
r
v
e
(
A
UC
)
cu
r
v
es
an
d
th
e
G
i
n
i
c
o
e
f
f
i
ci
e
n
t
.
T
h
e
b
es
t
-
p
e
r
f
o
r
m
i
n
g
m
o
d
e
l
w
a
s
s
e
l
e
ct
e
d
a
s
c
h
a
m
p
i
o
n
m
o
d
e
l
a
n
d
i
t
i
s
d
e
p
l
o
y
e
d
f
o
r
p
r
o
d
u
c
t
i
o
n
.
2
.
2
.
P
r
o
ce
s
s
f
lo
w
dia
g
ra
m
a
nd
a
lg
o
rit
hm
F
ig
u
r
e
1
illu
s
tr
ates
th
e
o
v
er
all
f
lo
wg
r
ap
h
o
f
th
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
,
o
u
tlin
in
g
th
e
s
eq
u
en
tial
p
r
o
ce
s
s
es
an
d
in
ter
ac
tio
n
s
a
m
o
n
g
its
k
ey
co
m
p
o
n
en
ts
.
T
h
e
s
tr
ea
m
in
g
d
ata
is
p
r
ep
ar
ed
f
ir
s
t
,
an
d
th
e
p
r
e
-
p
r
o
ce
s
s
in
g
u
p
to
lem
m
atiza
tio
n
is
ap
p
lied
with
th
e
tex
tu
al
r
ev
iews.
T
h
en
th
is
d
ata
i
s
p
ass
ed
th
r
o
u
g
h
a
s
tr
ea
m
in
g
p
ip
elin
e,
wh
ich
co
n
s
is
ts
o
f
th
e
s
tag
es,
to
k
en
izati
o
n
,
h
ash
in
g
,
I
DF
ca
lcu
latio
n
,
an
d
f
in
ally
th
e
ML
m
o
d
ellin
g
.
T
h
e
p
r
o
p
o
s
ed
ap
p
r
o
ac
h
f
o
r
m
a
k
in
g
t
h
e
p
r
ed
ictiv
e
an
aly
s
is
with
th
e
s
tr
ea
m
in
g
p
ip
elin
e
is
d
etailed
in
th
e
A
lg
o
r
ith
m
1
p
r
ed
ictiv
e
an
aly
tics
p
ip
elin
e.
Alg
o
r
ith
m
1
.
Pre
d
ictiv
e
an
aly
t
ics
p
ip
elin
e
Input: The Data frame ‘data’ consisting of unstructured review_body and star_rating
Output: The deployable machine learning model for intelligent prediction
//Preparing the dataset for modelling by removing neutral comments
1:
Load the data frame ‘data’ from the distributed storage facility.
2:
For each row ‘i’ in the data frame ‘data’
3:
if (i[star_rating] == ’3’) filter it
4:
el
se
if
((
i[
st
ar
_r
at
in
g]
==
’1
’)
||
(i
[s
ta
r_
ra
ti
ng
]
==
’2
’)
),
i[
y]
=0
,
wh
er
e
y
∈
Target
5:
else i[y] = 1, where y
∈
Y /* Y is a new column ‘Target’*/
6:
End of if
7:
End of for
//Pre
-
processing
the
entire
set
of
‘review_body’
i
n
data,
which
is
the
unst
ructured
textual
data
8:
Fo
r
ea
ch
‘r
ev
ie
w
_
bo
dy
’,
‘
X’
in
da
ta
,
pe
rf
or
m
th
e
f
ol
lo
wi
ng
tr
an
sf
or
ma
ti
on
s
fr
om
st
ep
9
to step 16 as pre
-
processing, and with each step a new column will be added to ‘data’
9:
lower_text = lower(X)
10:
text__non_asci = strip_non_ascii(lower_text)
11:
fixed_abbrev = fix_abbreviation(text_non_asci)
12:
removed_features = remove_puctuations_special symbols(fixed_abbrev)
13:
removed_stops = remove_stops(removed_features)
14:
tag_and_remove_pos = remove_tags(removed_stops)
15:
lemmatized_text = lemmatize(tag_and_remove_pos)
16:
End of for
17:
Construct the final data frame ‘data’ with the result of previous step 15 as
18:
data = data.select(lemmatized_text, date, Target) and
19:
set ‘text’ = lemmatized_text and ‘label’ = Target
20:
Di
vi
de
th
e
da
ta
fo
r
tr
ai
ni
ng
an
d
te
st
in
g
st
ag
es
as
70
%
da
ta
fo
r
tr
ai
ni
ng
an
d
30
%
fo
r
test as train_data=random_select(data,70) and test_data=random_select(data,30)
21:
Tr
ai
n
an
d
fi
t
th
e
mo
de
l
by
pa
ss
in
g
th
e
tr
ai
n
da
ta
th
ro
ug
h
th
e
st
re
am
in
g
pi
pe
li
ne
as
in
step 22 with the following stages, where ML is the machine learning algorithm used.
22:
Streaming_pipeline = Pipeline(stages = [tokenizer, hashing, IDF, ML])
23:
For each stage in the pipeline, perform the following Steps 25 to Step 37
24:
switch(stage)
25:
Case ‘tokenizer’:
26:
tokenize(‘text’) = {t: t
∈
T, where T is the text in the dataset}
27:
Case ‘hashing’:
//Deciding the term frequency by applying hashing trick to each term in the {t}
28:
For each token {t: t
∈
T}
29:
hashing(t)
=
{index(t,
V),
where
V
is
the
fixed
length
feature
vector}
30:
End for
//Hashing
is
a
transformer
which
takes
sets
of
terms
and
converts
them
into
fixed
-
le
ng
th
feature vectors
31:
Case ‘IDF’:
(
)
=
lo
g
(
/
(
ℎ
ℎ
(
)
+
1
)
)
//where n → total no of terms in V
32:
ℎ
ℎ
_
(
,
)
=
ℎ
ℎ
(
)
∗
lo
g
(
/
(
ℎ
ℎ
(
)
+
1
)
)
//whatever we are processing now is T
33:
Case ‘ML’:
34:
Ca
rr
y
ou
t
mo
de
l
co
ns
t
ru
ct
io
n
us
in
g
th
e
fo
l
lo
wi
ng
su
pe
rv
is
ed
ma
ch
i
ne
le
ar
ni
ng
classification
techniques
,
ML
=
{L
R,
SV
M,
NB
},
wh
er
e
LR
=
lo
gi
st
ic
re
gr
es
si
on
,
SV
M
=
support vector machines, and NB
=
naïve Bayes
35:
End switch
36:
End For
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
A
n
in
tel
lig
en
t A
I
fr
a
mewo
r
k
fo
r
b
u
s
in
ess
in
s
ig
h
t g
en
era
tio
n
f
r
o
m
s
tr
ea
min
g
…
(
B
en
ymo
l
Jo
s
e
)
4075
37:
Te
st
th
e
mo
de
ls
us
in
g
th
e
te
st
_d
at
a
an
d
ev
al
ua
te
th
e
re
su
lt
s
wi
th
AU
C
-
RO
C
an
d
Gi
ni
Coefficient.
38:
Select the champion model which is the deployable model that can work with stream data
39:
End
Fig
u
r
e
1
.
Flo
w
d
ia
g
r
am
u
s
ed
with
th
e
d
ata
p
r
o
ce
s
s
in
g
p
latf
o
r
m
3.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
3
.
1
.
Da
t
a
s
et
us
ed
a
nd
ex
perim
ent
a
l set
up
T
h
e
d
ataset
u
tili
ze
d
in
th
is
i
s
th
e
Am
az
o
n
r
ev
iews
f
o
r
s
p
o
r
ts
an
d
o
u
t
d
o
o
r
s
,
d
o
wn
l
o
a
d
ed
f
r
o
m
Kag
g
le.
I
t
co
n
tain
s
4
,
8
4
9
,
9
4
5
r
ec
o
r
d
s
,
a
p
p
r
o
x
im
ately
2
GB
in
s
ize,
an
d
in
clu
d
es
d
etail
s
o
f
s
p
o
r
ts
p
r
o
d
u
cts
alo
n
g
with
cu
s
to
m
er
f
ee
d
b
ac
k
an
d
s
u
m
m
ar
y
in
f
o
r
m
atio
n
.
T
h
e
an
aly
s
is
f
o
cu
s
es
o
n
th
e
u
n
s
tr
u
ctu
r
ed
tex
tu
al
r
ev
iews
r
ev
iew_
b
o
d
y
p
r
o
v
id
e
d
b
y
cu
s
to
m
e
r
s
,
alo
n
g
with
th
e
co
r
r
esp
o
n
d
in
g
s
tar
r
atin
g
f
o
r
v
ar
io
u
s
s
p
o
r
ts
p
r
o
d
u
cts
[
2
2
]
.
T
h
e
ex
p
e
r
im
en
t
s
wer
e
d
o
n
e
u
s
in
g
th
e
Data
b
r
i
ck
s
co
m
m
u
n
ity
e
d
itio
n
with
in
th
e
Ap
ac
h
e
S
p
ar
k
ec
o
s
y
s
tem
,
co
n
f
ig
u
r
e
d
with
6
GB
o
f
R
AM
,
0
.
