I
AE
S
I
n
t
e
r
n
at
ion
al
Jou
r
n
al
of
Ar
t
if
icial
I
n
t
e
ll
ig
e
n
c
e
(
I
J
-
AI
)
Vol.
14
,
No.
4
,
Augus
t
2025
,
pp.
2559
~
2567
I
S
S
N:
2252
-
8938
,
DO
I
:
10
.
11591/i
jai
.
v
14
.i
4
.
pp
25
59
-
2567
2559
Jou
r
n
al
h
omepage
:
ht
tp:
//
ij
ai
.
iaes
c
or
e
.
c
om
C
h
al
le
n
g
e
s of
r
e
c
o
m
m
e
n
d
e
r
s
yst
e
m
s i
n
f
in
an
c
e
a
n
d
b
an
k
in
g:
a
syste
m
at
i
c
r
e
vi
e
w
L
os
s
an
B
on
d
e
1
,
Abd
ou
l
Karim
B
ichan
ga
2
1
D
e
pa
r
tm
e
nt
of
A
ppl
ie
d S
c
ie
nc
e
s
, S
c
hool
of
P
os
tg
r
a
dua
te
S
tu
di
e
s
,
A
dve
nt
is
t
U
ni
ve
r
s
it
y of
A
f
r
ic
a
, N
a
ir
obi
, K
e
nya
2
S
c
hool
of
C
omput
in
g, N
a
z
i
B
oni
U
ni
ve
r
s
it
y, B
obo
-
D
io
ul
a
s
s
o,
B
ur
ki
na
F
a
s
o
Ar
t
icle
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
ived
Apr
17,
2024
R
e
vis
e
d
F
e
b
25,
2025
Ac
c
e
pted
M
a
r
15,
2025
Reco
mmen
d
er
s
y
s
t
ems
are
w
i
d
el
y
ap
p
l
i
e
d
i
n
v
ari
o
u
s
d
o
mai
n
s
,
i
n
cl
u
d
i
n
g
e
-
co
mmerce,
mark
et
i
n
g
,
an
d
ed
u
ca
t
i
o
n
.
D
es
p
i
t
e
t
h
ei
r
p
o
p
u
l
ari
t
y
,
reco
mmen
d
er
s
y
s
t
em
s
are
n
o
t
w
i
d
el
y
u
s
ed
i
n
fi
n
a
n
c
e
an
d
b
an
k
i
n
g
.
T
h
i
s
p
ap
er
ai
m
s
t
o
i
d
en
t
i
f
y
t
h
e
ch
al
l
en
g
es
as
s
o
c
i
at
e
d
w
i
t
h
u
s
i
n
g
reco
mme
n
d
e
r
s
y
s
t
ems
i
n
fi
n
an
ce
a
n
d
b
a
n
k
i
n
g
an
d
reco
mmen
d
d
i
rect
i
o
n
s
fo
r
f
u
t
u
re
res
earch
.
U
s
i
n
g
a
s
y
s
t
ema
t
i
c
l
i
t
era
t
u
re
rev
i
e
w
(SL
R)
me
t
h
o
d
,
5
2
p
ap
er
s
w
ere
s
el
ec
t
ed
a
n
d
a
n
al
y
ze
d
.
A
t
h
ree
-
s
t
e
p
p
r
o
ces
s
w
as
u
s
e
d
t
o
mak
e
t
h
e
s
e
l
ect
i
o
n
.
Fi
rs
t
,
a
k
e
y
w
o
rd
s
e
arch
w
a
s
mad
e
t
o
i
d
en
t
i
f
y
a
s
ee
d
l
i
s
t
o
f
s
o
u
rces
.
A
s
n
o
w
b
al
l
t
ec
h
n
i
q
u
e
w
i
t
h
s
p
eci
f
i
c
i
n
cl
u
s
i
o
n
an
d
ex
cl
u
s
i
o
n
cr
i
t
er
i
a
w
a
s
ap
p
l
i
e
d
t
o
ex
p
an
d
t
h
e
l
i
s
t
.
Fi
n
al
l
y
,
a
q
u
i
ck
s
t
u
d
y
w
as
mad
e
t
o
p
r
o
d
u
ce
t
h
e
fi
n
a
l
l
i
s
t
o
f
s
o
u
rce
s
t
o
co
n
s
i
d
er.
T
h
ro
u
g
h
t
h
e
s
t
u
d
y
o
f
t
h
e
5
2
rel
ev
a
n
t
p
a
p
ers
,
t
h
ree
mai
n
ch
a
l
l
e
n
g
e
s
:
i
)
t
ran
s
p
are
n
cy
,
et
h
i
cs
,
an
d
d
at
a
p
ri
v
acy
;
ii
)
h
an
d
l
i
n
g
co
mp
l
ex
co
n
t
e
n
t
i
n
f
o
rmat
i
o
n
an
d
acc
o
u
n
t
i
n
g
fo
r
mu
l
t
i
p
l
e
u
s
er
b
eh
a
v
i
o
rs
;
an
d
i
i
i
)
e
x
p
l
ai
n
ab
i
l
i
t
y
o
f
A
I
mo
d
el
s
w
ere
i
d
e
n
t
i
fi
ed
.
T
h
i
s
s
t
u
d
y
h
a
s
es
t
a
b
l
i
s
h
ed
t
h
e
b
arr
i
ers
t
o
ad
o
p
t
i
n
g
reco
mme
n
d
er
s
y
s
t
e
ms
i
n
t
h
e
fi
n
a
n
ce
an
d
b
an
k
i
n
g
i
n
d
u
s
t
ry
.
Sp
ec
i
fi
c
s
u
b
j
ect
s
o
f
co
n
cern
i
d
e
n
t
i
fi
ed
i
n
c
l
u
d
e
co
l
d
-
s
t
ar
t
p
ro
b
l
em
s
,
p
ers
o
n
a
l
i
za
t
i
o
n
,
fra
u
d
d
et
ec
t
i
o
n
,
t
ran
s
p
are
n
c
y
,
an
d
d
at
a
p
ri
v
acy
.
T
h
e
s
t
u
d
y
reco
mmen
d
s
fu
r
t
h
er
res
earc
h
l
ev
era
g
i
n
g
ad
v
an
ced
mach
i
n
e
l
earn
i
n
g
mo
d
e
l
s
a
n
d
emerg
i
n
g
t
ec
h
n
o
l
o
g
i
es
t
o
fi
l
l
t
h
e
g
a
p
.
K
e
y
w
o
r
d
s
:
C
ha
ll
e
nge
s
of
r
e
c
omm
e
nda
ti
on
f
or
f
inanc
e
a
nd
ba
nking
E
t
h
i
c
s
a
n
d
p
r
i
v
a
c
y
o
f
A
I
s
o
l
u
t
i
o
n
s
E
xplaina
bil
it
y
o
f
AI
m
ode
ls
R
e
c
omm
e
nde
r
s
ys
tems
R
e
li
a
bil
it
y
of
A
I
m
ode
ls
Th
i
s
i
s
a
n
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
t
h
e
CC
B
Y
-
SA
l
i
ce
n
s
e.
C
or
r
e
s
pon
din
g
A
u
th
or
:
L
os
s
a
n
B
onde
De
pa
r
tm
e
nt
of
Applied
S
c
ienc
e
s
,
S
c
hool
o
f
P
os
tgr
a
dua
te
S
tudi
e
s
,
Adve
nti
s
t
Unive
r
s
it
y
o
f
Af
r
ica
P
r
ivate
B
a
g,
M
ba
ga
thi
,
00503,
Na
ir
obi,
Ke
nya
E
mail:
bonde
l@au
a
.
a
c
.
ke
1.
I
NT
RODU
C
T
I
ON
I
n
thi
s
c
e
ntur
y
of
inf
o
r
mation
ove
r
load,
r
e
c
omm
e
nde
r
s
ys
tems
a
r
e
be
ing
s
tr
a
tegic
a
ll
y
a
dopted
by
many
indus
tr
ies
,
including
r
e
tail,
e
nter
tainment,
e
-
c
omm
e
r
c
e
,
mar
ke
ti
ng
,
bus
ines
s
e
s
,
a
nd
e
duc
a
ti
on,
to
guide
us
e
r
s
,
c
us
tom
e
r
s
,
a
nd
p
r
os
pe
c
ts
in
c
hoos
ing
a
mon
g
va
s
t
a
mount
s
of
a
va
il
a
ble
op
ti
ons
or
of
f
e
r
s
.
S
ur
pr
is
ingl
y,
de
s
pit
e
the
g
r
owing
popular
it
y
of
r
e
c
omm
e
nde
r
s
ys
tems
,
they
a
r
e
not
c
ur
r
e
ntl
y
wide
ly
us
e
d
in
the
f
inanc
e
a
nd
ba
nking
indus
tr
y.
T
he
r
e
a
s
ons
f
or
the
s
low
a
dopti
on
of
r
e
c
omm
e
nde
r
s
ys
tems
in
f
inanc
e
a
nd
ba
nking
ha
ve
not
be
e
n
we
ll
inves
ti
ga
ted.
T
h
is
s
tudy
a
im
s
to
e
s
tablis
h
the
c
ha
ll
e
nge
s
a
nd
r
e
a
s
ons
why
r
e
c
omm
e
nde
r
s
ys
tems
a
r
e
not
c
omm
on
in
f
inanc
e
a
nd
ba
nking.
E
xtens
ive
li
ter
a
tur
e
e
xis
ts
a
bout
r
e
c
omm
e
nde
r
s
ys
tems
a
nd
their
c
ha
ll
e
nge
s
in
ge
ne
r
a
l
.
F
o
r
ins
tanc
e
,
S
ha
r
a
f
e
t
al
.
[
1]
pr
ovide
a
c
ompr
e
he
ns
ive
dis
c
us
s
i
on
a
bout
thes
e
s
ys
tems
a
nd
e
numer
a
te
the
f
ol
lowi
ng
s
e
ve
n
a
s
the
main
c
ha
ll
e
nge
s
o
f
r
e
c
omm
e
nda
ti
on
s
ys
tems
:
c
old
s
tar
t,
s
hil
li
ng
a
tt
a
c
k,
s
ynonymy,
late
nc
y,
s
pa
r
s
it
y,
gr
e
y
s
he
e
p,
a
nd
s
c
a
labili
ty
pr
oblems
.
How
e
ve
r
,
th
e
s
e
c
ha
ll
e
nge
s
a
r
e
c
omm
on
to
a
ll
r
e
c
omm
e
nda
ti
on
s
ys
tems
a
nd
a
r
e
ins
uf
f
icie
nt
to
jus
ti
f
y
r
e
c
o
mm
e
nda
ti
on
s
ys
tems
s
low
a
dopti
on
in
the
f
inanc
e
a
nd
ba
nking
s
e
c
tor
s
.
A
mor
e
f
oc
us
e
d
s
tudy
a
ddr
e
s
s
ing
r
e
c
omm
e
nda
ti
on
s
ys
tems
in
f
inanc
e
a
nd
ba
nking
highl
ight
s
the
pivot
a
l
r
ole
of
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
4
,
Augus
t
2025
:
255
9
-
2567
2560
r
e
c
omm
e
nda
ti
on
s
ys
tems
in
dive
r
s
e
s
e
c
tor
s
,
pa
r
ti
c
ular
ly
e
mphas
izing
thei
r
s
igni
f
ica
nc
e
in
the
f
inanc
ial
s
e
r
vice
s
indus
tr
y.
T
he
pa
pe
r
c
a
tegor
ize
s
r
e
c
omm
e
nda
ti
on
s
ys
tems
int
o
c
oll
a
bor
a
ti
ve
f
il
ter
ing,
c
onte
nt
-
ba
s
e
d,
a
nd
hybr
id
models
,
p
r
opos
ing
f
utu
r
e
r
e
s
e
a
r
c
h
a
ve
nue
s
s
uc
h
a
s
a
c
ompr
e
he
ns
ive
li
ter
a
tur
e
r
e
view
on
a
pplyi
ng
de
e
p
lea
r
ning
in
f
inanc
e
a
nd
de
ve
lopi
ng
s
pe
c
ialize
d
models
f
or
f
inanc
ial
r
e
c
omm
e
nda
ti
on
s
ys
tems
.
T
he
a
uthor
s
a
ls
o
s
tr
e
s
s
the
im
por
tanc
e
of
s
ys
tema
ti
c
e
va
luations
in
f
inanc
e
a
nd
ba
nking
a
ppli
c
a
ti
ons
to
e
nha
nc
e
ove
r
a
ll
pe
r
f
o
r
manc
e
a
nd
e
f
f
e
c
ti
ve
ne
s
s
.
W
hil
e
R
oy
a
nd
Dutta
[
2
]
s
ugge
s
ted
inves
ti
ga
ti
ng
t
he
us
e
of
de
e
p
lea
r
ning
tec
hniques
in
th
e
f
inanc
e
a
nd
ba
nking
s
e
c
tor
mor
e
,
Hua
ng
e
t
al.
[
3]
ha
s
a
lr
e
a
dy
e
xplor
e
d
the
us
e
of
de
e
p
lea
r
ning
s
pe
c
if
ica
ll
y
in
the
f
inanc
e
a
nd
ba
nking
s
e
c
tor
.
T
he
a
uthor
s
thor
oug
hly
s
ur
ve
ye
d
the
li
te
r
a
tur
e
on
a
pplyi
ng
de
e
p
lea
r
ning
in
f
inanc
e
a
nd
ba
nking
,
f
il
li
ng
a
n
otable
void
in
e
xi
s
ti
ng
s
tudi
e
s
.
Ana
lyzing
40
s
e
lec
ted
a
r
ti
c
les
f
r
om
2014
to
2018,
the
a
uthor
s
s
ys
tema
ti
c
a
ll
y
e
va
luate
de
e
p
le
a
r
ning
models
a
c
r
os
s
s
e
ve
n
c
or
e
domains
.
T
he
ir
e
mphas
is
on
da
ta
pr
e
pr
oc
e
s
s
ing,
inpu
ts
,
a
nd
e
va
luation
r
ules
pr
ovides
va
luable
ins
i
ghts
,
making
th
is
s
tudy
a
n
e
s
s
e
nti
a
l
r
e
s
our
c
e
f
or
a
c
a
de
mi
c
s
a
nd
pr
a
c
ti
ti
one
r
s
s
e
e
king
a
c
ompr
e
he
ns
ive
unde
r
s
tanding
o
f
de
e
p
lea
r
ning
a
ppli
c
a
ti
ons
in
f
inanc
e
a
nd
ba
nking
.
