Applied Mathematics and Nonlinear Sciences,
Год журнала:
2025,
Номер
10(1)
Опубликована: Янв. 1, 2025
Abstract
The
rapid
expansion
of
cross-national
e-commerce
has
brought
significant
opportunities
and
challenges
in
understanding
diverse
consumer
behavior.
This
study
introduces
an
innovative
framework
combining
the
XLSTM
(Extended
Long
Short-Term
Memory)
model
with
K-means
clustering
to
analyze
user
behavior
optimize
conversion
rates
on
global
platforms.
extends
traditional
LSTM
models
by
incorporating
multi-dimensional
cell
states,
attention
mechanisms,
improved
memory
capabilities,
enabling
it
effectively
capture
complex
temporal
cross-cultural
patterns.
integration
enhances
process
providing
high-quality
embeddings
that
lead
well-defined
stable
clusters.
Through
comprehensive
evaluations,
combined
approach
demonstrates
superior
performance
across
key
metrics,
including
Silhouette
Score,
Davies-Bouldin
Index
(DBI),
Adjusted
Rand
(ARI),
compared
standalone
algorithms
LSTM-based
methods.
Feature
importance
analysis
further
identifies
coupon
usage,
visit
frequency,
product
category
interest
as
most
influential
factors
purchase
decisions.
findings
highlight
potential
this
methodology
improve
engagement
marketing
strategies
for
Geoenergy Science and Engineering,
Год журнала:
2024,
Номер
242, С. 213240 - 213240
Опубликована: Авг. 16, 2024
The
oil
and
gas
industry
is
changing.The
drive
towards
cleaner
safer
operations
becoming
increasingly
important.Researchers
are
looking
for
more
efficient
accurate
ways
to
detect
faults
that
could
lead
environmental
sustainability
issues.This
study
aims
enhance
the
safety
of
by
improving
existing
artificial
intelligence
approaches
automate
monitoring
detection
malfunctions.This
article
explores
application
deep
neural
networks
anomaly
in
flow
natural
offshore
wells,
proposing
an
innovative
approach
takes
advantage
power
Genetic
Algorithms
Gated
Recurrent
Units
(GRU).The
leveraging
malfunctions.Utilizing
a
comprehensive
dataset
from
3W
Petrobras
project,
which
includes
realtime
data
21
wells
collected
between
2012
2018,
research
focuses
on
detecting
various
anomalies
such
as
abrupt
increases
basic
sediment
water,
spurious
closures
downhole
valves,
severe
slugging,
instability,
rapid
productivity
loss,
quick
restrictions
production
choke,
scaling,
hydrate
formation
lines.The
methodology
integrates
Long
Short-Term
Memory
(LSTM)
GRU
backbones
with
genetic
algorithms
optimise
model
performance.Several
hyperparameter
optimisation
tools
were
explored
innovatively,
focusing
mainly
Algorithms,
it
was
possible
obtain
algorithm
2
stacked
better
comparative
performance
compared
what
reported
literature
producing
F1
equal
0.97.The
findings
demonstrate
potential
AI
improve
real-time
detection,
thereby
reducing
operational
risks
contributing
industry's
transition
greener
practices.It
also
underscores
importance
open
collaborative
efforts
advancing
applications
sector,
aligning
United
Nations'
Sustainable
Development
Goals
mitigate
climate
impact
promote
responsible
consumption
production.
Computers,
Год журнала:
2024,
Номер
13(9), С. 235 - 235
Опубликована: Сен. 17, 2024
This
paper
presents
a
comprehensive
and
comparative
study
of
solar
energy
forecasting
in
Morocco,
utilizing
four
machine
learning
algorithms:
Extreme
Gradient
Boosting
(XGBoost),
Machine
(GBM),
recurrent
neural
networks
(RNNs),
artificial
(ANNs).
The
is
conducted
using
smart
metering
device
designed
for
photovoltaic
system
at
an
industrial
site
Benguerir,
Morocco.
collects
usage
data
from
submeter
transmits
it
to
the
cloud
via
ESP-32
card,
enhancing
monitoring,
efficiency,
utilization.
Our
methodology
includes
analysis
resources,
considering
factors
such
as
location,
temperature,
irradiance
levels,
with
PVSYST
simulation
software
version
7.2,
employed
evaluate
performance
under
varying
conditions.