8
8
co
r
es,
Data
b
r
ick
s
r
u
n
tim
e
(
DB
R
)
6
.
4
with
Sp
ar
k
2
.
4
.
5
,
an
d
Scala
2
.
1
1
.
Ad
d
itio
n
ally
,
Go
o
g
le
C
o
lab
,
a
f
r
ee
clo
u
d
p
latf
o
r
m
b
y
Go
o
g
le,
was
u
s
ed
to
en
co
u
r
a
g
e
r
esear
ch
lev
er
ag
in
g
th
e
Go
o
g
le
C
lo
u
d
Platfo
r
m
(
GC
P)
f
o
r
co
m
p
u
t
atio
n
s
an
d
ev
alu
atio
n
s
,
o
f
f
er
i
n
g
b
en
e
f
its
s
u
ch
as
in
teg
r
atio
n
with
r
ep
o
r
tin
g
an
d
ex
tr
ac
t,
tr
an
s
f
o
r
m
,
an
d
lo
ad
(
ETL
)
to
o
ls
.
T
h
e
Sp
ar
k
p
ip
elin
in
g
im
p
lem
en
tatio
n
s
wer
e
ex
ec
u
te
d
u
s
in
g
Py
Sp
ar
k
o
n
a
GC
P
co
m
p
u
te
en
g
in
e
b
ac
k
en
d
,
with
1
2
.
7
2
GB
R
AM
an
d
1
0
7
.
7
7
GB
o
f
d
is
k
s
p
ac
e.
3
.
2
.
C
o
m
p
a
r
a
t
iv
e
per
f
o
r
ma
nce
e
v
a
lu
a
t
i
o
n
o
f
a
r
t
i
f
ic
ia
l in
t
el
li
g
en
ce
m
o
d
els
f
o
r
c
ha
m
p
i
o
n
mo
d
el
s
el
ec
t
i
o
n
T
o
ac
tiv
ate
an
d
p
er
f
o
r
m
th
e
in
tellig
en
t
p
r
o
ce
s
s
in
g
o
f
th
e
u
n
s
tr
u
ctu
r
ed
r
e
v
iews,
s
to
r
ed
in
th
e
d
is
tr
ib
u
ted
f
ile
s
y
s
tem
,
th
e
d
a
ta
is
p
ass
ed
th
r
o
u
g
h
a
s
tr
ea
m
in
g
p
ip
elin
e
.
Sp
ar
k
s
tr
ea
m
in
g
an
d
Sp
ar
k
ML
ar
e
u
s
ed
to
g
eth
er
to
p
er
f
o
r
m
th
e
p
r
ed
ictiv
e
a
n
aly
tics
,
an
d
in
th
e
f
i
n
al
s
tag
e
o
f
th
e
p
i
p
elin
e,
d
if
f
er
e
n
t
ML
class
if
icatio
n
m
o
d
els
lik
e
LR
,
SVM,
an
d
NB
,
ar
e
im
p
lem
en
ted
.
T
h
ese
th
r
ee
alg
o
r
ith
m
s
r
ep
r
esen
t
v
ar
io
u
s
asp
ec
ts
o
f
ML
s
u
ch
as
r
eg
r
ess
io
n
an
aly
s
is
,
wh
ich
is
b
ased
o
n
s
tatis
t
ical
m
o
d
ellin
g
,
wh
er
ea
s
SVM
p
lay
s
with
th
e
id
ea
o
f
s
tate
-
b
ased
an
alo
g
y
,
with
a
u
n
iq
u
e
way
o
f
ev
al
u
atio
n
[
2
6
]
.
T
h
e
B
ay
esian
th
e
o
r
em
b
ac
k
s
th
e
NB
alg
o
r
ith
m
th
at
wo
r
k
s
with
p
r
o
b
ab
ilis
tic
an
aly
s
is
[
2
7
]
.
T
h
e
tr
ain
in
g
d
ata
is
p
ass
ed
th
r
o
u
g
h
th
e
p
ip
elin
e
s
tag
es,
s
u
ch
as
to
k
en
izatio
n
,
h
ash
in
g
,
I
DF
,
an
d
f
in
ally
th
e
d
if
f
er
en
t
ML
alg
o
r
ith
m
s
ar
e
ap
p
lied
,
wh
er
e
th
e
p
ip
elin
es
p
er
f
o
r
m
d
i
f
f
er
en
tly
with
ea
ch
o
f
t
h
e
alg
o
r
ith
m
s
u
s
ed
.
T
h
e
p
e
r
f
o
r
m
an
ce
o
f
th
e
p
i
p
elin
ed
m
o
d
el
in
m
ak
in
g
p
r
e
d
ictio
n
s
with
th
e
test
d
ata
is
m
ea
s
u
r
ed
b
y
co
n
s
id
er
i
n
g
th
e
m
etr
ics
AUC
-
R
OC
an
d
th
e
Gin
i
co
ef
f
icien
t.
T
h
e
R
OC
g
r
ap
h
is
a
s
tan
d
ar
d
way
o
f
an
aly
zin
g
th
e
p
er
f
o
r
m
an
ce
o
f
ML
alg
o
r
ith
m
s
b
ased
o
n
its
ab
ilit
y
to
d
o
p
r
ed
ictio
n
s
.
Gin
i
co
ef
f
icien
t
is
an
o
th
er
p
er
f
o
r
m
a
n
ce
m
etr
ic
u
s
ed
in
class
if
icatio
n
p
r
o
b
lem
s
.
Gin
i
in
d
ex
ca
n
b
e
d
er
iv
ed
s
tr
a
ig
h
t
awa
y
f
r
o
m
th
e
AUC
-
R
O
C
,
an
d
it
is
a
n
in
d
icato
r
o
f
h
o
w
well
th
e
m
o
d
el
is
p
er
f
o
r
m
in
g
.
T
h
e
Gin
i
in
d
ex
f
o
r
th
e
p
r
o
p
o
s
ed
m
o
d
els
ca
n
b
e
ca
lcu
lated
as
in
(
3
)
.
T
h
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
p
ip
elin
e
with
th
e
f
o
llo
win
g
ML
alg
o
r
ith
m
s
is
ev
alu
ated
.
(
)
=
(
2
×
)
−
1
(
3
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
5
,
Octo
b
er
2
0
2
6
:
4
0
7
1
-
4
0
8
0
4076
3
.
2
.
1
.
L
o
g
is
t
ic
re
g
re
s
s
io
n m
o
del
I
t
r
ep
r
esen
ts
a
s
p
ec
ialized
f
o
r
m
o
f
g
e
n
er
alize
d
lin
ea
r
m
o
d
e
l
in
ten
d
ed
to
c
o
m
p
u
te
th
e
lik
elih
o
o
d
o
f
p
o
s
s
ib
le
o
u
tco
m
es.
W
ith
in
Sp
ar
k
ML
,
LR
p
r
o
v
id
es
s
u
p
p
o
r
t
f
o
r
b
o
t
h
two
-
class
p
r
ed
ictio
n
u
s
in
g
b
in
o
m
ial
LR
an
d
m
u
lti
-
ca
teg
o
r
y
p
r
ed
ictio
n
u
s
in
g
m
u
ltin
o
m
ial
L
R
[
2
8
]
.
Her
e,
th
is
ML
m
o
d
el
is
u
s
ed
f
o
r
m
ak
in
g
b
i
n
ar
y
class
p
r
ed
ictio
n
s
r
eg
a
r
d
in
g
th
e
ac
ce
p
tan
ce
o
f
a
p
r
o
d
u
ct,
b
ased
o
n
th
e
r
ev
iews.
W
ith
a
p
r
e
d
i
ctio
n
v
alu
e
‘
1
’
,
th
e
p
r
o
d
u
ct
is
ass
u
m
ed
to
b
e
ac
ce
p
tab
le
,
an
d
p
r
ed
ictio
n
v
alu
e
‘
0
’
d
en
o
tes its
u
n
ac
ce
p
ta
b
ilit
y
.
−
AUC
-
R
O
C
:
wh
en
th
is
s
tatis
t
ic
al
m
o
d
el
was
ap
p
lied
u
s
in
g
th
at
wo
r
k
f
lo
w,
th
e
r
esu
lts
s
h
o
wed
an
AUC
o
f
6
1
%.
T
h
e
R
OC
cu
r
v
e
o
f
L
R
p
ip
elin
e
is
s
h
o
wn
in
F
ig
u
r
e
2
.
T
h
e
R
OC
-
AUC
f
o
r
th
e
LR
m
o
d
el
s
h
o
ws
6
1
%
o
f
ac
cu
r
ac
y
,
w
h
ich
in
d
ic
ates
its
co
n
f
id
en
ce
in
p
r
ed
ictin
g
lab
els
ac
cu
r
ately
with
a
n
e
r
r
o
r
c
h
an
ce
o
f
3
9
%.
I
t
in
d
icate
s
th
at
wh
en
t
h
e
p
r
ed
ictio
n
s
ar
e
m
ad
e
u
s
in
g
th
is
ML
p
ip
elin
e,
6
1
%
o
f
p
r
ed
ictio
n
s
ar
e
m
ad
e
co
r
r
ec
tly
an
d
3
9
% o
f
p
r
ed
ictio
n
s
wer
e
wr
o
n
g
.