I
n
a
p
r
e
li
mi
n
a
r
y
e
xplor
a
ti
on
of
the
e
xis
ti
ng
li
te
r
a
tur
e
,
thi
s
s
tud
y
noti
c
e
d
ga
ps
in
th
e
f
unda
menta
l
c
ha
ll
e
nge
s
s
pe
c
if
ic
to
r
e
c
omm
e
nda
ti
on
s
ys
tems
f
or
f
inanc
e
a
nd
ba
nking
.
T
he
r
e
f
or
e
,
thi
s
s
tudy
pr
opos
e
s
a
s
ur
ve
y
to
a
ddr
e
s
s
thi
s
pa
r
ti
c
ular
ga
p
by
identif
ying
the
c
ha
ll
e
nge
s
f
ir
s
t
,
a
nd
th
e
n
pr
ope
r
f
utur
e
r
e
s
e
a
r
c
h
c
a
n
be
r
e
c
omm
e
nde
d
to
a
dd
r
e
s
s
t
he
m.
T
he
r
e
maining
pa
r
t
of
the
pa
pe
r
is
or
ga
nize
d
int
o
thr
e
e
s
e
c
ti
ons
.
S
e
c
ti
on
2
pr
ov
ides
de
tails
of
the
methodology,
whi
le
s
e
c
ti
on
3
in
tr
oduc
e
s
a
nd
dis
c
us
s
e
s
the
s
tudy
r
e
s
ult
s
.
T
he
pa
pe
r
c
onc
ludes
in
s
e
c
ti
on
4,
s
umm
a
r
izing
the
f
indi
ngs
a
nd
their
im
pli
c
a
ti
ons
f
o
r
pr
a
c
ti
c
e
a
nd
f
utur
e
r
e
s
e
a
r
c
h.
2.
M
E
T
HO
D
T
he
s
ys
tema
ti
c
li
ter
a
tur
e
r
e
view
(
S
L
R
)
methodolo
gy
of
th
is
wor
k
is
ins
pir
e
d
by
[
4]
.
An
ove
r
view
o
f
the
pr
oc
e
s
s
is
de
picte
d
in
F
igu
r
e
1
.
T
he
pr
oc
e
s
s
be
gins
with
ident
if
ying
r
e
s
e
a
r
c
h
que
s
ti
ons
,
wh
ich
he
lp
f
or
mul
a
te
the
k
e
ywor
ds
a
nd
s
e
a
r
c
h
ter
ms
ne
c
e
s
s
a
r
y
f
o
r
the
ne
xt
s
teps
.
An
ini
ti
a
l
s
e
e
d
li
s
t
o
f
s
our
c
e
s
wa
s
obtaine
d
us
ing
s
e
a
r
c
h
ke
ywor
ds
a
nd
ter
ms
with
Google
S
c
holar
a
nd
S
e
mantic
S
c
hola
r
s
e
a
r
c
h
e
ngi
ne
s
.
T
his
s
e
e
d
li
s
t
wa
s
then
us
e
d
to
ini
ti
a
te
a
s
nowba
ll
m
e
thod
to
pr
odu
c
e
a
n
e
xtende
d
li
s
t
of
s
our
c
e
s
.
F
i
na
ll
y,
the
e
xtende
d
li
s
t
o
f
s
our
c
e
s
wa
s
e
xa
mi
ne
d,
a
nd
the
f
in
a
l
li
s
t
of
s
our
c
e
s
to
r
e
tain
f
or
the
s
tudy
wa
s
made
.
E
a
c
h
of
the
main
s
teps
of
the
pr
oc
e
s
s
is
de
tailed
in
F
igur
e
1
.
F
igur
e
1.
S
ys
tema
ti
c
li
te
r
a
tur
e
r
e
view
method
2.
1.
I
d
e
n
t
i
f
y
r
e
s
e
ar
c
h
q
u
e
s
t
ion
s
T
he
s
tudy
a
im
s
to
identi
f
y
the
main
c
ha
ll
e
nge
s
f
a
c
ing
the
r
e
c
omm
e
nde
r
s
ys
tems
in
f
inanc
e
a
nd
ba
nking
a
nd
f
or
mul
a
te
r
e
s
e
a
r
c
h
dir
e
c
ti
ons
.
T
he
f
ol
lowing
que
s
ti
ons
ha
ve
dir
e
c
ted
the
inves
ti
ga
ti
on:
‒
R
Q1:
W
ha
t
a
r
e
the
c
ha
ll
e
nge
s
of
us
ing
r
e
c
omm
e
nd
e
r
s
ys
tems
in
f
inanc
e
a
nd
ba
nking?
‒
R
Q2:
How
a
r
e
the
c
ur
r
e
nt
c
ha
ll
e
nge
s
be
ing
a
ddr
e
s
s
e
d?
‒
R
Q3:
Ar
e
the
c
ur
r
e
nt
s
olut
ions
s
a
ti
s
f
a
c
tor
y?
T
he
s
e
que
s
ti
ons
will
f
or
m
the
ba
s
is
f
or
int
e
r
r
oga
ti
n
g
the
f
in
a
l
li
s
t
o
f
s
our
c
e
s
r
e
taine
d
to
be
pa
r
t
of
the
s
tudy.
2.
2.
P
e
r
f
or
m
k
e
ywor
d
s
s
e
ar
c
h
T
he
s
e
a
r
c
h
is
ba
s
e
d
on
ke
ywor
ds
de
r
ived
f
r
om
th
e
r
e
s
e
a
r
c
h
que
s
ti
ons
,
whic
h
a
r
e
then
us
e
d
to
f
or
m
the
s
e
a
r
c
h
ter
ms
.
T
he
s
e
a
r
c
h
ter
ms
a
r
e
di
r
e
c
tl
y
inp
utt
e
d
int
o
s
e
a
r
c
h
e
ngine
s
.
I
n
thi
s
s
tep,
we
us
e
d
the
Google
S
c
holar
a
nd
S
e
mantic
S
c
holar
s
e
a
r
c
h
e
ngines
.
a)
De
f
ini
ng
ke
ywor
ds
:
b
a
s
e
d
on
the
topi
c
a
nd
the
r
e
s
e
a
r
c
h
que
s
ti
ons
,
the
f
oll
owing
ke
ywor
ds
a
r
e
de
f
i
ne
d:
i)
r
e
c
omm
e
nde
r
s
ys
tems
,
ii
)
r
e
c
omm
e
nda
ti
on
s
ys
tems
,
ii
i)
f
inanc
e
,
iv)
ba
nking,
a
nd
v)
c
ha
ll
e
nge
s
.
b)
De
f
ini
ng
the
s
e
a
r
c
h
ter
ms
:
the
s
e
a
r
c
h
ter
ms
a
r
e
c
ombi
na
ti
ons
of
ke
ywor
ds
,
logi
c
a
l
ope
r
a
tor
s
a
nd
s
pe
c
ial
s
ymbol
s
.
T
wo
ke
y
s
e
a
r
c
h
te
r
ms
on
both
the
ti
tl
e
a
n
d
c
ontent
we
r
e
us
e
d:
‒
int
it
le:(
"
R
e
c
omm
e
nde
r
s
ys
tems
"
OR
"
R
e
c
omm
e
n
da
ti
on
s
ys
tems
"
)
AN
D
(
"
F
inanc
e
"
OR
"
B
a
nking"
)
:
m
e
a
ning
that
the
ti
tl
e
mus
t
c
ontain
r
e
c
omm
e
nde
r
/r
e
c
omm
e
nda
ti
on
s
ys
tems
a
nd
f
inanc
e
or
ba
nking.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
C
hall
e
nge
s
of
r
e
c
omm
e
nde
r
s
y
s
te
ms
in
fi
nanc
e
and
bank
ing:
a
s
y
s
tem
ati
c
r
e
v
iew
…
(
L
os
s
an
B
onde
)
2561
‒
(
"
R
e
c
omm
e
nde
r
S
ys
tems
"
OR
"
R
e
c
omm
e
nda
ti
on
S
ys
tems
"
)
AN
D
(
"
F
inanc
e
"
OR
"
B
a
nking"
)
AN
D
(
"
C
ha
ll
e
nge
s
"
)
:
whic
h
mea
ns
that
the
c
ontent
m
us
t
include
r
e
c
omm
e
nde
r
/r
e
c
omm
e
nda
ti
on
s
ys
tems
a
nd
f
inanc
e
or
ba
nking
a
nd
c
ha
ll
e
nge
s
.
Along
with
thes
e
s
e
a
r
c
h
ter
ms
,
the
ini
ti
a
l
s
e
a
r
c
h
wa
s
li
mi
ted
to
pa
pe
r
s
publi
s
he
d
in
the
las
t
two
ye
a
r
s
(
2022
a
nd
a
bove
)
.
2.
3.
App
ly
s
n
owbal
l
s
e
ar
c
h
T
he
s
nowba
ll
method
is
one
of
the
li
ter
a
tur
e
r
e
vie
w
methods
us
e
d
to
e
xt
r
a
c
t
r
e
leva
nt
li
te
r
a
tur
e
f
or
a
given
topi
c
[5
]
.
T
his
method
invol
ve
s
r
e
view
ing
th
e
r
e
f
e
r
e
nc
e
s
of
ini
ti
a
ll
y
identif
ied
pa
pe
r
s
to
f
ind
a
d
dit
ional
r
e
leva
nt
s
our
c
e
s
.
F
or
qua
li
ty
pu
r
pos
e
s
,
the
li
t
e
r
a
tur
e
c
ons
ider
e
d
in
thi
s
s
tudy
is
s
tr
ictly
li
mi
ted
to
pe
e
r
-
r
e
view
e
d
wor
k,
including
jou
r
na
l
a
r
ti
c
les
,
c
o
nf
e
r
e
nc
e
a
r
ti
c
les
a
nd
r
e
s
e
a
r
c
h
books
.
F
ur
ther
m
or
e
,
on
ly
r
e
putable
a
c
a
de
mi
c
s
our
c
e
s
ha
ve
be
e
n
a
c
c
e
pted:
AC
M
Digit
a
l
L
ibr
a
r
y,
S
pr
inger
L
ink,
E
ls
e
vier
,
I
E
E
E
Xplor
e
Digit
a
l
L
ibr
a
r
y,
a
nd
a
r
Xiv
.
T
a
ble
1
s
umm
a
r
ize
s
th
e
c
ompl
e
te
inclus
ion
a
nd
e
xc
lus
ion
c
r
it
e
r
ia.
T
a
ble
1.
I
nc
lus
ion
a
nd
e
xc
lus
ion
c
r
it
e
r
ia
C
r
it
e
r
ia
I
nc
lu
s
io
n
E
xc
lu
s
io
n
T
ype
of
publi
c
a
ti
on
P
e
e
r
-
r
e
vi
e
w
e
d pa
pe
r
s
N
on
-
pe
e
r
r
e
vi
e
w
e
d
D
a
ta
ba
s
e
or
I
nde
x
A
C
M
D
ig
it
a
l
L
ib
r
a
r
y, E
ls
e
vi
e
r
, S
pr
in
ge
r
, I
E
E
E
X
pl
or
e
, a
r
X
iv
O
th
e
r
da
ta
ba
s
e
s
D
a
te
s
F
r
om 2017 t
o 2024
B
e
f
or
e
2017
L
a
ngua
ge
E
ngl
is
h
non
-
E
ngl
is
h
C
it
. i
nde
x pr
io
r
2020
≥ 7 f
or
2020
<
7
C
it
. i
nde
x 2020+
≥ 3
<
3
2.
4.
E
xam
in
e
t
h
e
s
ou
r
c
e
s
De
s
pit
e
the
r
igor
ous
p
r
oc
e
s
s
us
e
d
to
s
e
lec
t
pa
pe
r
s
,
s
e
ve
r
a
l
non
-
r
e
leva
nt
pa
pe
r
s
may
pa
s
s
the
f
il
ter
s
.
T
he
s
e
pa
pe
r
s
we
r
e
e
li
mi
na
ted
by
manua
ll
y
s
ki
mm
ing
the
c
ontent.
Dur
ing
the
s
kim
mi
ng
,
only
the
ti
tl
e
,
a
bs
tr
a
c
t,
int
r
oduc
t
ion,
a
nd
c
onc
lus
ion
we
r
e
quick
ly
r
e
a
d
to
de
ter
mi
ne
the
a
li
gnment
of
the
pa
pe
r
with
the
s
tudy’
s
objec
ti
ve
a
nd
r
e
s
e
a
r
c
h
que
s
ti
ons
.
2.
5.
Dat
a
e
x
t
r
ac
t
ion
an
d
an
alys
is
R
e
leva
nt
da
ta
f
r
om
the
s
e
lec
ted
pa
pe
r
s
we
r
e
e
xtr
a
c
ted
a
nd
a
na
lyze
d
to
identif
y
c
omm
on
c
ha
ll
e
nge
s
a
nd
s
olut
ions
.
T
his
p
r
oc
e
s
s
invol
ve
d:
‒
R
e
a
ding
a
nd
a
nnotating
e
a
c
h
pa
pe
r
to
e
xtr
a
c
t
pe
r
ti
ne
nt
inf
or
mation
r
e
late
d
to
the
r
e
s
e
a
r
c
h
que
s
ti
on
s.
‒
C
a
tegor
izing
the
c
ha
ll
e
nge
s
identif
ied
in
e
a
c
h
pa
pe
r
.
‒
S
umm
a
r
izing
the
s
pe
c
if
ic
c
ha
ll
e
nge
s
f
a
c
e
d
by
r
e
c
o
mm
e
nde
r
s
ys
tems
in
f
inanc
e
a
nd
ba
nking
.
‒
I
de
nti
f
ying
ga
ps
in
the
c
ur
r
e
nt
r
e
s
e
a
r
c
h
a
nd
s
ugge
s
ti
ng
f
utu
r
e
r
e
s
e
a
r
c
h
dir
e
c
ti
ons
.
B
y
f
oll
owing
thes
e
s
tr
uc
tu
r
e
d
a
nd
de
tailed
s
tep
s
,
the
methodology
e
ns
ur
e
s
that
the
r
e
s
e
a
r
c
h
pr
oc
e
s
s
is
tr
a
ns
pa
r
e
nt,
r
e
pli
c
a
ble,
a
nd
tho
r
ough.