Additionally,
logger
developed
monitor
panel
production,
securely
storing
while
accurately
measuring
key
parameters
transmitting
them
reliable
communication
protocols.
An
intuitive
web
interface
also
created
visualization
analysis.
research
demonstrates
holistic
approach
devices
systems,
contributing
sustainable
utilization,
grid
development,
environmental
conservation
indicates
that
ANNs
are
most
effective
predictive
model
similar
scenarios,
demonstrating
lowest
RMSE
MAE
values,
along
highest
R2
value.
IEEE Access,
Год журнала:
2024,
Номер
12, С. 57917 - 57946
Опубликована: Янв. 1, 2024
Challenges
in
the
big
data
phenomenon
arise
due
to
existence
of
unstructured
text
data,
which
is
very
large,
comes
from
various
sources,
has
formats,
and
contains
much
noise.
The
complexity
makes
it
difficult
extract
useful
information.
Therefore,
a
process
needed
transform
into
structured
be
processed
further.
information
Extraction
(IE)
helps
relationships,
entities,
semantic
roles,
events
by
converting
them
output.
One
IE's
tasks
Semantic
Role
Labeling
(SRL),
crucial
function
identifying
roles
sentence
so
that
can
enrich
understanding
text.
However,
SRL
development
focuses
on
high-resource
especially
English.
limited
specific
low-resource
languages
or
domains
complex
challenge.
This
research
aims
conduct
systematic
study
for
both
language
domain-specific
contexts.
review
was
carried
out
systematically
using
Preferred
Reporting
Items
Systematic
Reviews
Meta-Analyses
(PRISMA)
model,
54
quality
papers
were
obtained
filtering
(from
2018
2023).
We
several
essential
points,
including
(1)
datasets
are
often
used
their
labeling
strategies
(2)
methods
have
currently
been
developed
learning
scenarios
when
dealing
with
(4)
evaluation
metrics,
(5)
application
tasks.
complemented
discussion
issues
potential
solutions
developing
help
researchers
develop
more
effectively
challenges
faced
data.
Welding International,
Год журнала:
2025,
Номер
unknown, С. 1 - 8
Опубликована: Янв. 2, 2025
In
this
study,
novel
approach
of
employing
a
recurrent
neural
network
(RNN)
model
to
predict
the
shear
strength
aluminium-low
carbon
steel
explosive
clads
is
attempted.
The
cladding
process
parameters,
namely
ratio
(ranging
from
0.6
1.4),
standoff
distance
5.1
9.1
mm),
were
varied
and
detailed
elsewhere.
Due
nonlinear
relationship
between
these
predicting
through
analytical
techniques
becomes
challenging,
making
computational
machine
learning
approaches
more
relevant.
RNN
was
trained
using
experimental
data
preliminary
experiments
in
Python
environment.
performance
then
evaluated
against
remaining
test
data.
demonstrated
high
prediction
accuracy,
achieving
coefficient
determination
(R2)
0.9762,
mean
squared
error
(MSE)
0.7571,
root
(RMSE)
0.87
absolute
(MAE)
0.72.
Advances in healthcare information systems and administration book series,
Год журнала:
2025,
Номер
unknown, С. 535 - 564
Опубликована: Янв. 17, 2025
The
rapid
increase
of
generative
AI
aligned
with
IoMT
promotes
the
medical
and
healthcare
industry
its
numerous
applications
such
as
in
precision
medicine
(PM),
drug
discovery,
disease
diagnosis,
etc.
In
past
decade,
acceptance
remoulded
ongoing
innovations
industry.
As
world
grapples
chronic
illnesses
lifestyle
disorders
burdening
hospitals
clinics,
succors
by
lessening
load.
Nevertheless,
limitations
like
data
security
privacy
are
a
cause
for
worry.
consists
sensor,
or
wearable
connected
via
internet
to
recorder.
This
chapter
attempts
comprehend
goals,
outcome,
future
perspective,
tribulations
alignment
Generative
IoMT.
integration
offers
transformative
outcomes,
including
enhanced
diagnostic
accuracy
through
precise
timely
analyses.
These
promise
better
outcomes
greater
efficiency,
but
it's
essential
prioritize
ethics,
protect
patient
data,
ensure
compliance
they
become
part
everyday
care.