−
Gin
i
co
ef
f
icien
t
:
ev
alu
atin
g
t
h
e
L
R
m
o
d
el
with
th
e
Gin
i
co
e
f
f
icien
t
g
av
e
a
s
co
r
e
o
f
0
.
2
2
,
wh
ich
im
p
lies
th
at
th
e
m
o
d
el
is
n
o
t
s
u
itab
l
e
f
o
r
p
r
o
d
u
ctio
n
.
Sin
ce
LR
wo
r
k
s
b
ased
o
n
lik
elih
o
o
d
a
n
d
is
p
r
o
n
e
to
o
u
tlier
s
,
it is
co
n
v
en
ien
tly
elim
in
ated
in
th
e
c
h
am
p
io
n
m
o
d
el
s
elec
tio
n
.
Fig
u
r
e
2
.
T
h
e
AUC
-
R
OC
o
f
L
R
3
.
2
.
2
.
Na
ïv
e
B
a
y
es
m
o
del
NB
cla
s
s
if
ier
s
ar
e
a
g
r
o
u
p
o
f
s
tr
aig
h
tf
o
r
war
d
p
r
o
b
ab
ilis
tic
m
u
lticlas
s
m
o
d
els
th
at
r
ely
o
n
B
ay
es’
th
eo
r
em
,
ass
u
m
in
g
th
at
all
f
ea
tu
r
es
ar
e
co
n
d
itio
n
ally
in
d
ep
e
n
d
en
t
o
f
o
n
e
an
o
t
h
er
.
W
ith
ju
s
t
o
n
e
p
ass
th
r
o
u
g
h
th
e
tr
ain
in
g
d
ataset,
th
e
alg
o
r
i
th
m
ca
lcu
lates
th
e
co
n
d
itio
n
al
p
r
o
b
ab
ilit
y
d
is
tr
ib
u
tio
n
o
f
ea
ch
f
ea
tu
r
e
f
o
r
ev
er
y
lab
el.
Du
r
in
g
p
r
ed
ictio
n
,
it
u
s
es
B
ay
es’
th
eo
r
em
to
d
eter
m
i
n
e
th
e
co
n
d
itio
n
al
p
r
o
b
a
b
ilit
y
d
is
tr
ib
u
tio
n
o
f
ea
c
h
lab
el
g
iv
en
a
n
ew
o
b
s
er
v
atio
n
[
2
8
].
−
AUC
-
R
O
C
:
wit
h
NB
clas
s
if
ie
r
s
th
e
s
tr
ea
m
in
g
p
ip
elin
in
g
g
a
v
e
an
AUC
v
alu
e
o
f
8
1
%.
A
UC
v
alu
e
0
.
8
1
m
ea
n
s
th
at
8
1
%
o
f
th
e
p
r
ed
ic
tio
n
s
ar
e
co
r
r
ec
t
an
d
th
at
o
n
l
y
1
9
%
o
f
p
r
ed
ictio
n
s
ar
e
wr
o
n
g
.
T
h
e
R
OC
cu
r
v
e
o
f
th
e
p
ip
elin
e
is
s
h
o
wn
in
Fig
u
r
e
3
.
−
Gin
i
co
ef
f
icien
t
:
NB
alg
o
r
ith
m
g
av
e
a
v
alu
e
o
f
0
.
6
2
as
its
Gin
i
co
ef
f
icien
t,
wh
ich
is
a
p
r
o
m
is
in
g
o
u
tco
m
e,
s
in
ce
th
e
s
tr
ateg
y
o
f
Gin
i
co
ef
f
icien
t
tech
n
iq
u
e
d
e
m
an
d
s
th
e
p
er
f
o
r
m
an
ce
o
f
th
e
m
o
d
el
m
u
s
t
b
e
g
r
ea
ter
t
h
an
0
.
6
0
[
2
9
]
,
[
3
0
]
.
Fig
u
r
e
3
.
T
h
e
AUC
-
R
OC
o
f
NB
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
A
n
in
tel
lig
en
t A
I
fr
a
mewo
r
k
fo
r
b
u
s
in
ess
in
s
ig
h
t g
en
era
tio
n
f
r
o
m
s
tr
ea
min
g
…
(
B
en
ymo
l
Jo
s
e
)
4077
3
.
2
.
3
.
Su
pp
o
rt
v
ec
t
o
r
m
a
chi
ne
m
o
del
An
SVM
cr
ea
tes
a
h
y
p
er
p
lan
e,
o
r
m
u
ltip
le
h
y
p
er
p
lan
e
s
,
with
in
a
h
ig
h
-
d
im
en
s
io
n
a
l
o
r
ev
en
in
f
in
ite
-
d
im
en
s
io
n
al
s
p
ac
e,
wh
ich
ca
n
b
e
u
s
ef
u
l
f
o
r
class
if
icatio
n
,
r
e
g
r
ess
io
n
,
a
n
d
o
th
e
r
t
ask
s
.
I
n
ess
en
ce
,
an
ef
f
ec
tiv
e
s
ep
ar
atio
n
is
o
b
tain
ed
b
y
s
elec
tin
g
th
e
h
y
p
e
r
p
la
n
e
th
at
m
ax
im
izes
th
e
d
is
ta
n
ce
to
th
e
clo
s
est
tr
ain
in
g
s
am
p
les
f
r
o
m
ea
c
h
cl
ass
.
Gen
er
ally
,
a
lar
g
er
m
a
r
g
i
n
r
esu
lts
in
lo
wer
g
en
e
r
aliza
tio
n
e
r
r
o
r
f
o
r
a
lin
ea
r
SVM
class
if
ier
[
3
1
]
.
−
AUC
-
R
O
C
:
th
e
R
OC
cu
r
v
e
o
f
th
e
SVM
p
ip
elin
e
is
s
h
o
wn
in
Fig
u
r
e
4
.
T
h
e
p
ip
elin
e
with
th
e
SVM
m
o
d
el
g
av
e
an
AUC
v
alu
e
o
f
7
2
%.
T
h
e
m
o
d
el’
s
p
r
o
b
ab
il
is
tic
m
etr
ics
in
d
icate
an
ac
cu
r
ac
y
o
f
7
2
%,
r
ef
lectin
g
th
e
SVM
m
o
d
el’
s
c
ap
ab
ilit
y
to
co
r
r
ec
tly
class
if
y
m
o
s
t
lab
els,
wh
ile
leav
in
g
a
r
em
ain
in
g
2
8
%
m
ar
g
in
f
o
r
p
o
ten
tial m
is
class
if
icatio
n
.
−
Gin
i
co
ef
f
icien
t
:
SVM
s
co
r
es
0
.
4
4
f
o
r
Gin
i
co
ef
f
icien
t,
wh
i
ch
im
p
lies
th
at
th
e
SVM
m
o
d
el
is
n
o
t
u
p
to
th
e
n
o
tch
f
o
r
p
r
o
d
u
ctio
n
i
n
th
i
s
co
n
tex
t sin
ce
it d
o
es n
o
t
b
r
ea
ch
th
e
th
r
esh
o
ld
o
f
0
.
6
0
.
Fig
u
r
e
4
.
T
h
e
AUC
-
R
OC
o
f
S
VM
3
.
2
.
4
.
Cha
m
pio
n
mo
del selec
t
io
n
T
h
e
R
OC
s
o
f
th
e
th
r
ee
alg
o
r
ith
m
s
u
s
ed
f
o
r
m
ak
in
g
in
t
ellig
en
t
p
r
ed
ictio
n
s
with
th
e
p
ip
elin
e
d
m
o
d
ellin
g
ar
e
co
m
b
i
n
ed
an
d
ar
e
s
h
o
wn
in
Fig
u
r
e
5
.
Fro
m
Fig
u
r
e
5
,
it
is
ev
id
en
t
th
at
t
h
e
NB
ML
m
o
d
el
p
er
f
o
r
m
s
b
etter
in
t
h
e
Sp
ar
k
ML
an
d
p
ip
elin
in
g
e
n
v
ir
o
n
m
en
t
f
o
r
u
n
s
tr
u
ctu
r
ed
tex
t
u
al
d
ata
an
aly
s
is
with
an
AUC
o
f
8
1
%.
i.e
.
,
wh
en
th
e
e
n
tire
test
d
ata
i
s
ev
alu
ated
,
th
is
s
tu
d
y
g
o
t
a
h
ar
m
o
n
i
z
ed
p
r
ed
ictio
n
an
d
lab
el
f
o
r
8
1
%
o
f
th
e
d
ata
,
w
h
er
ea
s
it
co
n
tr
ad
icted
th
e
la
b
el
with
p
r
e
d
ictio
n
f
o
r
t
h
e
r
em
ain
in
g
1
9
%
.