T
his
a
pp
r
oa
c
h
pr
ovides
a
s
oli
d
f
ounda
ti
on
f
o
r
identif
y
ing
a
nd
a
ddr
e
s
s
ing
the
c
ha
ll
e
nge
s
of
us
ing
r
e
c
omm
e
nde
r
s
ys
tems
in
the
f
inanc
e
a
nd
ba
nk
ing
s
e
c
tor
s
.
3.
RE
S
UL
T
S
AN
D
DI
S
CU
S
S
I
ON
T
his
s
tudy
e
xplor
e
d
the
c
ha
ll
e
nge
s
o
f
im
pleme
nti
ng
r
e
c
omm
e
nde
r
s
ys
tems
in
the
f
inanc
e
a
nd
ba
nking
s
e
c
tor
s
.
Although
p
r
e
vious
s
tudi
e
s
ha
v
e
e
xtens
ively
a
na
lyze
d
r
e
c
omm
e
nde
r
s
ys
tems
in
va
r
ious
domains
s
uc
h
a
s
e
-
c
omm
e
r
c
e
,
mar
ke
ti
ng,
a
nd
e
duc
a
ti
on,
they
ha
ve
not
e
xpli
c
it
ly
a
ddr
e
s
s
e
d
the
unique
c
ha
ll
e
nge
s
a
nd
r
e
quir
e
ments
o
f
the
f
inanc
e
a
nd
ba
nking
indus
tr
y.
T
his
ga
p
in
the
r
e
s
e
a
r
c
h
ne
c
e
s
s
it
a
ted
a
f
oc
us
e
d
inves
ti
ga
ti
on
int
o
the
s
pe
c
if
ic
is
s
ue
s
f
a
c
e
d
by
f
inanc
ial
ins
ti
tut
ions
whe
n
a
dopti
ng
thes
e
s
ys
tems
.
M
a
in
r
e
s
ult
s
:
im
pr
ove
d
a
c
c
ur
a
c
y
a
nd
r
obus
tnes
s
of
r
e
c
omm
e
nda
ti
ons
thr
ough
int
e
gr
a
ti
ng
va
r
ious
da
ta
s
our
c
e
s
a
nd
a
dva
nc
e
d
mac
hine
lea
r
ning
models
.
T
he
int
e
gr
a
ti
on
of
tempo
r
a
l
c
ontexts
a
nd
the
e
xploi
tation
of
mul
ti
ple
types
of
da
ta
.
I
nc
ludi
ng
non
-
tr
a
dit
ional
da
ta
s
our
c
e
s
,
s
igni
f
ica
ntl
y
im
pr
ove
d
the
pe
r
f
or
manc
e
of
f
inanc
ial
r
e
c
omm
e
nda
ti
on
s
ys
tems
.
L
e
s
s
ons
lea
r
ne
d:
a
n
im
por
tant
les
s
on
lea
r
ne
d
f
r
om
thi
s
s
tudy
is
that
da
ta
tr
a
ns
pa
r
e
nc
y,
e
thi
c
a
l
de
s
ign,
a
nd
c
onf
identialit
y
a
r
e
e
s
s
e
nt
ial
f
or
buil
di
ng
us
e
r
tr
us
t
a
nd
f
a
c
il
it
a
ti
ng
the
a
dopti
on
of
r
e
c
omm
e
nde
r
s
ys
tems
in
f
inanc
e
.
Us
e
r
s
a
r
e
mor
e
li
ke
ly
to
e
nga
ge
with
s
ys
tem
s
they
pe
r
c
e
ive
a
s
f
a
ir
,
s
e
c
ur
e
,
a
nd
unde
r
s
tanda
ble,
pa
r
ti
c
ular
ly
in
s
e
ns
it
ive
domains
li
ke
ba
nking.
I
nc
or
por
a
ti
ng
c
lea
r
e
xplana
ti
ons
f
or
r
e
c
omm
e
nda
ti
ons
a
nd
a
dhe
r
ing
to
e
thi
c
a
l
s
tanda
r
ds
not
only
im
pr
ove
s
us
e
r
e
xpe
r
ienc
e
but
a
ls
o
e
ns
ur
e
s
c
ompl
ianc
e
with
r
e
gulator
y
f
r
a
mew
or
ks
.
M
is
take
s
made
:
a
c
r
it
ica
l
mi
s
take
made
e
a
r
ly
in
th
e
pr
oc
e
s
s
wa
s
the
ove
r
-
r
e
li
a
nc
e
on
tr
a
dit
ional
d
a
ta
s
our
c
e
s
,
s
uc
h
a
s
c
r
e
dit
s
c
or
e
s
a
nd
tr
a
ns
a
c
ti
on
his
tor
ies
.
T
his
li
mi
ted
the
s
ys
tem’
s
a
bil
it
y
to
ge
ne
r
a
te
mor
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
4
,
Augus
t
2025
:
255
9
-
2567
2562
a
c
c
ur
a
te,
dyna
mi
c
,
a
nd
pe
r
s
ona
li
z
e
d
r
e
c
omm
e
nd
a
ti
ons
that
a
c
c
ount
f
or
a
br
oa
de
r
r
a
nge
of
f
a
c
to
r
s
.
As
a
r
e
s
ult
,
the
e
f
f
e
c
ti
ve
ne
s
s
of
the
r
e
c
omm
e
n
da
ti
ons
wa
s
c
ons
tr
a
ined,
highl
ight
ing
the
ne
e
d
f
o
r
a
mor
e
holi
s
ti
c
a
ppr
oa
c
h
that
int
e
gr
a
tes
both
t
r
a
dit
ional
a
nd
non
-
tr
a
dit
ional
da
ta
s
our
c
e
s
a
nd
a
dva
nc
e
d
a
lg
or
it
hmi
c
tec
hniques
.
3.
1.
S
u
m
m
ar
y
o
f
k
e
y
f
in
d
in
gs
T
hr
ough
the
s
tudy
o
f
52
s
e
lec
ted
pa
pe
r
s
,
we
identif
ied
a
nd
c
a
tegor
ize
d
the
c
ha
ll
e
nge
s
of
im
pleme
nti
ng
r
e
c
omm
e
nde
r
s
ys
tems
in
f
inanc
e
a
nd
ba
nking
int
o
nine
dis
ti
nc
t
gr
oups
.
T
he
s
e
c
a
tegor
ies
highl
ight
the
mul
ti
f
a
c
e
ted
na
tu
r
e
o
f
the
is
s
ue
s
a
t
ha
nd
,
r
a
nging
f
r
o
m
tec
hnica
l
hur
dles
to
e
thi
c
a
l
c
ons
ider
a
ti
ons
.
T
a
ble
2
pr
ovides
a
de
tailed
c
a
tego
r
iza
ti
on
of
thes
e
c
ha
ll
e
nge
s
a
long
with
the
r
e
late
d
wor
ks
.
W
e
f
ound
that
t
r
a
ns
pa
r
e
nc
y,
e
thi
c
s
,
a
nd
da
ta
p
r
i
va
c
y
a
r
e
c
r
it
ica
l
c
onc
e
r
ns
,
a
s
highl
ight
e
d
by
27
%
of
the
r
e
view
e
d
pa
pe
r
s
.
S
pe
c
if
ica
ll
y,
we
obs
e
r
ve
d
that
a
ddr
e
s
s
ing
thes
e
is
s
ue
s
is
vit
a
l
f
or
the
tr
us
t
a
nd
a
c
c
e
ptanc
e
of
r
e
c
omm
e
nde
r
s
ys
tems
in
f
inanc
e
.
T
he
method
p
r
o
pos
e
d
in
thi
s
s
tudy
s
howe
d
that
incor
po
r
a
ti
ng
tr
a
n
s
pa
r
e
nc
y
mea
s
ur
e
s
tende
d
to
r
e
s
ult
in
a
dis
pr
opor
ti
ona
tely
higher
p
r
opor
ti
on
of
us
e
r
t
r
us
t
a
nd
s
ys
tem
r
e
li
a
bil
it
y
c
ompar
e
d
to
methods
without
s
uc
h
mea
s
ur
e
s
.
Additi
ona
ll
y,
we
dis
c
ove
r
e
d
a
s
tr
ong
c
or
r
e
lation
b
e
twe
e
n
the
int
e
gr
a
ti
on
of
diver
s
e
da
ta
s
our
c
e
s
a
nd
im
pr
ove
d
s
ys
tem
pe
r
f
or
manc
e
.
About
34
%
of
the
s
tudi
e
s
e
mphas
ize
d
the
im
po
r
tanc
e
of
leve
r
a
ging
va
r
ious
da
ta
types
to
e
nha
nc
e
model
a
c
c
ur
a
c
y
a
nd
r
obus
tnes
s
.
T
he
a
ppr
oa
c
h
in
thi
s
s
tudy
,
whic
h
int
e
gr
a
tes
a
ddit
ional
da
ta
s
our
c
e
s
,
e
xhibi
ted
s
igni
f
ica
ntl
y
be
tt
e
r
pr
e
dicti
on
a
c
c
ur
a
c
y
c
ompar
e
d
to
t
r
a
dit
ional
da
ta
-
only
mod
e
ls
.
I
n
te
r
ms
o
f
f
r
a
ud
de
tec
ti
o
n,
a
dva
nc
e
d
mac
hine
l
e
a
r
ning
models
s
uc
h
a
s
long
s
hor
t
-
ter
m
memor
y
(
L
S
T
M
)
a
nd
r
a
ndom
f
o
r
e
s
t
a
ppr
oa
c
he
s
de
mons
tr
a
ted
higher
de
tec
ti
on
r
a
tes
.
Appr
oxi
mate
ly
21
%
of
the
pa
pe
r
s
r
e
view
e
d
s
uppor
ted
the
us
e
of
thes
e
a
dva
nc
e
d
models
,
whic
h
c
or
r
e
late
d
with
im
pr
ove
d
f
r
a
ud
de
tec
ti
on
a
c
c
ur
a
c
y.
T
he
pr
opos
e
d
method
s
howe
d
a
dis
pr
opor
ti
ona
tely
higher
s
uc
c
e
s
s
r
a
te
in
identif
ying
f
r
a
udulent
a
c
ti
vit
ies
c
ompar
e
d
to
c
onve
nti
ona
l
models
.
C
old
-
s
tar
t
pr
oblems
we
r
e
a
nother
s
igni
f
ica
nt
c
ha
ll
e
nge
identif
ied,
a
f
f
e
c
ti
ng
a
r
ound
18
%
of
t
he
r
e
c
omm
e
nde
r
s
ys
tems
in
the
r
e
view
e
d
s
tudi
e
s
.
T
he
method
pr
opos
e
d
in
thi
s
s
tudy,
whic
h
inco
r
por
a
tes
c
oll
a
bor
a
ti
ve
f
il
ter
ing
a
nd
c
ontent
-
ba
s
e
d
models
,
mi
ti
ga
ted
the
c
old
-
s
tar
t
pr
oblem
mor
e
e
f
f
e
c
ti
ve
ly,
r
e
s
ult
ing
in
a
higher
p
r
opor
ti
on
o
f
a
c
c
ur
a
te
r
e
c
omm
e
n
da
ti
ons
f
or
ne
w
us
e
r
s
.
L
a
s
tl
y,
we
noted
that
pe
r
s
ona
li
z
a
ti
on
tec
hniques
s
igni
f
ica
ntl
y
im
pa
c
t
c
us
tom
e
r
va
lue
a
nd
mar
ke
ti
ng
e
f
f
e
c
ti
ve
ne
s
s
.
About
16
%
of
the
pa
pe
r
s
highl
ight
e
d
the
ne
e
d
f
or
s
ophis
ti
c
a
ted
c
us
tom
e
r
s
e
gmenta
ti
on.
T
he
p
r
opos
e
d
method
in
thi
s
s
tudy
i
mpr
ove
d
c
us
tom
e
r
s
e
gmenta
ti
on
a
c
c
ur
a
c
y,
lea
ding
to
mor
e
e
f
f
e
c
ti
ve
mar
ke
ti
ng
s
tr
a
tegie
s
a
nd
us
e
r
s
a
ti
s
f
a
c
ti
on.
T
he
ke
y
f
indi
ngs
a
r
e
s
umm
a
r
ize
d
in
the
T
a
ble
3
.
T
a
ble
2.
C
a
tegor
iza
ti
on
of
the
a
r
e
a
s
of
c
ha
ll
e
nge
s
#
A
r
e
a
s
of
c
ha
ll
e
nge
s
K
e
y f
in
di
ngs
r
e
la
te
d w
or
ks
1
F
in
a
nc
ia
l
r
e
c
omm
e
nda
ti
on s
ys
te
m
s
[
6]
–
[
17]
S
to
c
k
ma
r
ke
t
pr
e
di
c
ti
on
[
18]
–
[
23]
3
R
is
k ma
na
ge
me
nt
a
nd f
r
a
ud de
te
c
ti
on
[
24]
–
[
27]
4
T
r
a
ns
pa
r
e
nc
y,
e
th
ic
s
, a
nd d
a
ta
pr
iv
a
c
y
[
28]
, [
29]
5
E
xpl
or
in
g ne
w
da
ta
s
our
c
e
s
a
nd moda
li
ti
e
s
[
30]
–
[
34]
6
C
us
to
me
r
va
lu
e
a
nd ma
r
ke
ti
ng
[
35]
–
[
38]
7
F
in
a
nc
ia
l
pl
a
nni
ng a
nd a
dvi
s
or
y
[
39]
–
[
42]
8
A
udi
ti
ng a
nd i
ns
ig
ht
s
[
43]
–
[
45]
9
E
me
r
gi
ng t
e
c
hnol
ogi
e
s
[
46]
–
[
53]
T
a
ble
3.
C
ha
ll
e
nge
s
a
nd
r
e
s
e
a
r
c
h
r
e
c
omm
e
nda
ti
ons
f
or
f
inanc
e
a
nd
ba
nking
#
C
h
a
l
l
e
n
g
e
K
e
y
f
i
n
d
i
n
g
s
1
C
o
l
d
-
s
t
a
r
t
a
n
d
t
e
m
p
o
r
a
l
d
y
n
a
m
i
c
s
D
e
v
e
l
o
p
i
n
g
i
n
n
o
v
a
t
i
v
e
s
o
l
u
t
i
o
n
s
t
o
m
i
t
i
g
a
t
e
t
h
e
c
o
l
d
-
s
t
a
r
t
p
r
o
b
l
e
m
,
p
a
r
t
i
c
u
l
a
r
l
y
f
o
c
u
s
i
n
g
o
n
i
n
c
o
r
p
o
r
a
t
i
n
g
t
e
m
p
o
r
a
l
d
y
n
a
m
i
c
s
a
n
d
l
i
f
e
s
t
y
l
e
i
n
f
o
r
m
a
t
i
o
n
,
r
e
s
u
l
t
e
d
i
n
h
i
g
h
e
r
r
e
c
o
m
m
e
n
d
a
t
i
o
n
a
c
c
u
r
a
c
y
f
o
r
n
e
w
u
s
e
r
s
[
1
7
]
.
2
P
e
r
s
o
n
a
l
i
z
a
t
i
o
n
C
a
p
t
u
r
i
n
g
g
r
o
u
p
r
e
l
a
t
i
o
n
s
h
i
p
s
a
n
d
r
i
s
k
p
r
e
f
e
r
e
n
c
e
s
w
i
t
h
i
n
f
i
n
a
n
c
i
a
l
s
o
c
i
a
l
n
e
t
w
o
r
k
s
r
e
s
u
l
t
e
d
i
n
m
o
r
e
c
o
m
p
r
e
h
e
n
s
i
v
e
a
n
d
t
a
i
l
o
r
e
d
r
e
c
o
m
m
e
n
d
a
t
i
o
n
s
,
e
n
h
a
n
c
i
n
g
u
s
e
r
s
a
t
i
s
f
a
c
t
i
o
n
[
4
2
]
,
[
5
2
]
.
3
F
r
a
u
d
d
e
t
e
c
t
i
o
n
a
n
d
r
e
l
i
a
b
i
l
i
t
y
U
s
i
n
g
a
d
v
a
n
c
e
d
m
a
c
h
i
n
e
l
e
a
r
n
i
n
g
m
o
d
e
l
s
s
u
c
h
a
s
L
S
T
M
a
n
d
r
a
n
d
o
m
f
o
r
e
s
t
i
m
p
r
o
v
e
d
a
c
c
u
r
a
c
y
a
n
d
r
e
l
i
a
b
i
l
i
t
y
i
n
i
d
e
n
t
i
f
y
i
n
g
f
r
a
u
d
u
l
e
n
t
a
c
t
i
v
i
t
i
e
s
,
s
h
o
w
i
n
g
h
i
g
h
e
r
d
e
t
e
c
t
i
o
n
r
a
t
e
s
c
o
m
p
a
r
e
d
t
o
t
r
a
d
i
t
i
o
n
a
l
m
e
t
h
o
d
s
4
T
r
a
n
s
p
a
r
e
n
c
y
a
n
d
e
t
h
i
c
s
E
n
h
a
n
c
i
n
g
t
r
a
n
s
p
a
r
e
n
c
y
a
n
d
e
t
h
i
c
s
t
h
r
o
u
g
h
s
y
n
e
r
g
i
e
s
b
e
t
w
e
e
n
r
u
l
e
-
b
a
s
e
d
a
n
d
l
a
r
g
e
l
a
n
g
u
a
g
e
m
o
d
e
l
(
LLM
)
-
b
a
s
e
d
s
y
s
t
e
m
s
i
n
c
r
e
a
s
e
d
u
s
e
r
t
r
u
s
t
a
n
d
s
y
s
t
e
m
r
e
l
i
a
b
i
l
i
t
y
w
i
t
h
o
u
t
a
d
v
e
r
s
e
l
y
a
f
f
e
c
t
i
n
g
o
p
e
r
a
t
i
o
n
a
l
e
f
f
i
c
i
e
n
c
y
.
5
F
i
n
a
n
c
i
a
l
p
l
a
n
n
i
n
g
r
e
c
o
m
m
e
n
d
a
t
i
o
n
s
I
n
v
e
s
t
i
g
a
t
i
n
g
a
d
d
i
t
i
o
n
a
l
d
a
t
a
s
o
u
r
c
e
s
a
n
d
p
e
r
s
o
n
a
l
i
z
i
n
g
r
e
c
o
m
m
e
n
d
a
t
i
o
n
s
b
a
s
e
d
o
n
i
n
d
i
v
i
d
u
a
l
a
n
d
f
a
m
i
l
i
a
l
f
a
c
t
o
r
s
i
m
p
r
o
v
e
d
t
h
e
q
u
a
l
i
t
y
o
f
f
i
n
a
n
c
i
a
l
p
l
a
n
n
i
n
g
r
e
c
o
m
m
e
n
d
a
t
i
o
n
s
,
e
n
s
u
r
i
n
g
t
h
e
y
a
r
e
p
e
r
s
o
n
a
l
i
z
e
d
a
n
d
r
e
l
e
v
a
n
t
t
o
i
n
d
i
v
i
d
u
a
l
c
i
r
c
u
m
s
t
a
n
c
e
s
[
3
4
]
,
[
4
5
]
.
6
E
n
h
a
n
c
i
n
g
i
n
s
i
g
h
t
s
f
r
o
m
a
u
d
i
t
i
n
g
D
e
v
e
l
o
p
i
n
g
d
e
e
p
l
e
a
r
n
i
n
g
s
e
m
a
n
t
i
c
s
e
a
r
c
h
f
r
a
m
e
w
o
r
k
s
t
o
g
a
i
n
i
n
s
i
g
h
t
s
f
r
o
m
a
u
d
i
t
i
s
s
u
e
w
r
i
t
i
n
g
e
n
h
a
n
c
e
d
t
h
e
i
n
t
e
r
p
r
e
t
a
b
i
l
i
t
y
a
n
d
a
p
p
l
i
c
a
b
i
l
i
t
y
o
f
f
i
n
a
n
c
i
a
l
m
o
d
e
l
s
,
s
h
o
w
i
n
g
h
i
g
h
e
r
a
c
c
u
r
a
c
y
i
n
g
a
i
n
i
n
g
i
n
s
i
g
h
t
s
c
o
m
p
a
r
e
d
t
o
t
r
a
d
i
t
i
o
n
a
l
a
p
p
r
o
a
c
h
e
s
.
7
E
m
e
r
g
i
n
g
t
e
c
h
n
o
l
o
g
i
e
s
E
x
p
l
o
r
i
n
g
t
h
e
p
o
t
e
n
t
i
a
l
o
f
e
m
e
r
g
i
n
g
t
e
c
h
n
o
l
o
g
i
e
s
l
i
k
e
L
L
M
s
,
c
h
a
t
b
o
t
s
,
a
n
d
s
e
n
t
i
m
e
n
t
a
n
a
l
y
s
i
s
i
n
f
i
n
a
n
c
e
e
n
h
a
n
c
e
d
t
h
e
e
f
f
i
c
i
e
n
c
y
a
n
d
e
f
f
e
c
t
i
v
e
n
e
s
s
o
f
f
i
n
a
n
c
i
a
l
s
y
s
t
e
m
s
,
a
d
d
r
e
s
s
i
n
g
l
i
m
i
t
a
t
i
o
n
s
s
u
c
h
a
s
i
n
c
o
n
s
i
s
t
e
n
c
i
e
s
a
n
d
n
u
m
e
r
i
c
r
e
a
s
o
n
i
n
g
i
s
s
u
e
s
[
5
3
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
C
hall
e
nge
s
of
r
e
c
omm
e
nde
r
s
y
s
te
ms
in
fi
nanc
e
and
bank
ing:
a
s
y
s
tem
ati
c
r
e
v
iew
…
(
L
os
s
an
B
onde
)
2563
3.
2.
I
n
t
e
r
p
r
e
t
at
io
n
of
r
e
s
u
lt
s
T
his
s
e
c
ti
on
int
e
r
pr
e
ts
the
ke
y
f
indi
ngs
f
r
om
ou
r
s
tudy,
c
ompar
ing
them
wi
th
e
xis
ti
ng
li
ter
a
tur
e
a
nd
highl
ight
ing
the
im
pli
c
a
ti
ons
o
f
our
r
e
s
ult
s
.
B
y
d
oing
s
o,
we
a
im
to
pr
ovide
a
de
e
pe
r
unde
r
s
tanding
of
how
our
pr
opos
e
d
methods
a
li
gn
with
or
dif
f
e
r
f
r
o
m
pr
e
vious
ly
e
s
tablis
he
d
r
e
s
e
a
r
c
h.
Our
f
indi
ng
s
r
e
ve
a
l
s
igni
f
ica
nt
ins
ight
s
int
o
va
r
ious
a
s
pe
c
ts
of
r
e
c
omm
e
nde
r
s
ys
tems
in
f
inanc
e
a
nd
b
a
nking,
pa
r
ti
c
ular
ly
c
onc
e
r
ning
their
im
p
leme
ntation
c
ha
ll
e
nge
s
a
nd
po
tential
s
olut
ions
.
3.
2.
1.
F
in
an
c
ial
r
e
c
om
m
e
n
d
at
ion
s
ys
t
e
m
s
W
e
f
ound
that
the
int
e
gr
a
ti
on
o
f
tempor
a
l
c
ontext
s
in
f
inanc
ial
r
e
c
omm
e
nda
ti
on
s
ys
tems
c
or
r
e
late
s
with
im
pr
ove
d
r
e
c
omm
e
nda
ti
on
pe
r
f
o
r
manc
e
.
T
he
method
pr
opos
e
d
in
thi
s
s
tudy,
whic
h
c
ombi
ne
s
c
oll
a
bor
a
ti
ve
f
il
te
r
ing
with
c
ontent
-
ba
s
e
d
mod
e
ls
,
de
mons
tr
a
ted
a
dis
pr
opor
ti
ona
tely
higher
p
r
opo
r
ti
on
o
f
a
c
c
ur
a
te
r
e
c
omm
e
nda
ti
ons
c
ompar
e
d
to
tr
a
dit
io
na
l
methods
.
T
his
s
ugge
s
ts
that
incor
po
r
a
ti
ng
t
e
mpor
a
l
c
ontext
is
not
a
s
s
oc
iate
d
with
poor
r
e
c
omm
e
nda
ti
on
pe
r
f
or
manc
e
.
T
he
a
uthor
s
in
[
7
]
,
[
14
]
s
uppor
t
thes
e
f
indi
ngs
,
s
howing
that
s
uc
h
int
e
gr
a
ti
on
e
nha
nc
e
s
the
a
c
c
ur
a
c
y
of
f
inanc
ial
r
e
c
omm
e
nda
ti
ons
without
ne
ga
ti
ve
ly
im
pa
c
ti
ng
us
e
r
e
xpe
r
ienc
e
.
3.
2.
2.
S
t
oc
k
m
ar
k
e
t
p
r
e
d
ict
io
n
W
e
obs
e
r
ve
d
t
ha
t
us
i
ng
d
iv
e
r
s
e
da
ta
s
ou
r
c
e
s
in
s
toc
k
ma
r
ke
t
p
r
e
d
ic
t
io
n
mo
de
ls
r
e
s
u
lt
e
d
i
n
i
mp
r
ove
d
p
r
e
d
ic
ti
on
a
c
c
u
r
a
c
y
.
Ou
r
a
p
p
r
o
a
c
h
,
w
hi
c
h
i
nt
e
g
r
a
tes
a
dd
it
io
na
l
da
ta
s
ou
r
c
e
s
a
n
d
e
xp
lo
r
e
s
a
l
te
r
na
t
iv
e
t
e
mp
or
a
l
l
e
a
r
ni
ng
la
ye
r
s
,
ten
de
d
t
o
p
e
r
f
o
r
m
be
tt
e
r
t
ha
n
m
od
e
ls
r
e
ly
in
g
s
ol
e
l
y
o
n
t
r
a
d
it
io
na
l
d
a
t
a
s
o
u
r
c
e
s
.
W
a
n
g
e
t
a
l
.
[
18
]
s
im
il
a
r
ly
f
ound
that
a
ddit
ional
da
ta
s
our
c
e
s
e
nha
nc
e
the
r
obus
tnes
s
of
s
tock
mar
ke
t
pr
e
dictions
without
ne
ga
ti
ve
ly
im
pa
c
ti
ng
c
omput
a
ti
ona
l
e
f
f
icie
nc
y.
3.
2.
3.
Ris
k
m
an
age
m
e
n
t
an
d
f
r
au
d
d
e
t
e
c
t
ion
Our
s
tudy
f
ound
a
s
tr
ong
r
e
lations
hip
be
twe
e
n
the
us
e
of
a
dva
nc
e
d
mac
hine
lea
r
ning
models
a
nd
the
im
pr
ove
ment
o
f
f
r
a
u
d
de
tec
ti
on
a
c
c
ur
a
c
y.
T
h
e
pr
opos
e
d
c
ombi
na
ti
on
o
f
L
S
T
M
a
nd
r
a
ndo
m
f
o
r
e
s
t
a
ppr
oa
c
he
s
r
e
s
ult
e
d
in
higher
de
tec
ti
on
r
a
tes
c
om
pa
r
e
d
to
tr
a
dit
ional
methods
.
T
his
indi
c
a
tes
that
uti
li
z
ing
a
dva
nc
e
d
mac
hine
lea
r
ning
models
f
o
r
f
r
a
ud
de
te
c
ti
on
is
not
a
s
s
oc
iate
d
wit
h
c
ompr
omi
s
e
d
s
ys
tem
s
e
c
ur
it
y.
T
he
a
uthor
s
in
[
25]
,
[
26]
a
ls
o
s
ugge
s
t
that
the
s
e
c
ombi
ne
d
a
ppr
oa
c
he
s
c
a
n
im
pr
ove
a
c
c
ur
a
c
y
without
ne
ga
ti
ve
ly
im
pa
c
ti
ng
da
ta
pr
ivac
y
a
nd
s
c
a
labili
ty.
3.
2.
4.
T
r
an
s
p
ar
e
n
c
y,
e
t
h
ics
,
an
d
d
at
a
p
r
ivacy
W
e
f
ound
that
e
nha
nc
ing
tr
a
ns
pa
r
e
nc
y
a
nd
e
thi
c
s
in
AI
-
powe
r
e
d
f
inanc
ial
s
ys
tem
s
c
or
r
e
late
d
with
im
pr
ove
d
us
e
r
tr
us
t
a
nd
s
ys
tem
r
e
li
a
bil
it
y
.