T
h
e
SVM
an
d
LR
m
o
d
els
also
p
er
f
o
r
m
with
a
c
o
m
p
etin
g
AUC
o
f
7
2
%
an
d
6
1
%
wh
ich
d
o
es
n
o
t
im
p
ly
th
a
t
th
ese
m
o
d
els
ar
e
in
co
m
p
eten
t,
in
s
tead
,
th
ese
al
g
o
r
ith
m
s
ar
e
also
ad
ap
tab
le
s
in
ce
th
e
m
o
d
els
s
h
o
w
AUC
o
v
er
th
e
th
r
esh
o
ld
v
alu
e,
wh
ic
h
is
5
0
%.
W
ith
th
i
s
an
aly
s
is
,
th
is
s
tu
d
y
s
elec
ted
th
e
NB
m
o
d
el
as
th
e
ch
am
p
io
n
,
wh
ich
is
u
s
ed
f
o
r
d
ep
lo
y
m
e
n
t
p
u
r
p
o
s
es
.
Fig
u
r
e
5
.
C
o
m
p
a
r
is
o
n
o
f
R
OC
s
o
f
d
if
f
er
e
n
t M
L
m
o
d
els
Mo
d
els
wh
ich
ar
e
h
a
v
in
g
t
h
e
Gin
i
co
ef
f
icien
t
a
b
o
v
e
6
0
%
ar
e
co
n
s
id
er
ed
as
g
o
o
d
m
o
d
els
[
2
9
]
,
[
3
0
]
wh
ich
ca
n
b
e
ta
k
en
as
t
h
e
ch
a
m
p
io
n
m
o
d
el
an
d
b
e
d
ep
lo
y
ed
.
T
h
e
f
o
llo
win
g
g
r
a
p
h
in
Fig
u
r
e
6
d
em
o
n
s
tr
ates
a
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
5
,
Octo
b
er
2
0
2
6
:
4
0
7
1
-
4
0
8
0
4078
co
m
p
ar
is
o
n
o
f
Gin
i
co
ef
f
icien
t
o
b
tain
ed
f
o
r
d
i
f
f
er
en
t
ML
alg
o
r
ith
m
s
.
Fo
r
NB
m
o
d
el,
th
is
s
tu
d
y
g
o
t
Gin
i
co
ef
f
icien
t a
s
6
2
%,
wh
ich
is
ab
o
v
e
th
e
m
a
r
g
in
o
f
6
0
% a
n
d
h
en
ce
it c
an
b
e
co
n
s
id
er
e
d
as th
e
ch
am
p
io
n
m
o
d
el
.
Acc
o
r
d
in
g
to
C
h
en
et
a
l.
[
3
2
]
,
th
e
two
k
ey
o
b
s
tacle
s
in
tap
p
in
g
b
ig
d
ata
ar
e
its
u
n
s
tr
u
ctu
r
ed
n
at
u
r
e
a
n
d
t
h
e
v
o
lu
m
e
o
f
th
e
d
ata
t
o
b
e
an
al
y
ze
d
.
T
h
is
p
a
p
er
p
r
o
p
o
s
es
th
a
t
NL
P
tech
n
iq
u
es
ca
n
b
e
ap
p
l
ied
with
in
b
ig
d
ata
p
latf
o
r
m
s
s
u
ch
as
Sp
ar
k
to
a
n
aly
ze
u
n
s
tr
u
ctu
r
ed
s
tr
ea
m
in
g
d
ata
a
n
d
u
n
co
v
er
v
alu
ab
le
p
atter
n
s
th
at
s
u
p
p
o
r
t
b
u
s
in
ess
in
tellig
en
ce
.
Acc
o
r
d
in
g
to
Kam
al
et
a
l.
[
1
1
]
,
th
e
h
u
g
e
d
ata
ca
n
b
e
p
r
o
ce
s
s
ed
in
r
ea
l
tim
e
b
y
u
s
in
g
Sp
ar
k
,
an
d
th
ey
h
av
e
u
s
ed
f
ew
s
tr
u
ctu
r
ed
q
u
e
r
y
lan
g
u
ag
e
(
SQL
)
q
u
er
ies
t
o
m
in
e
i
n
f
o
r
m
atio
n
f
r
o
m
th
e
b
ig
d
ata.
Her
e,
te
x
t
m
in
in
g
an
d
NL
P
ar
e
im
p
lem
e
n
ted
to
ap
p
r
eh
en
d
th
e
u
n
s
tr
u
ct
u
r
ed
tex
tu
al
d
ata
o
f
s
tr
ea
m
in
g
n
atu
r
e.
T
h
e
d
ata
o
f
s
tr
ea
m
in
g
n
atu
r
e
is
p
r
o
ce
s
s
ed
an
d
th
en
th
e
cu
r
r
en
tly
d
is
co
v
e
r
ed
d
ata
ar
e
ap
p
e
n
d
ed
t
o
it.
Fig
u
r
e
6
.
Gin
i c
o
ef
f
icien
t
v
al
u
e
o
f
d
if
f
er
en
t M
L
m
o
d
els
4.
CO
NCLU
SI
O
N
I
n
th
is
p
ap
er
,
t
o
tak
e
ad
v
a
n
t
ag
e
o
f
in
tellig
en
t
p
r
o
ce
s
s
in
g
o
n
u
n
s
tr
u
ct
u
r
ed
tex
t
u
al
d
ata
with
th
e
v
elo
city
f
ac
to
r
,
a
m
o
d
el
is
d
ev
elo
p
ed
u
s
in
g
Sp
ar
k
ML
p
ip
elin
in
g
co
n
ce
p
ts
.
T
h
e
p
r
o
p
o
s
ed
m
o
d
el
h
as
th
e
ca
p
ab
ilit
ies
to
p
r
o
ce
s
s
u
n
s
tr
u
ctu
r
ed
d
ata
with
p
r
o
ce
d
u
r
al
p
r
o
ce
s
s
in
g
f
lex
i
b
ilit
y
f
o
r
im
p
lem
en
tin
g
ar
tific
ial
in
tellig
en
ce
.
T
h
e
d
ata
,
wh
ich
is
s
p
r
ea
d
g
lo
b
ally
o
v
er
v
ar
io
u
s
clu
s
ter
s
,
ca
n
b
e
m
in
e
d
in
a
r
ea
l
-
tim
e
f
ash
io
n
,
s
in
ce
it
is
p
o
s
s
ib
le
th
at
th
e
d
at
a
ca
n
b
e
f
r
o
m
a
liv
e
s
o
u
r
ce
o
r
f
r
o
m
a
s
tr
ea
m
ab
le
s
o
u
r
ce
.
I
t
h
elp
s
to
im
p
lem
e
n
t
p
r
ed
ictiv
e
an
al
y
tics
with
p
r
o
c
ed
u
r
al
-
lev
el
p
r
o
g
r
am
m
in
g
f
le
x
ib
ilit
y
o
v
er
s
tr
ea
m
d
ata
with
o
u
t
an
y
ad
d
itio
n
al
en
g
in
ee
r
in
g
u
p
g
r
ad
es.
T
h
e
a
u
th
o
r
s
s
u
g
g
est
th
at
b
u
s
in
ess
in
tellig
en
ce
ca
n
b
e
en
h
an
ce
d
b
y
ca
p
t
u
r
in
g
an
d
an
aly
zin
g
lar
g
e
v
o
l
u
m
es
o
f
d
ata
in
r
ea
l
tim
e.
Stre
am
p
r
o
ce
s
s
in
g
is
in
ar
g
u
ab
ly
o
n
e
o
f
th
e
m
o
s
t
v
alu
ab
le
co
m
p
o
n
en
ts
o
f
th
e
p
r
o
p
o
s
ed
s
y
s
tem
.
B
y
en
ab
lin
g
r
ea
l
-
tim
e
co
m
p
u
tatio
n
s
,
it e
q
u
ip
s
b
u
s
in
e
s
s
es
with
th
e
ab
ilit
y
to
g
e
n
e
r
at
e
im
m
e
d
i
ate
i
n
s
i
g
h
ts
,
o
p
ti
m
iz
e
o
p
e
r
a
ti
o
n
s
,
a
n
d
m
a
k
e
p
r
o
ac
t
iv
e
d
e
cisi
o
n
s
b
ase
d
o
n
liv
e
d
a
ta
s
t
r
ea
m
s
.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
Au
th
o
r
s
s
tate
n
o
f
u
n
d
in
g
in
v
o
lv
ed
.
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
al
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
B
en
y
m
o
l Jo
s
e
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Sajim
o
n
Ab
r
ah
am
✓
✓
✓
✓
✓
✓
✓
✓
Nik
ith
a
Sajith
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
A
n
in
tel
lig
en
t A
I
fr
a
mewo
r
k
fo
r
b
u
s
in
ess
in
s
ig
h
t g
en
era
tio
n
f
r
o
m
s
tr
ea
min
g
…
(
B
en
ymo
l
Jo
s
e
)
4079
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
Au
th
o
r
s
s
tate
n
o
co
n
f
lict o
f
in
t
er
est.
DATA AV
AI
L
AB
I
L
I
T
Y
T
h
e
d
ata
th
at
s
u
p
p
o
r
t
th
e
f
in
d
in
g
s
o
f
th
is
s
tu
d
y
ar
e
o
p
e
n
ly
av
ailab
le
in
Kag
g
le
at
h
ttp
s
://www.
k
ag
g
le.
co
m
/d
atasets
/b
ea
g
lelee
/am
az
o
n
-
r
ev
iews
-
us
-
s
p
o
r
ts
-
v1
-
00
-
ts
v
,
r
ef
er
e
n
c
e
[
22
].
RE
F
E
R
E
NC
E
S
[
1
]
G
.
M
o
h
a
n
r
a
j
,
R
.
K
.
N
a
d
e
sh
,
J
.
Ja
g
a
n
n
a
t
h
a
n
,
S
.
A
d
i
t
y
a
n
,
a
n
d
V
.
S
a
t
h
i
y
a
m
o
o
r
t
h
i
,
“
M
o
n
i
t
o
r
i
n
g
i
n
c
i
d
e
n
t
r
e
s
p
o
n
s
e
u
si
n
g
r
e
a
l
-
t
i
m
e
a
n
a
l
y
t
i
c
s,
”
i
n
2
n
d
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
I
T
I
n
n
o
v
a
t
i
o
n
s
a
n
d
K
n
o
w
l
e
d
g
e
D
i
sc
o
v
e
ry,
I
T
I
K
D
2
0
2
4
,
2
0
2
5
,
p
p
.
1
–
8
,
d
o
i
:
1
0
.
1
1
0
9
/
I
TI
K
D
6
3
5
7
4
.
2
0
2
5
.
1
1
0
0
4
8
2
4
.
[
2
]
M
d
A
.
A
l
a
m,
A
.
R
.
N
a
b
i
l
,
A
.
A
.
M
i
n
t
o
o
,
A
.
I
sl
a
m
,
“
R
e
a
l
-
t
i
m
e
a
n
a
l
y
t
i
c
s
i
n
st
r
e
a
mi
n
g
b
i
g
d
a
t
a
:
t
e
c
h
n
i
q
u
e
s
a
n
d
a
p
p
l
i
c
a
t
i
o
n
s
,
”
J
o
u
rn
a
l
o
f
S
c
i
e
n
c
e
a
n
d
En
g
i
n
e
e
r
i
n
g
R
e
se
a
rc
h
,
v
o
l
.
1
,
n
o
.
1
,
O
c
t
.
2
0
2
4
,
d
o
i
:
1
0
.
7
0
0
0
8
/
j
e
s
e
r
.
v
1
i
0
1
.
5
6
.
[
3
]
A
.
V
o
g
e
l
,
S
.
H
e
n
n
i
n
g
,
E.
W
o
h
l
f
e
i
l
,
O
.
Er
t
l
,
a
n
d
R
.
R
a
b
i
s
e
r
,
“
A
c
o
m
p
r
e
h
e
n
s
i
v
e
b
e
n
c
h
mar
k
i
n
g
a
n
a
l
y
si
s
o
f
f
a
u
l
t
r
e
c
o
v
e
r
y
i
n
st
r
e
a
m
p
r
o
c
e
ss
i
n
g
f
r
a
m
e
w
o
r
k
s,”
i
n
D
EB
S
2
0
2
4
-
1
8
t
h
A
C
M
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
D
i
s
t
ri
b
u
t
e
d
a
n
d
Ev
e
n
t
-
B
a
sed
S
y
s
t
e
m
s
,
J
u
n
.
2
0
2
4
,
p
p
.
1
7
1
–
1
8
2
,
d
o
i
:
1
0
.
1
1
4
5
/
3
6
2
9
1
0
4
.
3
6
6
6
0
4
0
.
[
4
]
S
.
H
e
n
n
i
n
g
a
n
d
W
.
H
a
sse
l
b
r
i
n
g
,
“
B
e
n
c
h
mar
k
i
n
g
sc
a
l
a
b
i
l
i
t
y
o
f
s
t
r
e
a
m
p
r
o
c
e
ssi
n
g
f
r
a
m
e
w
o
r
k
s
d
e
p
l
o
y
e
d
a
s
m
i
c
r
o
serv
i
c
e
s
i
n
t
h
e
c
l
o
u
d
,
”
J
o
u
rn
a
l
o
f
S
y
s
t
e
m
s
a
n
d
S
o
f
t
w
a
re
,
v
o
l
.
2
0
8
,
O
c
t
.
2
0
2
4
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
j
ss.
2
0
2
3
.
1
1
1
8
7
9
.
[
5
]
R
.
N
.
S
.
W
i
d
o
d
o
,
H
.
A
b
e
,
a
n
d
K
.
K
a
t
o
,
“
H
a
d
o
o
p
d
a
t
a
r
e
d
u
c
t
i
o
n
f
r
a
mew
o
r
k
:
a
p
p
l
y
i
n
g
d
a
t
a
r
e
d
u
c
t
i
o
n
a
t
t
h
e
D
F
S
l
a
y
e
r
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
9
,
p
p
.
1
5
2
7
0
4
–
1
5
2
7
1
7
,
2
0
2
1
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
1
.
3
1
2
7
4
9
9
.
[
6
]
M
.
U
.
D
e
mi
r
e
z
e
n
a
n
d
T.
S
.
N
a
v
r
u
z
,
“
P
e
r
f
o
r
m
a
n
c
e
a
n
a
l
y
s
i
s
o
f
l
a
m
b
d
a
a
r
c
h
i
t
e
c
t
u
r
e
-
b
a
s
e
d
b
i
g
-
d
a
t
a
s
y
st
e
ms
o
n
a
i
r
/
g
r
o
u
n
d
su
r
v
e
i
l
l
a
n
c
e
a
p
p
l
i
c
a
t
i
o
n
w
i
t
h
A
D
S
-
B
da
t
a
,
”
S
e
n
so
rs
,
v
o
l
.
2
3
,
n
o
.
1
7
,
A
u
g
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
3
1
7
7
5
8
0
.
[
7
]
M
.
F
r
a
g
k
o
u
l
i
s,
P
.
C
a
r
b
o
n
e
,
V
.
K
a
l
a
v
r
i
,
a
n
d
A
.
K
a
t
si
f
o
d
i
mo
s,
“
A
s
u
r
v
e
y
o
n
t
h
e
e
v
o
l
u
t
i
o
n
o
f
s
t
r
e
a
m
p
r
o
c
e
ssi
n
g
sy
st
e
ms,
”
VLD
B
J
o
u
rn
a
l
,
v
o
l
.
3
3
,
n
o
.
2
,
p
p
.
5
0
7
–
5
4
1
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s0
0
7
7
8
-
0
2
3
-
0
0
8
1
9
-
8.
[
8
]
A
.
A
.
M
u
n
sh
i
a
n
d
Y
.
A
.
-
R
.
I
.
M
o
h
a
med
,
“
D
a
t
a
l
a
k
e
l
a
mb
d
a
a
r
c
h
i
t
e
c
t
u
r
e
f
o
r
sm
a
r
t
g
r
i
d
s
b
i
g
d
a
t
a
a
n
a
l
y
t
i
c
s
,
”
I
E
EE
A
c
c
e
ss
,
v
o
l
.
6
,
p
p
.
4
0
4
6
3
-
4
0
4
7
1
,
2
0
1
8
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
1
8
.
2
8
5
8
2
5
6
.
[
9
]
Z.
H
a
s
a
n
,
H
.
J.
X
i
n
g
,
a
n
d
M
.
I
.
M
a
g
r
a
y
,
“
B
i
g
d
a
t
a
mac
h
i
n
e
l
e
a
r
n
i
n
g
u
si
n
g
A
p
a
c
h
e
S
p
a
r
k
ml
l
i
b
,
”
Me
s
o
p
o
t
a
m
i
a
n
J
o
u
rn
a
l
o
f
Bi
g
D
a
t
a
,
v
o
l
.
2
0
2
2
,
p
p
.
1
–
1
1
,
2
0
2
2
,
d
o
i
:
1
0
.
5
8
4
9
6
/
M
J
B
D
/
2
0
2
2
/
0
0
1
.
[
1
0
]
U
.
S
u
t
h
a
k
a
r
,
L.
M
a
g
n
o
n
i
,
D
.
R
.
S
m
i
t
h
,
a
n
d
A
.
K
h
a
n
,
“
O
p
t
i
mi
s
e
d
l
a
mb
d
a
a
r
c
h
i
t
e
c
t
u
r
e
f
o
r
mo
n
i
t
o
r
i
n
g
sc
i
e
n
t
i
f
i
c
i
n
f
r
a
st
r
u
c
t
u
r
e
,
”
I
EEE
T
r
a
n
s
a
c
t
i
o
n
s
o
n
P
a
ra
l
l
e
l
a
n
d
D
i
st
r
i
b
u
t
e
d
S
y
st
e
m
s
,
v
o
l
.
3
2
,
n
o
.
6
,
p
p
.
1
3
9
5
–
1
4
0
8
,
2
0
2
1
,
d
o
i
:
1
0
.
1
1
0
9
/
TPD
S
.
2
0
1
7
.
2
7
7
2
2
4
1
.
[
1
1
]
M
.
V
.
K
a
m
a
l
,
P
.
D
i
l
e
e
p
,
a
n
d
D
.