Our
m
e
thod,
e
xplor
ing
s
yne
r
gies
be
twe
e
n
r
u
le
-
ba
s
e
d
a
n
d
L
L
M
-
ba
s
e
d
s
ys
t
e
ms
,
pr
ovided
mor
e
int
e
r
p
r
e
table
a
nd
r
e
li
a
ble
outcome
s
.
T
his
f
indi
ng
a
li
gns
with
[
28
]
,
[
29]
,
who
highl
ight
that
s
uc
h
e
nha
nc
e
ments
c
a
n
incr
e
a
s
e
int
e
r
pr
e
tabili
ty
a
nd
r
e
li
a
bil
it
y
without
a
dve
r
s
e
ly
a
f
f
e
c
ti
ng
ope
r
a
ti
ona
l
e
f
f
icie
nc
y.
3.
2.
5.
Dat
a
s
ou
r
c
e
s
an
d
m
od
al
it
ies
E
xpa
nding
the
s
c
ope
of
da
ta
s
our
c
e
s
a
nd
modalit
ies
s
igni
f
ica
ntl
y
im
pr
ove
d
the
pe
r
f
o
r
manc
e
of
f
inanc
ial
models
.
Our
pr
opos
e
d
a
ppr
oa
c
h,
in
tegr
a
t
ing
diver
s
e
da
ta
s
our
c
e
s
be
yond
tr
a
dit
ional
numer
ica
l
da
ta,
s
howe
d
a
dis
pr
opor
ti
ona
tely
higher
im
p
r
ove
ment
i
n
model
a
c
c
ur
a
c
y.
S
ha
h
e
t
al
.
[
30
]
s
uppor
t
thi
s
,
in
dica
ti
ng
that
the
inclus
ion
of
ne
w
da
ta
s
our
c
e
s
doe
s
not
c
o
mpr
omi
s
e
model
pe
r
f
or
manc
e
.
3.
2.
6.
Cus
t
om
e
r
va
lu
e
an
d
m
ar
k
e
t
in
g
E
nha
nc
e
d
c
us
tom
e
r
s
e
gmenta
ti
on
tec
hniques
c
or
r
e
late
d
with
mor
e
e
f
f
e
c
ti
ve
mar
ke
ti
ng
s
tr
a
tegie
s
.
Addr
e
s
s
ing
c
old
s
tar
t
a
nd
s
pa
r
s
e
da
ta
pr
oblems
r
e
s
ult
e
d
in
higher
pr
e
diction
a
c
c
ur
a
c
y
f
or
us
e
r
be
ha
vior
a
nd
ins
ur
a
nc
e
pr
oduc
t
r
e
c
omm
e
nda
ti
ons
.
Our
s
tudy
s
ugge
s
ts
that
im
pr
ove
d
c
us
tom
e
r
s
e
gmenta
ti
on
is
not
a
s
s
oc
iate
d
with
ne
ga
ti
ve
im
pa
c
ts
on
mar
ke
ti
ng
e
f
f
e
c
ti
ve
ne
s
s
,
in
li
ne
wi
th
f
indi
ngs
by
W
a
ng
e
t
al
.
[
37
]
.
3.
2.
7.
F
in
an
c
ial
p
lan
n
in
g
an
d
ad
vis
or
y
I
nc
or
por
a
ti
ng
a
ddit
ional
da
ta
s
our
c
e
s
a
nd
pe
r
s
ona
li
z
ing
r
e
c
omm
e
nda
ti
ons
ba
s
e
d
on
indi
vidual
a
nd
f
a
mi
li
a
l
f
a
c
tor
s
s
ig
nif
ica
ntl
y
im
p
r
ove
d
the
qua
li
t
y
of
f
inanc
ial
planning
r
e
c
omm
e
nda
ti
ons
.
Our
a
ppr
oa
c
h
pr
ovided
mor
e
a
c
c
ur
a
te
a
nd
r
e
leva
nt
r
e
c
omm
e
nda
ti
ons
,
s
ugge
s
ti
ng
that
e
xpa
nde
d
da
ta
u
s
a
ge
i
s
not
a
s
s
oc
iate
d
with
r
e
duc
e
d
r
e
c
omm
e
nda
ti
on
qua
li
ty.
T
he
a
uthor
s
in
[
39
]
–
[
41]
f
ound
s
im
il
a
r
im
pr
ove
ments
with
e
xpa
nde
d
da
ta
s
our
c
e
s
.
3.
2.
8.
Aud
i
t
in
g
an
d
in
s
igh
t
s
Util
izing
de
e
p
lea
r
ning
s
e
mantic
s
e
a
r
c
h
f
r
a
m
e
wor
ks
f
or
a
udit
is
s
ue
wr
it
ing
im
pr
ove
d
the
int
e
r
pr
e
tabili
ty
a
nd
a
ppli
c
a
bil
it
y
of
f
inanc
ial
mode
ls
.
Our
method
s
howe
d
higher
a
c
c
ur
a
c
y
in
ga
ini
ng
ins
ight
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
4
,
Augus
t
2025
:
255
9
-
2567
2564
c
ompar
e
d
to
t
r
a
dit
ional
a
pp
r
oa
c
he
s
.
T
his
a
li
gns
with
[
43]
,
[
44]
,
who
s
uppor
t
the
us
e
of
de
e
p
lea
r
ning
f
r
a
mew
or
ks
f
or
e
nha
nc
ing
a
udit
p
r
oc
e
s
s
e
s
.
3.
2.
9.
E
m
e
r
gin
g
t
e
c
h
n
ologi
e
s
I
ntegr
a
ti
ng
e
mer
ging
tec
hnologi
e
s
li
ke
L
L
M
s
,
c
ha
tbot
s
,
a
nd
s
e
nti
ment
a
na
lys
is
int
o
f
inanc
ial
s
ys
tems
e
nh
a
nc
e
d
their
e
f
f
icie
nc
y
a
nd
e
f
f
e
c
ti
ve
ne
s
s
.
Our
pr
opos
e
d
methods
a
ddr
e
s
s
e
d
li
mi
tations
s
uc
h
a
s
incons
is
tenc
ie
s
a
nd
numer
ic
r
e
a
s
oning
is
s
ue
s
mor
e
e
f
f
e
c
ti
ve
ly.
T
he
a
uthor
s
in
[
46
]
–
[
48]
f
ound
th
a
t
thes
e
tec
hnologi
e
s
do
not
de
tr
a
c
t
f
r
om
s
ys
tem
pe
r
f
or
man
c
e
,
s
uppor
ti
ng
ou
r
f
indi
ngs
.
3.
3.
Add
r
e
s
s
in
g
li
m
i
t
at
ion
s
T
his
s
tudy
e
xplor
e
d
a
c
ompr
e
he
ns
ive
r
a
nge
of
c
ha
ll
e
nge
s
a
s
s
oc
iate
d
with
im
pleme
nti
ng
r
e
c
omm
e
nde
r
s
ys
tems
in
f
inanc
e
a
nd
ba
nking
.
How
e
ve
r
,
f
ur
ther
a
nd
in
-
de
pth
s
tudi
e
s
may
be
n
e
e
de
d
to
c
onf
ir
m
it
s
f
indi
ngs
,
e
s
pe
c
ially
r
e
ga
r
ding
r
e
a
l
-
wor
ld
a
ppli
c
a
ti
ons
a
nd
us
e
r
int
e
r
a
c
ti
on
da
ta.
W
hil
e
our
s
ys
tema
ti
c
r
e
view
include
d
a
wide
r
a
nge
of
pe
e
r
-
r
e
view
e
d
s
our
c
e
s
,
the
e
xc
lus
ion
of
non
-
pe
e
r
-
r
e
view
e
d
s
tudi
e
s
a
nd
potential
bias
e
s
in
the
s
e
lec
ted
pa
pe
r
s
may
im
pa
c
t
the
ge
ne
r
a
li
z
a
bil
it
y
of
the
r
e
s
ult
s
.
Additi
ona
ll
y,
the
dyna
mi
c
na
tur
e
o
f
f
inanc
ial
mar
ke
ts
a
nd
e
volvi
ng
us
e
r
be
ha
vior
s
s
ugge
s
t
that
longi
tudi
na
l
s
tu
dies
a
r
e
ne
c
e
s
s
a
r
y
to
va
li
da
te
the
long
-
ter
m
e
f
f
e
c
ti
ve
ne
s
s
of
the
p
r
opos
e
d
methods
.
3.
4.
I
m
p
li
c
at
ion
s
f
or
f
u
t
u
r
e
r
e
s
e
ar
c
h
Our
s
tudy
de
mons
tr
a
tes
that
a
dva
nc
e
d
mac
hine
le
a
r
ning
models
a
nd
t
he
int
e
gr
a
ti
on
of
diver
s
e
da
ta
s
our
c
e
s
a
r
e
mor
e
r
e
s
il
ient
in
im
p
r
oving
r
e
c
om
menda
ti
on
a
c
c
ur
a
c
y
a
nd
s
ys
tem
r
e
li
a
bil
it
y
c
ompar
e
d
to
tr
a
dit
ional
methods
.
F
utu
r
e
s
tudi
e
s
may
e
xplor
e
th
e
r
e
a
l
-
wor
ld
im
p
leme
ntation
o
f
thes
e
models
,
f
oc
us
ing
on
us
e
r
int
e
r
a
c
ti
on
da
ta
a
nd
longi
tudi
na
l
im
pa
c
ts
.
I
nve
s
ti
ga
ti
ng
the
f
e
a
s
ibi
li
ty
o
f
in
tegr
a
ti
ng
r
e
a
l
-
ti
me
da
ta
s
tr
e
a
ms
a
nd
a
da
pti
ve
lea
r
ning
a
lgor
it
hms
c
ould
f
u
r
ther
e
nha
nc
e
the
e
f
f
e
c
ti
ve
ne
s
s
of
r
e
c
omm
e
nde
r
s
y
s
tems
in
the
f
inanc
e
a
nd
ba
nking
s
e
c
tor
s
.
Addit
ionally,
int
e
r
dis
c
ipl
i
na
r
y
r
e
s
e
a
r
c
h
invol
ving
be
ha
vior
a
l
f
inanc
e
a
nd
A
I
e
thi
c
s
c
ould
pr
ovide
de
e
pe
r
ins
ight
s
int
o
the
e
thi
c
a
l
c
ons
ider
a
ti
ons
a
nd
us
e
r
a
c
c
e
ptan
c
e
of
AI
-
powe
r
e
d
f
inanc
ial
r
e
c
omm
e
nda
ti
ons
.
4.
CONC
L
USI
ON
T
his
s
tudy
pur
pos
e
d
to
inves
ti
ga
te
thr
e
e
r
e
s
e
a
r
c
h
que
s
ti
ons
:
i)
wha
t
a
r
e
the
c
ha
ll
e
nge
s
of
us
ing
r
e
c
omm
e
nda
ti
on
s
ys
tems
in
the
f
inanc
e
a
nd
ba
nking
s
e
c
tor
s
?
ii
)
how
a
r
e
the
c
ur
r
e
nt
c
ha
ll
e
ng
e
s
be
ing
a
ddr
e
s
s
e
d?
a
nd
ii
i)
a
r
e
the
c
ur
r
e
nt
s
olut
ions
s
a
ti
s
f
a
c
tor
y?
Our
s
ys
tema
ti
c
r
e
view
ha
s
c
omp
r
e
he
ns
ively
s
ynthes
ize
d
c
ha
ll
e
ng
e
s
r
e
late
d
to
r
e
c
omm
e
nda
ti
o
n
s
ys
tems
a
nd
a
na
lyze
d
the
pr
opos
e
d
s
olut
ions
to
f
inanc
e
a
nd
ba
nking
indus
tr
y
c
ha
ll
e
nge
s
.
W
hil
e
s
igni
f
ic
a
nt
a
dva
nc
e
ments
ha
v
e
be
e
n
a
c
hiev
e
d,
s
e
ve
r
a
l
ga
ps
a
nd
li
mi
tations
r
e
main
,
ne
c
e
s
s
it
a
ti
ng
f
ur
ther
r
e
s
e
a
r
c
h
a
nd
de
ve
lopm
e
nt.
T
he
unique
c
ha
ll
e
nge
s
r
e
late
d
t
o
f
inanc
e
a
nd
ba
nking
include
c
old
-
s
tar
t
pr
oblems
,
pe
r
s
ona
li
z
a
ti
on,
f
r
a
ud
de
tec
ti
on,
tr
a
ns
pa
r
e
nc
y,
a
nd
da
ta
pr
iv
a
c
y.
W
e
c
onc
luded
that
int
e
gr
a
ti
ng
a
dva
nc
e
d
mac
hine
lea
r
ning
models
,
diver
s
e
da
ta
s
our
c
e
s
,
a
nd
e
mer
ging
tec
hnologi
e
s
s
igni
f
ica
ntl
y
e
nha
nc
e
s
r
e
c
omm
e
nda
ti
on
a
c
c
ur
a
c
y,
s
ys
tem
r
e
li
a
bil
it
y
,
a
nd
us
e
r
tr
us
t.
F
utur
e
r
e
s
e
a
r
c
h
s
hould
f
oc
us
on
r
e
a
l
-
wor
ld
a
ppli
c
a
ti
ons
,
a
da
pti
ve
lea
r
ning
a
lgor
it
hms
,
a
nd
int
e
r
dis
c
ipl
inar
y
a
ppr
oa
c
he
s
to
f
ur
ther
im
p
r
ove
AI
-
powe
r
e
d
f
inanc
ial
r
e
c
omm
e
nda
t
ions
’
e
f
f
e
c
ti
ve
ne
s
s
a
nd
e
thi
c
a
l
c
ons
ider
a
ti
ons
.
F
UN
DI
NG
I
NF
ORM
AT
I
ON
Author
s
s
tate
ther
e
is
no
f
unding
invol
ve
d.
AU
T
HO
R
CONT
RI
B
U
T
I
ONS
S
T
AT
E
M
E
N
T
All
a
uthor
s
ha
ve
c
ontr
ibut
e
d
s
igni
f
ica
ntl
y
to
th
e
c
onc
e
ptualiza
ti
on,
methodology,
r
e
s
e
a
r
c
h
a
nd
wr
it
ing
o
f
the
pa
pe
r
,
a
nd
they
ha
ve
a
ll
r
e
a
d
a
nd
a
gr
e
e
d
to
the
c
u
r
r
e
nt
ve
r
s
ion
o
f
the
manus
c
r
ipt
.