V
a
s
u
ma
t
i
,
“
S
p
a
r
k
s
t
r
e
a
m
i
n
g
f
o
r
p
r
e
d
i
c
t
i
v
e
b
u
s
i
n
e
ss
i
n
t
e
l
l
i
g
e
n
c
e
,
”
i
n
S
o
f
t
C
o
m
p
u
t
i
n
g
a
n
d
S
i
g
n
a
l
Pro
c
e
ssi
n
g
,
S
i
n
g
a
p
o
r
e
:
S
p
r
i
n
g
e
r
,
2
0
1
9
,
p
p
.
2
8
9
–
2
9
8
,
d
o
i
:
1
0
.
1
0
0
7
/
9
7
8
-
981
-
13
-
3
3
9
3
-
4
_
3
0
.
[
1
2
]
A
.
S
v
y
a
t
k
o
v
s
k
i
y
,
K
.
I
mai
,
M
.
K
r
o
e
g
e
r
,
a
n
d
Y
.
S
h
i
r
a
i
t
o
,
“
La
r
g
e
-
sc
a
l
e
t
e
x
t
p
r
o
c
e
ss
i
n
g
p
i
p
e
l
i
n
e
w
i
t
h
A
p
a
c
h
e
S
p
a
r
k
,
”
i
n
2
0
1
6
I
E
EE
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
B
i
g
D
a
t
a
,
Bi
g
D
a
t
a
2
0
1
6
,
D
e
c
.
2
0
1
6
,
p
p
.
3
9
2
8
–
3
9
3
5
,
d
o
i
:
1
0
.
1
1
0
9
/
B
i
g
D
a
t
a
.
2
0
1
6
.
7
8
4
1
0
6
8
.
[
1
3
]
A
.
A
h
me
d
,
R
.
X
i
,
M
.
H
o
u
,
S
.
A
.
S
h
a
h
,
a
n
d
S
.
H
a
me
e
d
,
“
H
a
r
n
e
ssi
n
g
b
i
g
d
a
t
a
a
n
a
l
y
t
i
c
s
f
o
r
h
e
a
l
t
h
c
a
r
e
:
a
c
o
mp
r
e
h
e
n
si
v
e
r
e
v
i
e
w
o
f
f
r
a
mew
o
r
k
s,
i
m
p
l
i
c
a
t
i
o
n
s,
a
p
p
l
i
c
a
t
i
o
n
s,
a
n
d
i
m
p
a
c
t
s
,
”
I
EEE
Ac
c
e
ss
,
v
o
l
.
1
1
,
p
p
.
1
1
2
8
9
1
–
1
1
2
9
2
8
,
2
0
2
3
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
3
.
3
3
2
3
5
7
4
.
[
1
4
]
E.
A
z
h
i
r
,
M
.
H
o
sse
i
n
z
a
d
e
h
,
F
.
K
h
a
n
,
a
n
d
A
.
M
o
s
a
v
i
,
“
P
e
r
f
o
r
m
a
n
c
e
e
v
a
l
u
a
t
i
o
n
o
f
q
u
e
r
y
p
l
a
n
r
e
c
o
mm
e
n
d
a
t
i
o
n
w
i
t
h
A
p
a
c
h
e
H
a
d
o
o
p
a
n
d
A
p
a
c
h
e
S
p
a
r
k
,
”
M
a
t
h
e
m
a
t
i
c
s
,
v
o
l
.
1
0
,
n
o
.
1
9
,
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
ma
t
h
1
0
1
9
3
5
1
7
.
[
1
5
]
L.
R
.
N
a
i
r
,
S
.
D
.
S
h
e
t
t
y
,
a
n
d
S
.
D
.
S
h
e
t
t
y
,
“
A
p
p
l
y
i
n
g
s
p
a
r
k
-
b
a
se
d
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
m
o
d
e
l
o
n
st
r
e
a
m
i
n
g
b
i
g
d
a
t
a
f
o
r
h
e
a
l
t
h
s
t
a
t
u
s
p
r
e
d
i
c
t
i
o
n
,
”
C
o
m
p
u
t
e
rs
a
n
d
E
l
e
c
t
ri
c
a
l
En
g
i
n
e
e
ri
n
g
,
v
o
l
.
6
5
,
p
p
.
3
9
3
–
3
9
9
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
e
l
e
c
e
n
g
.
2
0
1
7
.
0
3
.
0
0
9
.
[
1
6
]
T.
K
o
l
a
j
o
,
O
.
D
a
r
a
mo
l
a
,
a
n
d
A
.
A
d
e
b
i
y
i
,
“
B
i
g
d
a
t
a
st
r
e
a
m
a
n
a
l
y
s
i
s:
a
sy
s
t
e
ma
t
i
c
l
i
t
e
r
a
t
u
r
e
r
e
v
i
e
w
,
”
J
o
u
rn
a
l
o
f
Bi
g
D
a
t
a
,
v
o
l
.
6
,
n
o
.
1
,
2
0
1
9
,
d
o
i
:
1
0
.
1
1
8
6
/
s
4
0
5
3
7
-
0
1
9
-
0
2
1
0
-
7.
[
1
7
]
K
.
J.
M
a
t
t
e
u
ssi
,
J.
C
.
S
.
D
.
A
n
j
o
s
,
V
.
R
.
Q
.
Le
i
t
h
a
r
d
t
,
a
n
d
C
.
F
.
R
.
G
e
y
e
r
,
“
P
e
r
f
o
r
m
a
n
c
e
e
v
a
l
u
a
t
i
o
n
a
n
a
l
y
s
i
s
o
f
S
p
a
r
k
S
t
r
e
a
m
i
n
g
b
a
c
k
p
r
e
ss
u
r
e
f
o
r
d
a
t
a
-
i
n
t
e
n
s
i
v
e
p
i
p
e
l
i
n
e
s,
”
S
e
n
s
o
rs
,
v
o
l
.
2
2
,
n
o
.
1
3
,
J
u
n
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
s
2
2
1
3
4
7
5
6
.
[
1
8
]
A
p
a
c
h
e
S
p
a
r
k
,
“
A
p
a
c
h
e
S
p
a
r
k
,
”
sp
a
r
k
.
a
p
a
c
h
e
.
o
r
g
.
A
c
c
e
sse
d
:
Ju
n
.
1
9
,
2
0
2
5
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s:
/
/
s
p
a
r
k
.
a
p
a
c
h
e
.
o
r
g
/
[
1
9
]
F
.
C
e
r
e
z
o
,
C
.
E.
C
u
e
st
a
,
J.
C
.
M
.
-
H
e
r
r
a
n
z
,
a
n
d
B
.
V
e
l
a
,
“
D
e
c
o
n
st
r
u
c
t
i
n
g
t
h
e
l
a
m
b
d
a
a
r
c
h
i
t
e
c
t
u
r
e
:
a
n
e
x
p
e
r
i
e
n
c
e
r
e
p
o
r
t
,
”
i
n
2
0
1
9
I
EEE
I
n
t
.
C
o
n
f
.
S
o
f
t
w
a
re
Ar
c
h
i
t
e
c
t
u
r
e
C
o
m
p
a
n
i
o
n
(
I
C
S
A
-
C)
,
2
0
1
9
,
p
p
.
1
9
6
–
2
0
1
,
d
o
i
:
1
0
.
1
1
0
9
/
I
C
S
A
-
C
.
2
0
1
9
.
0
0
0
4
2
.
[
2
0
]
A
maz
o
n
W
e
b
S
e
r
v
i
c
e
s
,
“
S
t
r
e
a
mi
n
g
d
a
t
a
,”
a
w
s.a
maz
o
n
.
c
o
m
.
A
c
c
e
ss
e
d
:
J
u
n
.
2
1
,
2
0
2
5
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
a
w
s.
a
ma
z
o
n
.
c
o
m/
s
t
r
e
a
m
i
n
g
-
d
a
t
a
/
[
2
1
]
A
p
a
c
h
e
S
p
a
r
k
,
“
M
L
p
i
p
e
l
i
n
e
s
-
S
p
a
r
k
3
.
0
.
0
d
o
c
u
me
n
t
a
t
i
o
n
,
”
s
p
a
r
k
.
a
p
a
c
h
e
.
o
r
g
.
A
c
c
e
sse
d
:
Ju
n
.
2
1
,
2
0
2
5
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
sp
a
r
k
.
a
p
a
c
h
e
.
o
r
g
/
d
o
c
s
/
l
a
t
e
st
/
m
l
-
p
i
p
e
l
i
n
e
.
h
t
m
l
[
2
2
]
B
.
L
e
e
,
“
A
ma
z
o
n
r
e
v
i
e
w
s
U
S
–
s
p
o
r
t
s
v
1
.
0
0
TSV
,
”
k
a
g
g
l
e
.
c
o
m.
A
c
c
e
sse
d
:
Ja
n
.
8
,
2
0
2
5
.
[
O
n
l
i
n
e
]
.
A
v
a
i
l
a
b
l
e
:
h
t
t
p
s
:
/
/
w
w
w
.
k
a
g
g
l
e
.
c
o
m
/
d
a
t
a
se
t
s/
b
e
a
g
l
e
l
e
e
/
a
ma
z
o
n
-
r
e
v
i
e
w
s
-
us
-
sp
o
r
t
s
-
v1
-
00
-
t
s
v
[
2
3
]
R
.
R
a
n
j
a
n
,
S
.
G
a
r
g
,
A
.
R
.
K
h
o
s
k
b
a
r
,
E.
S
o
l
a
i
ma
n
,
P
.
J
a
mes
,
a
n
d
D
.
G
e
o
r
g
a
k
o
p
o
u
l
o
s,
“
O
r
c
h
e
st
r
a
t
i
n
g
b
i
g
d
a
t
a
a
n
a
l
y
s
i
s w
o
r
k
f
l
o
w
s,
”
I
EEE
C
l
o
u
d
C
o
m
p
u
t
i
n
g
,
v
o
l
.
4
,
n
o
.
3
,
p
p
.
2
0
–
2
8
,
M
a
y
2
0
1
7
,
d
o
i
:
1
0
.
1
1
0
9
/
M
C
C
.
2
0
1
7
.
5
5
.
[
2
4
]
J.
C
.
-
R
o
d
r
í
g
u
e
z
e
t
a
l
.
,
“
A
p
a
c
h
e
h
i
v
e
:
f
r
o
m
ma
p
r
e
d
u
c
e
t
o
e
n
t
e
r
p
r
i
se
-
g
r
a
d
e
b
i
g
d
a
t
a
w
a
r
e
h
o
u
si
n
g
,
”
i
n
S
I
G
MOD
'
1
9
:
2
0
1
9
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
Ma
n
a
g
e
m
e
n
t
o
f
D
a
t
a
,
2
0
1
9
,
p
p
.
1
7
7
3
–
1
7
8
6
,
d
o
i
:
1
0
.
1
1
4
5
/
3
2
9
9
8
6
9
.
3
3
1
4
0
4
5
.
[
2
5
]
K
.
P
a
p
i
n
e
n
i
,
“
W
h
y
i
n
v
e
r
s
e
d
o
c
u
me
n
t
f
r
e
q
u
e
n
c
y
?
,
”
i
n
2
n
d
Me
e
t
i
n
g
o
f
t
h
e
N
o
r
t
h
Am
e
r
i
c
a
n
C
h
a
p
t
e
r
o
f
t
h
e
Asso
c
i
a
t
i
o
n
f
o
r
C
o
m
p
u
t
a
t
i
o
n
a
l
L
i
n
g
u
i
s
t
i
c
s,
N
AA
C
L
2
0
0
1
,
2
0
0
1
,
p
p
.
1
–
8
,
d
o
i
:
1
0
.
3
1
1
5
/
1
0
7
3
3
3
6
.
1
0
7
3
3
4
0
.
[
2
6
]
N
.
A
.
-
Te
j
e
r
a
,
M
.
G
a
marr
a
,
J.
I
.
V
é
l
e
z
,
a
n
d
E.
Zu
r
e
k
,
“
A
d
i
st
a
n
c
e
-
b
a
se
d
k
e
r
n
e
l
f
o
r
c
l
a
ss
i
f
i
c
a
t
i
o
n
v
i
a
s
u
p
p
o
r
t
v
e
c
t
o
r
m
a
c
h
i
n
e
s
,
”
Fro
n
t
i
e
rs
i
n
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
7
,
F
e
b
.
2
0
2
4
,
d
o
i
:
1
0
.
3
3
8
9
/
f
r
a
i
.
2
0
2
4
.
1
2
8
7
8
7
5
.
[
2
7
]
Z.
Y
a
n
g
e
t
a
l
.
,
“
A
n
e
w
t
h
r
e
e
-
w
a
y
i
n
c
r
e
me
n
t
a
l
n
a
i
v
e
B
a
y
e
s
c
l
a
ss
i
f
i
e
r
,
”
El
e
c
t
ro
n
i
c
s
,
v
o
l
.
1
2
,
n
o
.
7
,
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
e
l
e
c
t
r
o
n
i
c
s1
2
0
7
1
7
3
0
.
[
2
8
]
A
.
K
.
G
á
r
a
t
e
e
sc
a
mi
l
l
a
,
A
.
H
.
E
l
H
a
ss
a
n
i
,
a
n
d
E.
A
n
d
r
e
s,
“
B
i
g
d
a
t
a
e
x
e
c
u
t
i
o
n
t
i
me
b
a
se
d
o
n
s
p
a
r
k
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
l
i
b
r
a
r
i
e
s
,
”
i
n
I
C
C
BD
C
'
1
9
:
2
0
1
9
3
rd
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
C
l
o
u
d
a
n
d
Bi
g
D
a
t
a
C
o
m
p
u
t
i
n
g
,
A
u
g
.
2
0
1
9
,
p
p
.
7
8
–
83
,
d
o
i
:
1
0
.
1
1
4
5
/
3
3
5
8
5
0
5
.
3
3
5
8
5
1
9
.
[
2
9
]
E.
S
c
h
e
c
h
t
ma
n
a
n
d
G
.
S
c
h
e
c
h
t
m
a
n
,
“
Th
e
r
e
l
a
t
i
o
n
sh
i
p
b
e
t
w
e
e
n
G
i
n
i
t
e
r
m
i
n
o
l
o
g
y
a
n
d
t
h
e
R
O
C
c
u
r
v
e
,
”
Me
t
ro
n
,
v
o
l
.
7
7
,
n
o
.
3
,
p
p
.
1
7
1
–
1
7
8
,
D
e
c
.
2
0
1
9
,
d
o
i
:
1
0
.
1
0
0
7
/
s4
0
3
0
0
-
019
-
0
0
1
6
0
-
7.
[
3
0
]
C
.
G
i
n
i
,
“
O
n
t
h
e
me
a
su
r
e
me
n
t
o
f
c
o
n
c
e
n
t
r
a
t
i
o
n
a
n
d
v
a
r
i
a
b
i
l
i
t
y
o
f
c
h
a
r
a
c
t
e
r
s,”
M
e
t
r
o
n
–
I
n
t
e
rn
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
S
t
a
t
i
s
t
i
c
s
,
v
o
l
.
63
,
n
o
.
1
,
p
p
.
1
–
3
8
,
2
0
0
5
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
1
5
,
No
.
5
,
Octo
b
er
2
0
2
6
:
4
0
7
1
-
4
0
8
0
4080
[
3
1
]
S
.
G
h
o
s
h
,
A
.
D
a
sg
u
p
t
a
,
a
n
d
A
.
S
w
e
t
a
p
a
d
ma
,
“
A
st
u
d
y
o
n
s
u
p
p
o
r
t
v
e
c
t
o
r
mac
h
i
n
e
b
a
s
e
d
l
i
n
e
a
r
a
n
d
n
o
n
-
l
i
n
e
a
r
p
a
t
t
e
r
n
c
l
a
ss
i
f
i
c
a
t
i
o
n
,
”
i
n
2
0
1
9
I
n
t
e
rn
a
t
i
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
I
n
t
e
l
l
i
g
e
n
t
S
u
s
t
a
i
n
a
b
l
e
S
y
st
e
m
s
(
I
C
I
S
S
)
,
2
0
1
9
,
p
p
.
2
4
–
28
,
d
o
i
:
1
0
.
1
1
0
9
/
I
S
S
1
.
2
0
1
9
.
8
9
0
8
0
1
8
.
[
3
2
]
Y
.
C
h
e
n
,
C
.
Li
,
a
n
d
H
.
W
a
n
g
,
“
B
i
g
d
a
t
a
a
n
d
p
r
e
d
i
c
t
i
v
e
a
n
a
l
y
t
i
c
s
f
o
r
b
u
s
i
n
e
ss
i
n
t
e
l
l
i
g
e
n
c
e
:
a
b
i
b
l
i
o
g
r
a
p
h
i
c
s
t
u
d
y
(
2
0
0
0
–
2
0
2
1
)
,
”
Fo
re
c
a
st
i
n
g
,
v
o
l
.
4
,
n
o
.
4
,
p
p
.
7
6
7
–
7
8
6
,
S
e
p
.
2
0
2
2
,
d
o
i
:
1
0
.
3
3
9
0
/
f
o
r
e
c
a
s
t
4
0
4
0
0
4
2
.
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
Dr
.
Be
n
y
m
o
l
J
o
se
p
u
rse
d
m
a
ste
rs
in
Co
m
p
u
ter
S
c
ien
c
e
fro
m
Bh
a
ra
th
i
d
a
sa
n
Un
iv
e
rsity
,
T
h
iru
c
h
irap
p
a
ll
i
,
Ta
m
il
Na
d
u
i
n
1
9
9
9
;
a
n
M
.
P
h
i
l.
in
Co
m
p
u
ter
S
c
ien
c
e
fro
m
M
a
d
u
ra
i
Ka
m
a
ra
j
Un
iv
e
rsity
,
M
a
d
u
ra
i,
Tam
il
Na
d
u
in
t
h
e
y
e
a
r
2
0
1
4
;
a
n
d
P
h
.