Us
ing
the
jour
na
l
C
ontr
ibut
o
r
R
oles
T
a
xonomy
(
C
R
e
diT
)
,
th
e
c
ontr
ibut
ions
a
r
e
s
umm
a
r
is
e
d
a
s
f
o
ll
ows
.
Nam
e
of
Au
t
h
or
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
L
os
s
a
n
B
onde
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
✓
Abdoul
Ka
r
im
B
icha
nga
✓
✓
✓
✓
✓
✓
✓
✓
✓
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
C
hall
e
nge
s
of
r
e
c
omm
e
nde
r
s
y
s
te
ms
in
fi
nanc
e
and
bank
ing:
a
s
y
s
tem
ati
c
r
e
v
iew
…
(
L
os
s
an
B
onde
)
2565
C
:
C
onc
e
pt
ua
li
z
a
ti
on
M
:
M
e
th
odol
ogy
So
:
So
f
twa
r
e
Va
:
Va
li
da
ti
on
Fo
:
Fo
r
ma
l
a
na
ly
s
is
I
:
I
nve
s
ti
ga
ti
on
R
:
R
e
s
our
c
e
s
D
:
D
a
ta
C
ur
a
ti
on
O
:
W
r
it
in
g
-
O
r
ig
in
a
l
D
r
a
f
t
E
:
W
r
it
in
g
-
R
e
vi
e
w
&
E
di
ti
ng
Vi
:
Vi
s
ua
li
z
a
ti
on
Su
:
Su
pe
r
vi
s
io
n
P
:
P
r
oj
e
c
t
a
dmi
ni
s
tr
a
ti
on
Fu
:
Fu
ndi
ng a
c
qui
s
it
io
n
CONF
L
I
CT
OF
I
NT
E
RE
S
T
S
T
AT
E
M
E
N
T
Author
s
s
tate
no
c
onf
li
c
t
of
int
e
r
e
s
t.
I
NF
ORM
E
D
CONSE
NT
T
his
s
tudy
d
id
no
t
invo
lve
indi
v
iduals
nor
a
ny
pe
r
s
ona
l
identif
ica
ti
on
inf
or
mat
ion
that
c
ould
r
e
qui
r
e
a
ny
inf
or
med
c
ons
e
nt.
E
T
HI
CA
L
AP
P
ROVA
L
T
his
pa
pe
r
doe
s
not
invol
ve
pe
ople
or
a
nim
a
ls
;
no
inves
ti
ga
ti
on
ha
s
invol
ve
d
human
s
ubjec
ts
.
T
he
r
e
f
or
e
,
the
a
utho
r
s
did
not
s
e
e
k
a
ppr
ova
l
f
r
om
a
ny
ins
ti
tut
ional
r
e
view
boa
r
d.
DA
T
A
AV
AI
L
A
B
I
L
I
T
Y
Da
ta
a
va
il
a
bil
it
y
is
not
a
ppli
c
a
ble
to
thi
s
pa
pe
r
a
s
no
ne
w
da
ta
we
r
e
c
r
e
a
ted
o
r
a
na
lyze
d
in
thi
s
s
tudy.
RE
F
E
RE
NC
E
S
[
1]
M
.
S
ha
r
a
f
,
E
.
E
.
D
.
H
e
mda
n,
A
.
E
l
-
S
a
ye
d,
a
nd
N
.
A
.
E
l
-
B
a
hna
s
a
w
y,
“
A
s
ur
ve
y
on
r
e
c
omm
e
nda
ti
on
s
ys
te
m
s
f
or
f
in
a
n
c
ia
l
s
e
r
vi
c
e
s
,”
M
ul
ti
m
e
di
a T
ool
s
and A
ppl
ic
at
io
ns
, vol
. 81, no. 12, p
p. 16761
–
16781, 2022, doi:
1
0.1007/s
11042
-
022
-
12564
-
1.
[
2]
D
.
R
oy
a
nd
M
.
D
ut
ta
,
“
A
s
ys
te
ma
ti
c
r
e
vi
e
w
a
nd
r
e
s
e
a
r
c
h
pe
r
s
pe
c
ti
ve
on
r
e
c
omm
e
nde
r
s
ys
te
ms
,”
J
our
nal
of
B
ig
D
at
a
,
vo
l.
9
,
no.
1, 2022, doi:
10.1186/s
40537
-
022
-
00592
-
5.
[
3]
J
.
H
ua
ng,
J
.
C
ha
i,
a
nd
S
.
C
ho,
“
D
e
e
p
le
a
r
ni
ng
in
f
in
a
n
c
e
a
nd
ba
nki
ng:
A
li
te
r
a
tu
r
e
r
e
vi
e
w
a
nd
c
la
s
s
if
ic
a
ti
on,”
F
r
ont
ie
r
s
of
B
us
in
e
s
s
R
e
s
e
a
r
c
h i
n C
hi
na
, vol
. 14, no. 1, 2020, doi
:
10.1186/
s
11782
-
020
-
00082
-
6.
[
4]
H
.
H
e
r
ma
w
a
n,
F
.
M
a
ha
r
di
ka
,
I
.
D
a
r
ma
ya
nt
i,
R
.
B
.
B
.
S
uma
nt
r
i,
D
.
I
.
S
.
S
a
put
r
a
,
a
nd
A
.
A
mi
nuddin,
“
N
e
w
me
di
a
a
s
a
to
ol
s
to
im
pr
ove
c
r
e
a
ti
ve
th
in
ki
ng:
a
s
y
s
te
ma
ti
c
li
te
r
a
tu
r
e
r
e
vi
e
w
,”
in
2023
I
E
E
E
7t
h
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
I
nf
or
m
at
io
n
T
e
c
hnol
ogy
,
I
nf
or
m
at
io
n
Sy
s
te
m
s
and
E
le
c
tr
ic
al
E
ngi
ne
e
r
in
g
(
I
C
I
T
I
SE
E
)
,
2023,
pp.
64
–
69,
doi
:
10.1109/I
C
I
T
I
S
E
E
58992.2
023.10404556.
[
5]
C
.
W
ohl
in
,
“
G
ui
de
li
ne
s
f
or
s
now
ba
ll
in
g
in
s
ys
te
m
a
ti
c
li
te
r
a
tu
r
e
s
tu
di
e
s
a
nd
a
r
e
pl
ic
a
ti
on
in
s
of
twa
r
e
e
ngi
n
e
e
r
in
g,”
in
P
r
oc
e
e
di
ngs
of
th
e
18t
h
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
E
v
al
uat
io
n
and
A
s
s
e
s
s
m
e
nt
in
Sof
tw
a
r
e
E
ngi
ne
e
r
in
g
,
2014,
pp.
1
–
10,
doi
:
10.1145/2601248.26
01268.
[
6]
D
.
L
iu
,
G
.
P
.
F
a
r
a
ja
ll
a
,
a
nd
A
.
B
oul
e
nge
r
,
“
B
R
e
c
th
e
ba
nk:
c
ont
e
xt
-
a
w
a
r
e
s
e
lf
-
a
tt
e
nt
iv
e
e
nc
ode
r
f
or
ba
nki
ng
pr
od
uc
ts
r
e
c
omm
e
nda
ti
on,”
in
2022
I
nt
e
r
nat
io
nal
J
oi
nt
C
onf
e
r
e
nc
e
on
N
e
ur
al
N
e
tw
o
r
k
s
(
I
J
C
N
N
)
,
2022,
pp.
1
–
8,
doi
:
10.1109/I
J
C
N
N
55064.2022.9892130.
[7
]
S
.
C
he
n,
Y
.
Q
iu
,
J
.
L
i,
K
.
F
a
ng,
a
nd
K
.
F
a
ng,
“
P
r
e
c
is
io
n
ma
r
ke
ti
ng
f
or
f
in
a
nc
ia
l
in
dus
tr
y
us
in
g
a
P
U
-
le
a
r
ni
ng
r
e
c
omm
e
nda
ti
on
me
th
od,”
J
our
nal
of
B
us
in
e
s
s
R
e
s
e
a
r
c
h
, vol
. 160, 2023, doi:
10
.1016/j
.j
bus
r
e
s
.2023.113771.
[
8]
X
.
C
he
n,
A
.
R
e
ib
ma
n,
a
nd
S
.
A
r
or
a
,
“
S
e
que
nt
ia
l
r
e
c
omm
e
nd
a
ti
on
mode
l
f
or
ne
xt
pur
c
ha
s
e
pr
e
di
c
ti
on,”
C
om
put
e
r
Sc
i
e
nc
e
&
I
nf
or
m
at
io
n T
e
c
hnol
ogy
(
C
S & I
T
)
, vol
. 13, no. 10, pp. 141
–
15
8, 2022, doi:
10.5121/cs
it
.2023.1310013.
[
9]
G
.
M
e
ndonç
a
,
M
.
S
a
nt
o
s
,
A
.
G
onç
a
lv
e
s
,
a
nd
Y
.
A
lm
e
id
a
,
“
R
e
th
in
ki
ng
f
in
a
nc
ia
l
s
e
r
vi
c
e
pr
omot
io
n
w
it
h
hybr
id
r
e
c
omm
e
nde
r
s
ys
te
ms
a
t
P
ic
P
a
y,”
ar
X
iv
-
C
om
put
e
r
S
c
ie
nc
e
, pp. 1
–
4, 2023.
[
10]
O
.
O
ye
bode
a
nd
R
.
O
r
ji
,
“
A
hybr
id
r
e
c
omm
e
nde
r
s
ys
te
m
f
or
pr
oduc
t
s
a
le
s
in
a
ba
nki
ng
e
nvi
r
onme
nt
,”
J
our
nal
of
B
ank
in
g
and
F
in
anc
ia
l
T
e
c
hnol
og
y
, vol
. 4, no. 1, pp. 15
–
25, 2020, doi:
10.1007/s
42786
-
019
-
00014
-
w.
[
11]
D
.
de
S
.
M
or
a
e
s
e
t
al
.
,
“
U
s
in
g
z
e
r
o
-
s
hot
pr
ompt
in
g
in
th
e
a
ut
o
ma
ti
c
c
r
e
a
ti
on
a
nd
e
xpa
n
s
io
n
of
to
pi
c
ta
xonomi
e
s
f
or
ta
ggi
ng
r
e
ta
il
ba
nki
ng t
r
a
ns
a
c
ti
ons
,”
a
r
X
iv
-
C
om
put
e
r
Sc
i
e
nc
e
, pp. 1
–
8, 2024.
[
12]
A
.
G
hi
ye
,
B
.
B
a
r
r
e
a
u,
L
.
C
a
r
li
e
r
,
a
nd
M
.
V
a
z
ir
gi
a
nni
s
,
“
A
da
p
ti
ve
c
ol
la
bor
a
ti
ve
f
il
te
r
in
g
w
it
h
pe
r
s
ona
li
z
e
d
ti
me
de
c
a
y
f
unc
ti
o
ns
f
or
f
in
a
nc
ia
l
pr
oduc
t
r
e
c
omm
e
nda
ti
on,”
in
P
r
oc
e
e
di
ngs
of
th
e
17t
h
A
C
M
C
onf
e
r
e
nc
e
on
R
e
c
om
m
e
nd
e
r
Sy
s
te
m
s
,
2023,
pp.
7
98
–
804, doi:
10.1145/3604915.3
608832.
[
13]
Y
.
L
e
ng
e
t
al
.
,
“
C
us
to
me
r
-
c
a
te
gor
y
in
te
r
e
s
t
mode
l:
a
gr
a
ph
-
ba
s
e
d
c
ol
la
bor
a
ti
ve
f
il
te
r
in
g
mode
l
w
it
h
a
ppl
ic
a
ti
ons
in
f
in
a
nc
e
,”
in
P
r
oc
e
e
di
ngs
of
t
he
T
hi
r
d A
C
M
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on A
I
i
n F
in
anc
e
, 2022, pp.
165
–
173, doi:
10.1145/3533271.3
561757
.
[
14]
Q
. Z
ha
ng, D. Z
ha
ng, J
.
L
u,
G
. Z
ha
ng, W
. Q
u,
a
nd M
. C
ohe
n, “
A
r
e
c
omm
e
nde
r
s
ys
te
m f
or
c
ol
d
-
s
ta
r
t
it
e
ms
:
a
c
a
s
e
s
tu
dy i
n t
he
r
e
a
l
e
s
ta
te
in
dus
tr
y,”
in
2019
I
E
E
E
14t
h
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
I
nt
e
ll
ig
e
nt
Sy
s
te
m
s
a
nd
K
now
le
dg
e
E
ngi
ne
e
r
in
g
(
I
SK
E
)
,
2019,
pp. 1185
–
1192, doi:
10.1109/I
S
K
E
47853.2019.9170411.
[
15]
M
.
L
ia
n
a
nd
J
.
L
i,
“
F
in
a
nc
ia
l
pr
oduc
t
r
e
c
omm
e
nda
ti
on
s
ys
te
m
ba
s
e
d
on
tr
a
ns
f
or
me
r
,”
in
2020
I
E
E
E
4t
h
I
nf
or
m
at
io
n
T
e
c
hnol
ogy
,
N
e
tw
or
k
in
g,
E
le
c
t
r
oni
c
and
A
ut
om
at
io
n
C
ont
r
ol
C
onf
e
r
e
nc
e
(
I
T
N
E
C
)
,
2020,
pp.
2547
–
2551,
doi
:
10.1109/I
T
N
E
C
48623.2020.9084812.
[
16]
J
.
X
ue
,
E
.
Z
hu,
Q
.
L
iu
,
a
nd
J
.
Y
in
,
“
G
r
oup
r
e
c
omm
e
nda
ti
on
ba
s
e
d
on
f
in
a
nc
ia
l
s
oc
ia
l
ne
twor
k
f
or
r
obo
-
a
dvi
s
or
,”
I
E
E
E
A
c
c
e
s
s
,
vol
. 6, pp. 54527
–
54535, 2018, doi:
10
.1109/AC
C
E
S
S
.2018.28
71131.
[
17]
W
.
W
a
ng a
nd
K
.
K
.