D
.
in
C
o
m
p
u
ter
S
c
ien
c
e
fro
m
M
a
h
a
tma
G
a
n
d
h
i
Un
iv
e
rsity
,
Ko
tt
a
y
a
m
,
Ke
ra
la,
In
d
ia
i
n
2
0
2
1
on
t
h
e
to
p
ic
“
Un
stru
c
tu
re
d
d
a
ta
m
in
i
n
g
i
n
b
i
g
d
a
ta:
a
No
S
QL
p
e
rsp
e
c
ti
v
e
”
.
S
h
e
is
c
u
rre
n
tl
y
wo
rk
i
n
g
a
s
a
n
a
ss
o
c
iate
p
ro
fe
ss
o
r
in
t
h
e
De
p
a
rtme
n
t
o
f
Co
m
p
u
ter Ap
p
li
c
a
ti
o
n
s,
M
a
rian
Co
ll
e
g
e
K
u
tt
i
k
k
a
n
a
m
Au
to
n
o
m
o
u
s,
I
d
u
k
k
i
,
Ke
ra
la,
I
n
d
i
a
.
S
h
e
h
a
s
p
u
b
li
sh
e
d
m
a
n
y
p
a
p
e
rs
in
jo
u
rn
a
ls
a
n
d
c
o
n
fe
re
n
c
e
p
ro
c
e
e
d
in
g
s
,
i
n
c
lu
d
in
g
IEE
E,
El
s
e
v
ier,
a
n
d
AC
M
,
a
n
d
m
o
st
o
f
t
h
e
m
a
re
in
d
e
x
e
d
b
y
S
c
o
p
u
s
.
He
r
m
a
in
re
se
a
rc
h
wo
r
k
f
o
c
u
se
s
o
n
u
n
str
u
c
tu
re
d
d
a
ta
m
i
n
in
g
,
m
a
c
h
in
e
lea
rn
in
g
,
No
S
QL
d
a
tab
a
se
s
,
a
n
d
b
ig
d
a
ta
a
n
a
l
y
ti
c
s
.
S
h
e
h
a
s
2
6
y
e
a
rs
o
f
tea
c
h
in
g
e
x
p
e
rien
c
e
a
n
d
8
y
e
a
rs
o
f
re
se
a
rc
h
e
x
p
e
rien
c
e
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
b
e
n
y
m
o
l.
j
o
se
@m
a
rian
c
o
ll
e
g
e
.
o
rg
.
Dr
.
S
a
ji
m
o
n
A
b
r
a
h
a
m
is
a
M
CA,
M
.
S
c
.
M
a
th
e
m
a
ti
c
s,
M
BA,
P
h
.
D
.
in
Co
m
p
u
ter
S
c
ien
c
e
.
He
h
a
s
b
e
e
n
wo
rk
i
n
g
a
s
a
fa
c
u
lt
y
m
e
m
b
e
r
in
Co
m
p
u
ter
Ap
p
li
c
a
ti
o
n
s
a
n
d
IT,
S
c
h
o
o
l
o
f
M
a
n
a
g
e
m
e
n
t
a
n
d
B
u
sin
e
ss
S
t
u
d
ies
,
M
a
h
a
tma
G
a
n
d
h
i
Un
i
v
e
rsit
y
,
K
o
tt
a
y
a
m
,
Ke
ra
la,
I
n
d
ia.
He
c
u
rre
n
tl
y
h
o
ld
s
th
e
a
d
d
it
io
n
a
l
c
h
a
rg
e
o
f
d
irec
to
r
(H
o
n
)
,
u
n
i
v
e
rsity
c
e
n
ter
f
o
r
in
ter
n
a
ti
o
n
a
l
c
o
o
p
e
ra
ti
o
n
.
He
wa
s
p
re
v
io
u
sly
wo
rk
i
n
g
a
s
a
sy
ste
m
s
a
n
a
ly
st
in
th
e
I
n
stit
u
te
o
f
Hu
m
a
n
Re
so
u
rc
e
De
v
e
lo
p
m
e
n
t,
a
fa
c
u
lt
y
m
e
m
b
e
r
o
f
C
o
m
p
u
ter
A
p
p
li
c
a
ti
o
n
s
i
n
M
a
rian
Co
ll
e
g
e
,
Ku
tt
ik
k
a
n
a
m
,
a
n
d
a
d
a
tab
a
se
a
rc
h
it
e
c
t
in
Ro
y
a
l
Un
iv
e
rsity
o
f
Bh
u
tan
u
n
d
e
r
th
e
c
o
lo
m
b
o
p
lan
o
n
d
e
p
u
tati
o
n
t
h
ro
u
g
h
th
e
M
i
n
i
stry
o
f
Ex
tern
a
l
Affa
irs,
G
o
v
t.
o
f
In
d
ia.
His
re
se
a
rc
h
a
re
a
in
c
lu
d
e
s
d
a
ta
sc
ien
c
e
,
sp
a
ti
o
-
tem
p
o
ra
l
d
a
tab
a
se
s,
m
o
b
il
it
y
m
in
i
n
g
,
se
n
ti
m
e
n
t
a
n
a
ly
sis,
b
ig
d
a
ta
a
n
a
ly
ti
c
s,
e
-
lea
rn
in
g
,
d
a
tab
a
se
a
n
d
d
a
ta
m
in
in
g
,
d
a
ta
m
a
n
a
g
e
m
e
n
t,
we
b
c
li
c
k
stre
a
m
a
n
a
ly
sis,
a
n
d
b
u
si
n
e
ss
a
n
a
ly
ti
c
s
.
H
e
h
a
s
p
u
b
li
s
h
e
d
1
1
0
a
rti
c
les
in
n
a
ti
o
n
a
l,
in
tern
a
ti
o
n
a
l
jo
u
rn
a
ls
,
a
n
d
c
o
n
fe
re
n
c
e
p
ro
c
e
e
d
i
n
g
s
.
He
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
sa
ji
m
a
b
ra
h
a
m
@re
d
iffma
il
.
c
o
m
.
Nik
ith
a
S
a
jith
is
a
q
u
a
li
ty
e
n
g
in
e
e
r
a
t
LT
M
Li
m
it
e
d
i
n
Be
n
g
a
lu
ru
,
s
p
e
c
ializin
g
i
n
so
ftwa
re
q
u
a
li
ty
a
ss
u
ra
n
c
e
a
n
d
e
m
e
rg
in
g
tec
h
n
o
lo
g
ies
.
S
h
e
h
o
ld
s
a
Ba
c
h
e
lo
r
o
f
Tec
h
n
o
lo
g
y
i
n
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
En
g
in
e
e
rin
g
fr
o
m
Ch
rist
De
e
m
e
d
to
b
e
Un
i
v
e
rsity
,
Be
n
g
a
lu
ru
,
Ka
rn
a
tak
a
wh
e
re
sh
e
a
lso
se
rv
e
d
a
s
th
e
v
ice
c
h
a
ir
o
f
t
h
e
IE
EE
Wo
m
e
n
in
E
n
g
i
n
e
e
rin
g
S
tu
d
e
n
t
Ch
a
p
ter.
W
h
il
e
wo
rk
in
g
a
s
a
q
u
a
li
ty
e
n
g
i
n
e
e
r,
sh
e
is
a
lso
p
u
r
su
in
g
a
h
y
b
rid
M
a
ste
r
o
f
Tec
h
n
o
l
o
g
y
in
Arti
ficia
l
In
telli
g
e
n
c
e
a
n
d
Da
ta
S
c
ien
c
e
En
g
in
e
e
r
in
g
fr
o
m
In
d
ia
n
In
stit
u
te
o
f
Tec
h
n
o
l
o
g
y
P
a
tn
a
,
I
n
d
ia.
He
r
te
c
h
n
ica
l
i
n
tere
sts
in
c
lu
d
e
c
o
m
p
u
t
e
r
v
isio
n
,
m
a
c
h
in
e
lea
rn
i
n
g
,
a
n
d
th
e
a
p
p
li
c
a
ti
o
n
o
f
i
n
telli
g
e
n
t
a
u
to
m
a
ti
o
n
in
q
u
a
li
t
y
e
n
g
in
e
e
ri
n
g
.
S
h
e
h
a
s
c
o
n
tri
b
u
ted
t
o
m
u
lt
ip
le
p
r
o
jec
ts
i
n
v
o
lv
i
n
g
m
o
d
e
l
d
e
v
e
l
o
p
m
e
n
t,
p
e
rfo
rm
a
n
c
e
e
v
a
lu
a
ti
o
n
,
a
n
d
th
e
in
teg
ra
ti
o
n
o
f
AI
-
d
ri
v
e
n
s
o
lu
ti
o
n
s
i
n
to
so
ft
wa
re
wo
rk
flo
ws
.
S
h
e
c
o
n
ti
n
u
e
s
to
p
u
rs
u
e
re
se
a
rc
h
a
t
th
e
in
ters
e
c
ti
o
n
o
f
so
f
twa
re
q
u
a
li
t
y
,
AI,
a
n
d
e
m
e
rg
i
n
g
d
ig
it
a
l
tec
h
n
o
l
o
g
ies
.
S
h
e
c
a
n
b
e
c
o
n
tac
ted
a
t
e
m
a
il
:
n
ik
it
h
a
sa
ji
th
2
0
2
0
@g
m
a
il
.
c
o
m
.
Evaluation Warning : The document was created with Spire.PDF for Python.