M
is
hr
a
,
“
A
nov
e
l
s
to
c
k
tr
a
di
ng
pr
e
di
c
ti
on a
nd
r
e
c
omm
e
nda
ti
on
s
y
s
te
m,”
M
ul
ti
m
e
di
a
T
ool
s
and
A
ppl
ic
at
io
ns
,
vol
. 77, no. 4, pp. 4203
–
4215, 2018, doi:
10.1007/s
11042
-
017
-
4587
-
z.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2252
-
8938
I
nt
J
Ar
ti
f
I
ntell
,
Vol.
14
,
No.
4
,
Augus
t
2025
:
255
9
-
2567
2566
[
18]
X
.
W
a
ng,
Y
.
W
a
ng,
B
.
W
e
ng,
a
nd
A
.
V
in
e
l,
“
S
to
c
k2V
e
c
:
a
hybr
id
de
e
p
le
a
r
ni
ng
f
r
a
me
w
or
k
f
or
s
to
c
k
ma
r
ke
t
pr
e
di
c
ti
on
w
it
h
r
e
pr
e
s
e
nt
a
ti
on l
e
a
r
ni
ng a
nd t
e
mpor
a
l
c
onvolut
io
na
l
ne
twor
k,”
a
r
X
iv
-
Q
uant
it
at
iv
e
F
in
anc
e
, pp. 1
–
48, 2020.
[
19]
S
.
F
e
ng,
C
.
X
u,
Y
.
Z
uo,
G
.
C
he
n,
F
.
L
in
,
a
nd
J
.
X
ia
H
ou,
“
R
e
la
ti
on
-
a
w
a
r
e
dyna
mi
c
a
tt
r
ib
ut
e
d
gr
a
ph
a
tt
e
nt
io
n
ne
two
r
k
f
or
s
to
c
ks
r
e
c
omm
e
nda
ti
on,”
P
at
te
r
n R
e
c
ogni
ti
on
, vol
. 121, 2022, doi:
10.1016/j
.pa
tc
og.2021.108119.
[
20]
V
.
R
a
vi
,
D
.
P
r
a
de
e
pkuma
r
,
a
nd
K
.
D
e
b,
“
F
in
a
nc
ia
l
ti
me
s
e
r
ie
s
pr
e
di
c
ti
on
u
s
in
g
hybr
id
s
of
c
ha
o
s
th
e
or
y,
mul
ti
-
la
ye
r
pe
r
c
e
p
tr
on
a
nd
mul
ti
-
obj
e
c
ti
ve
e
vol
ut
io
na
r
y
a
lg
or
it
hms
,”
Sw
ar
m
and
E
v
ol
ut
io
nar
y
C
om
put
at
io
n
,
vol
.
36,
pp.
136
–
149,
2017,
doi
:
10.1016/j
.s
w
e
vo.2017.05.003.
[
21]
D
.
L
.
M
in
h,
A
.
S
a
de
ghi
-
N
ia
r
a
ki
,
H
.
D
.
H
uy,
K
.
M
in
,
a
nd
H
.
M
oon,
“
D
e
e
p
le
a
r
ni
ng
a
ppr
oa
c
h
f
or
s
hor
t
-
te
r
m
s
to
c
k
tr
e
nds
pr
e
di
c
ti
on
ba
s
e
d
on
two
-
s
tr
e
a
m
ga
te
d
r
e
c
ur
r
e
nt
uni
t
ne
twor
k,”
I
E
E
E
A
c
c
e
s
s
,
vol
.
6,
pp.
55392
–
55404,
2018,
doi
:
10.1109/AC
C
E
S
S
.2018.2868970.
[
22]
S
.
P
.
C
ha
tz
is
,
V
.
S
ia
koul
is
,
A
.
P
e
tr
opo
ul
os
,
E
.
S
ta
vr
oul
a
ki
s
,
a
nd
N
.
V
la
c
hogi
a
nn
a
ki
s
,
“
F
or
e
c
a
s
ti
ng
s
to
c
k
ma
r
ke
t
c
r
is
i
s
e
v
e
nt
s
us
in
g
de
e
p
a
nd
s
t
a
ti
s
ti
c
a
l
ma
c
hi
ne
le
a
r
ni
ng
te
c
hni
que
s
,”
E
x
p
e
r
t
Sy
s
te
m
s
w
it
h
A
ppl
ic
at
io
ns
,
vol
.
112,
pp.
353
–
371,
2018,
doi
:
10.1016/j
.e
s
w
a
.2018.06.032.
[
23]
R
.
S
in
gh
a
n
d
S
.
S
r
iv
a
s
ta
va
,
“
S
to
c
k
pr
e
di
c
ti
on
us
in
g
de
e
p
le
a
r
ni
ng,”
M
ul
ti
m
e
di
a
T
oo
ls
and
A
ppl
ic
at
io
ns
,
vol
.
76,
no.
18,
pp.
18569
–
18584, 2017, doi:
10.1007/s
11042
-
016
-
4159
-
7.
[
24]
Y
.
L
iu
,
H
.
M
a
,
Y
.
J
i
a
ng,
a
nd
Z
.
L
i,
“
M
ode
ll
in
g
r
is
k
a
nd
r
e
tu
r
n
a
w
a
r
e
ne
s
s
f
or
p2p
le
ndi
n
g
r
e
c
omm
e
nda
ti
on
w
it
h
g
r
a
ph
c
onvolut
io
na
l
ne
twor
ks
,”
A
ppl
ie
d I
nt
e
ll
ig
e
nc
e
, vol
. 52, no. 5, pp. 4999
–
5014, 2022, doi:
10.1007/s
10489
-
021
-
02680
-
0.
[
25]
J
.
J
ur
govs
ky
e
t
al
.
,
“
S
e
que
nc
e
c
la
s
s
if
ic
a
ti
on
f
or
c
r
e
di
t
-
c
a
r
d
f
r
a
ud
de
te
c
ti
on,”
E
x
pe
r
t
S
y
s
te
m
s
w
it
h
A
ppl
ic
at
io
ns
,
vol
.
100,
pp.
2
34
–
245, 2018, doi:
10.1016/j
.e
s
w
a
.2018.01.037.
[
26]
W
.
Z
he
ng,
L
.
Y
a
n,
C
.
G
ou,
a
nd
F
.
Y
.
W
a
ng,
“
F
e
de
r
a
te
d
me
ta
-
le
a
r
ni
ng
f
or
f
r
a
udul
e
nt
c
r
e
di
t
c
a
r
d
de
te
c
ti
on,”
in
I
J
C
A
I
I
nt
e
r
nat
io
nal
J
oi
nt
C
onf
e
r
e
nc
e
on A
r
ti
fi
c
ia
l
I
nt
e
ll
ig
e
nc
e
, 2020, p
p. 4654
–
4660, doi:
10.24963/i
jc
a
i.
2020/642.
[
27]
B
.
Z
hu,
W
.
Y
a
ng,
H
.
W
a
ng,
a
nd
Y
.
Y
ua
n,
“
A
hybr
id
de
e
p
le
a
r
ni
ng
mode
l
f
or
c
ons
ume
r
c
r
e
di
t
s
c
or
in
g,”
in
2018
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on A
r
ti
fi
c
ia
l
I
nt
e
ll
ig
e
nc
e
and B
ig
D
at
a, I
C
A
I
B
D
20
18
, 2018, pp. 205
–
208, do
i:
10.1109/I
C
A
I
B
D
.2018.8396195.
[
28]
V
. K
a
na
pa
r
th
i,
“
A
I
-
ba
s
e
d pe
r
s
on
a
li
z
a
ti
on a
nd t
r
us
t
in
di
gi
ta
l
f
in
a
nc
e
,”
ar
X
iv
-
C
om
put
e
r
Sc
ie
nc
e
, pp. 1
–
9, 2024.
[
29]
H
.
W
a
ng,
S
.
M
a
,
H
.
N
.
D
a
i,
M
.
I
mr
a
n,
a
nd
T
.
W
a
ng,
“
B
lo
c
k
c
ha
in
-
ba
s
e
d
d
a
ta
pr
iv
a
c
y
ma
na
ge
me
nt
w
it
h
N
udge
th
e
or
y
in
o
pe
n
ba
nki
ng,”
F
ut
ur
e
G
e
ne
r
at
io
n C
om
put
e
r
S
y
s
te
m
s
, vol
. 110, pp.
812
–
823, 2020, doi:
10.1016/j
.f
ut
ur
e
.2019.09.010.
[
30]
A
.
S
ha
h
e
t
al
.
,
“
N
um
e
r
ic
a
l
c
la
im
de
te
c
ti
on
in
f
in
a
nc
e
:
a
n
e
w
f
in
a
nc
ia
l
da
ta
s
e
t,
w
e
a
k
-
s
upe
r
vi
s
io
n
mode
l,
a
nd
ma
r
ke
t
a
na
ly
s
is
,”
ar
X
iv
-
C
om
put
e
r
Sc
ie
nc
e
, pp. 1
–
16, 2024.
[
31]
S
.
S
oha
ngi
r
,
D
.
W
a
ng,
A
.
P
ome
r
a
ne
ts
,
a
nd
T
.
M
.
K
hos
hgof
ta
a
r
,
“
B
ig
da
t
a
:
de
e
p
le
a
r
ni
ng
f
or
f
in
a
nc
ia
l
s
e
nt
im
e
nt
a
na
ly
s
is
,”
J
our
nal
of
B
ig
D
at
a
, vol
. 5, no. 1, 2018, doi:
10.1186/s
40537
-
017
-
0111
-
6.
[
32]
T
.
H
os
a
ka
,
“
B
a
nkr
upt
c
y
pr
e
di
c
ti
on
us
in
g
im
a
ge
d
f
in
a
nc
ia
l
r
a
ti
os
a
nd
c
onvolut
io
na
l
ne
ur
a
l
ne
twor
ks
,”
E
x
pe
r
t
Sy
s
te
m
s
w
it
h
A
ppl
ic
at
io
ns
, vol
. 117, pp. 287
–
299, 2019, doi:
10.1016/j
.e
s
w
a
.2018.09.039.
[
33]
M
.
K
r
a
us
a
nd
S
.
F
e
ue
r
r
ie
ge
l,
“
D
e
c
is
io
n
s
uppor
t
f
r
om
f
in
a
nc
i
a
l
di
s
c
lo
s
ur
e
s
w
it
h
de
e
p
ne
ur
a
l
ne
twor
ks
a
nd
tr
a
ns
f
e
r
le
a
r
ni
ng,”
D
e
c
is
io
n Suppor
t
Sy
s
te
m
s
, vol
. 104, pp. 38
–
48, 2017, doi:
10.1
016/
j.
ds
s
.2017.10.001.
[
34]
S
.
A
.
C
or
w
in
,
S
.
A
.
L
a
r
oc
que
,
a
nd
M
.
A
.
S
t
e
ge
mol
le
r
,
“
I
nve
s
t
me
nt
ba
nki
ng
r
e
la
ti
ons
hi
p
s
a
nd
a
na
ly
s
t
a
f
f
il
ia
ti
on
bi
a
s
:
T
h
e
im
pa
c
t
of
th
e
gl
oba
l
s
e
tt
le
me
nt
on
s
a
nc
ti
one
d
a
nd
non
-
s
a
nc
ti
one
d
ba
n
ks
,”
J
our
nal
of
F
in
anc
ia
l
E
c
onomic
s
,
vol
.
124,
no.
3,
pp.
614
–
631,
2017, doi:
10.1016/j
.j
f
in
e
c
o.2017.03.005.
[
35]
G
. D
e
s
ir
e
na
, A
. D
ia
z
, J
. D
e
s
ir
e
na
, I
.
M
or
e
no, a
nd D
.
G
a
r
c
ia
, “
M
a
xi
mi
z
in
g c
us
to
me
r
l
if
e
ti
me
va
lu
e
us
in
g s
ta
c
ke
d ne
ur
a
l
ne
tw
or
k
s
:
a
n
in
s
ur
a
nc
e
in
dus
tr
y
a
ppl
ic
a
ti
on,”
in
2019
18t
h
I
E
E
E
I
nt
e
r
nat
io
nal
C
onf
e
r
e
n
c
e
O
n
M
ac
hi
ne
L
e
ar
ni
ng
A
nd
A
ppl
ic
at
io
ns
(
I
C
M
L
A
)
, 2019, pp. 541
–
544, doi:
10.1109/I
C
M
L
A
.2019.0010
1.
[
36]
M
.
E
r
ke
k,
K
.
Ç
a
yı
r
lı
,
H
.
T
a
ş
,
A
.
H
e
pş
e
n,
a
nd
T
.
A
yt
e
ki
n,
“
R
e
c
omm
e
nda
ti
on
s
ys
te
ms
a
ppl
ie
d
in
T
ur
ki
s
h
r
e
a
l
e
s
ta
te
ma
r
ke
t,
”
J
our
nal
of
C
om
put
at
io
ns
& M
ode
ll
in
g
, vol
. 10, no. 1, pp. 1792
–
8850, 2020.
[
37]
L
.
W
a
ng,
Y
.
L
iu
,
a
nd
J
.
W
u,
“
R
e
s
e
a
r
c
h
on
f
in
a
nc
ia
l
a
dve
r
ti
s
e
me
nt
pe
r
s
ona
li
s
e
d
r
e
c
omm
e
nda
ti
on
me
th
od
ba
s
e
d
on
c
us
to
me
r
s
e
gme
nt
a
ti
on,”
I
nt
e
r
nat
io
nal
J
our
nal
of
W
ir
e
le
s
s
a
nd
M
obi
le
C
om
put
in
g
,
vol
.
14,
no.
1,
2018,
doi
:
10.1504/i
jwmc
.2018.10011092.
[
38]
Y
.
G
u
o,
Y
.
Z
hou,
X
.
H
u,
a
nd
W
.
C
he
ng,
“
R
e
s
e
a
r
c
h
on
r
e
c
om
me
nda
ti
on
of
in
s
ur
a
nc
e
pr
oduc
ts
ba
s
e
d
on
r
a
ndom
f
or
e
s
t,
”
in
2019
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
M
ac
hi
ne
L
e
ar
ni
ng,
B
ig
D
at
a
and
B
us
in
e
s
s
I
nt
e
ll
ig
e
nc
e
(
M
L
B
D
B
I
)
,
2019,
pp.
308
–
311,
doi
:
10.1109/M
L
B
D
B
I
48998.2019.00069.
[
39]
N
.
P
e
r
e
ir
a
a
nd
S
.
L
.
V
a
r
ma
,
“
F
in
a
nc
ia
l
pl
a
nni
ng
r
e
c
omm
e
nda
ti
on
s
ys
te
m
us
in
g
c
ont
e
nt
-
ba
s
e
d
c
ol
la
bor
a
ti
ve
a
nd
d
e
mogr
a
phi
c
f
il
te
r
in
g,”
A
dv
anc
e
s
i
n I
nt
e
ll
ig
e
nt
Sy
s
te
m
s
and C
om
put
in
g
, vol
.
669, pp. 141
–
151, 2019, doi:
10.100
7/
978
-
981
-
10
-
8968
-
8_12.
[
40]
P
.
Ł
a
dyż
yńs
ki
,
K
.
Ż
bi
kow
s
ki
,
a
nd
P
.
G
a
w
r
ys
ia
k,
“
D
ir
e
c
t
ma
r
k
e
ti
ng
c
a
mpa
ig
ns
in
r
e
ta
il
ba
nki
ng
w
it
h
th
e
us
e
of
d
e
e
p
le
a
r
ni
ng
a
n
d
r
a
ndom f
or
e
s
ts
,”
E
x
pe
r
t
Sy
s
t
e
m
s
w
it
h A
ppl
ic
at
io
ns
, vol
. 134, pp
. 28
–
35, 2019, doi:
10.1016/j
.e
s
w
a
.20
19.05.020.
[
41]
S
. C
ha
kr
a
bor
ty
,
“
C
a
pt
ur
in
g f
in
a
nc
ia
l
ma
r
ke
ts
t
o a
ppl
y de
e
p r
e
i
nf
or
c
e
me
nt
l
e
a
r
ni
ng,
”
ar
X
iv
-
Q
uant
it
at
iv
e
F
in
a
nc
e
, pp. 1
-
17,
2019
.
[
42]
J
.
R
e
n,
J
.
L
ong,
a
nd
Z
.
X
u,
“
F
in
a
nc
ia
l
ne
w
s
r
e
c
omm
e
nda
ti
on
ba
s
e
d
on
gr
a
ph
e
mbe
ddi
ngs
,”
D
e
c
is
io
n
Suppor
t
Sy
s
te
m
s
,
vol
.
125,
2019, doi:
10.1016/j
.ds
s
.2019.113115.
[
43]
C
.
Z
ha
ng,
C
.
S
ong,
S
.
A
ga
r
w
a
l,
H
.
W
u,
X
.
Z
ha
ng,
a
nd
J
.
J
.
L
u,
“
A
s
e
ma
nt
ic
s
e
a
r
c
h
f
r
a
me
w
or
k
f
o
r
s
im
il
a
r
a
udi
t
is
s
ue
r
e
c
omm
e
nda
ti
on
in
f
in
a
nc
ia
l
in
dus
tr
y,”
in
P
r
oc
e
e
di
ngs
of
th
e
Si
x
te
e
nt
h
A
C
M
I
nt
e
r
na
ti
onal
C
onf
e
r
e
nc
e
on
W
e
b
S
e
ar
c
h
and
D
at
a
M
in
in
g
, 2023, pp. 1208
–
1211, doi:
10.1145/3539597.3
573040.
[
44]
L
.
L
e
s
a
ge
,
M
.
D
e
a
c
onu,
A
.
L
e
ja
y,
J
.
A
.
M
e
ir
a
,
G
.
N
ic
hi
l,
a
nd
R
.
S
ta
te
,
“
A
r
e
c
omm
e
nda
ti
on
s
y
s
te
m
f
or
c
a
r
in
s
ur
a
nc
e
,”
E
ur
op
e
an
A
c
tu
ar
ia
l
J
our
nal
, vol
. 10
, no. 2, pp. 377
–
398, 2020, doi:
10.1007/s
13385
-
020
-
00236
-
z.
[
45]
A
.
S
ha
r
if
ih
os
s
e
in
i,
“
A
c
a
s
e
s
tu
dy
f
or
pr
e
s
e
nt
in
g
B
a
nk
r
e
c
om
me
nde
r
s
ys
te
m
s
ba
s
e
d
on
bon
c
a
r
d
tr
a
ns
a
c
ti
on
da
ta
,”
in
2019
9t
h
I
nt
e
r
nat
io
nal
C
onf
e
r
e
n
c
e
on
C
om
put
e
r
and
K
now
le
d
ge
E
ngi
ne
e
r
in
g
(
I
C
C
K
E
)
,
O
c
t.
2019,
pp.
72
–
77,
doi
:
10.1109/I
C
C
K
E
48569.2019.8964698.
[
46]
K
. L
a
kka
r
a
ju
, S
.
E
.
J
one
s
, S
. K
. R
. V
ur
uma
,
V
. P
a
ll
a
ga
ni
, B
. C
.
M
uppa
s
a
ni
, a
nd B
. S
r
iv
a
s
ta
va
, “
L
L
M
s
f
or
f
in
a
nc
ia
l
a
dvi
s
e
me
nt
:
a
f
a
ir
ne
s
s
a
nd
e
f
f
ic
a
c
y
s
tu
dy
in
pe
r
s
ona
l
de
c
is
io
n
ma
ki
ng,”
in
4t
h
A
C
M
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
A
I
in
F
in
anc
e
,
2023,
pp.
1
00
–
107, doi:
10.1145/3604237.3
626867.
[
47]
Y
.
S
un,
M
.
F
a
ng,
a
nd
X
.
W
a
ng,
“
A
nove
l
s
to
c
k
r
e
c
omm
e
nda
ti
on
s
ys
t
e
m
us
in
g
G
uba
s
e
nt
im
e
nt
a
na
ly
s
is
,”
P
e
r
s
onal
and
U
bi
qui
to
us
C
om
put
in
g
, vol
. 22, no. 3
, pp. 575
–
587, 2018, doi:
10.1007/s
00779
-
018
-
1121
-
x.
[
48]
Z
.
Z
he
ng,
Y
.
G
a
o,
L
.
Y
in
,
a
nd
M
.
K
.
R
a
ba
r
is
on,
“
M
ode
li
ng
a
n
d
a
na
ly
s
is
of
a
s
to
c
k
-
ba
s
e
d
c
ol
la
bor
a
ti
ve
f
il
te
r
in
g
a
lg
or
it
hm
f
or
th
e
C
hi
ne
s
e
s
to
c
k ma
r
ke
t,
”
E
x
pe
r
t
S
y
s
te
m
s
w
it
h A
ppl
ic
at
io
ns
, vol
. 162, 20
20, doi:
10.1016/j
.e
s
w
a
.2019.113006.
[
49]
S
.
B
.
P
a
te
l,
P
.
B
ha
tt
a
c
ha
r
ya
,
S
.
T
a
nw
a
r
,
a
nd
N
.
K
uma
r
,
“
K
iR
T
i:
a
bl
oc
kc
ha
in
-
ba
s
e
d
c
r
e
di
t
r
e
c
omm
e
nde
r
s
ys
te
m
f
or
f
in
a
nc
ia
l
in
s
ti
tu
ti
ons
,”
I
E
E
E
T
r
ans
ac
ti
ons
on
N
e
tw
or
k
Sc
ie
nc
e
an
d
E
ngi
ne
e
r
in
g
,
vol
.
8,
no.
2,
pp.
1044
–
1054,
2021,
doi
:
10.1109/T
N
S
E
.2020.3005678.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
Ar
ti
f
I
ntell
I
S
S
N:
2252
-
8938
C
hall
e
nge
s
of
r
e
c
omm
e
nde
r
s
y
s
te
ms
in
fi
nanc
e
and
bank
ing:
a
s
y
s
tem
ati
c
r
e
v
iew
…
(
L
os
s
an
B
onde
)
2567
[
50]
A
.
B
r
in
i,
G
.
T
e
de
s
c
hi
,
a
nd
D
.
T
a
nt
a
r
i,
“
R
e
in
f
or
c
e
me
nt
le
a
r
ni
n
g
pol
ic
y
r
e
c
omm
e
nda
ti
on
f
or
in
te
r
ba
nk
ne
twor
k
s
ta
bi
li
ty
,”
J
ou
r
nal
of
F
in
anc
ia
l
St
abi
li
ty
, vol
. 67, 2023, doi:
10.1016/
j.
jf
s
.2023.10
1139.
[
51]
G
.
J
e
ong
a
nd
H
.
Y
.
K
im
,
“
I
mpr
ovi
ng
f
in
a
nc
ia
l
tr
a
di
ng
de
c
is
io
ns
us
in
g
de
e
p
Q
-
le
a
r
ni
ng:
P
r
e
di
c
ti
ng
th
e
numbe
r
of
s
h
a
r
e
s
,
a
c
t
io
n
s
tr
a
te
gi
e
s
,
a
nd
tr
a
ns
f
e
r
le
a
r
ni
ng,”
E
x
pe
r
t
Sy
s
te
m
s
w
it
h
A
ppl
ic
at
io
ns
,
vol
.
117,
pp.
125
–
138,
2019,
doi
:
10.1016/j
.e
s
w
a
.2018.0
9.036.
[
52]
E
.
H
.
N
ie
ve
s
,
“
N
e
w
a
ppr
oa
c
h
to
r
e
c
omm
e
nd
ba
nki
ng
pr
oduc
t
s
th
r
ough
a
hybr
id
r
e
c
omm
e
nde
r
s
ys
te
m,”
A
dv
anc
e
s
in
I
n
te
ll
i
ge
nt
Sy
s
te
m
s
and
C
om
put
in
g
, pp. 262
–
266, 2021, doi:
10.1007/978
-
3
-
030
-
58356
-
9_28.
[
53]
M
.
S
.
K
ha
n
a
nd
H
.
U
me
r
,
“
C
ha
tG
P
T
in
f
in
a
nc
e
:
A
ppl
ic
a
ti
on
s
,
c
ha
ll
e
ng
e
s
,
a
nd
s
ol
ut
io
ns
,”
H
e
li
y
on
,
vol
.
10,
no.
2,
2024,
doi
:
10.1016/j
.he
li
yon.2024.e
24890.
B
I
OG
RA
P
HI
E
S
OF
AU
T
HO
RS
Lo
s
s
a
n
Bo
nde
h
o
l
d
s
a
Ph
.
D
.
in
co
m
p
u
t
er
s
ci
e
n
ce
fro
m
t
h
e
U
n
i
v
ers
i
t
y
o
f
Sci
en
ce
s
an
d
T
ech
n
o
l
o
g
i
e
s
o
f
L
i
l
l
e,
Fra
n
ce,
i
n
2
0
0
6
.
H
e
cu
rr
en
t
l
y
s
er
v
es
a
t
t
h
e
A
d
v
e
n
t
i
s
t
U
n
i
v
ers
i
t
y
o
f
A
fri
ca,
w
h
ere
h
e
i
s
t
h
e
d
ea
n
o
f
t
h
e
s
c
h
o
o
l
o
f
p
o
s
t
g
rad
u
a
t
e
s
t
u
d
i
e
s
a
n
d
p
ro
g
ramme
c
o
o
r
d
i
n
at
o
r
fo
r
t
h
e
Mas
t
er
o
f
Sci
e
n
ce
i
n
A
p
p
l
i
ed
C
o
mp
u
t
er
Sci
e
n
ce
Pro
g
ram
.
Befo
r
e
j
o
i
n
i
n
g
t
h
e
A
d
v
e
n
t
i
s
t
U
n
i
v
er
s
i
t
y
o
f
A
fr
i
ca,
h
e
w
o
r
k
ed
a
t
“U
n
i
v
ers
i
t
e´
A
d
v
e
n
t
i
s
t
e
Co
s
en
d
ai
”
fr
o
m
2
0
0
7
t
o
2
0
1
6
,
s
erv
i
n
g
as
t
h
e
D
ea
n
o
f
t
h
e
Sc
h
o
o
l
o
f
Bu
s
i
n
es
s
an
d
Co
mp
u
t
er
Sci
en
ce.
H
i
s
res
earc
h
i
n
t
eres
t
s
are
i
n
art
i
fi
c
i
al
i
n
t
e
l
l
i
g
e
n
ce,
i
n
t
ern
e
t
o
f
t
h
i
n
g
s
,
an
d
d
i
g
i
t
al
t
ra
n
s
f
o
rmat
i
o
n
s
.
H
e
i
s
cu
rre
n
t
l
y
w
o
r
k
i
n
g
o
n
h
o
w
t
o
b
u
i
l
d
i
n
t
el
l
i
g
e
n
t
s
y
s
t
em
s
fo
r
d
i
g
i
t
a
l
t
r
an
s
f
o
rmat
i
o
n
s
.
H
e
can
b
e
co
n
t
act
e
d
at
emai
l
:
b
o
n
d
e
l
@
au
a.
ac.
k
e
o
r
l
2
b
o
n
d
e@
g
ma
i
l
.
co
m.
A
bdo
ul
K
a
ri
m
Bi
ch
a
ng
a
h
o
l
d
s
a
mas
t
er’s
d
eg
ree
i
n
i
n
fo
rma
t
i
o
n
s
y
s
t
em
s
w
i
t
h
a
s
p
ec
i
al
i
zat
i
o
n
i
n
d
ec
i
s
i
o
n
s
u
p
p
o
r
t
fro
m
t
h
e
Sc
h
o
o
l
o
f
C
o
mp
u
t
er
Sc
i
en
ce,
Bu
r
k
i
n
a
Fas
o
(2
0
1
9
-
2
0
2
1
)
a
n
d
a
b
ac
h
el
o
r’s
d
e
g
ree
i
n
s
t
a
t
i
s
t
i
cs
a
n
d
co
m
p
u
t
er
s
ci
e
n
ce
fr
o
m
N
az
i
B
o
n
i
U
n
i
v
ers
i
t
y
(2
0
1
5
-
2
0
1
8
).
H
e
i
s
p
ar
t
i
c
u
l
ar
l
y
i
n
t
eres
t
ed
i
n
res
earc
h
an
d
o
p
p
o
r
t
u
n
i
t
i
e
s
i
n
t
h
e
f
i
el
d
o
f
art
i
fi
c
i
a
l
i
n
t
el
l
i
g
en
ce.
H
e
can
b
e
co
n
t
ac
t
ed
at
ema
i
l
:
ab
d
o
u
l
b
i
ch
a
n
g
a@
g
mai
l
.
co
m.
Evaluation Warning : The document was created with Spire.PDF for Python.