PeerJ Computer Science,
Journal Year:
2024,
Volume and Issue:
10, P. e2564 - e2564
Published: Dec. 23, 2024
The
complex
environments
and
unpredictable
states
within
transportation
networks
have
a
significant
impact
on
their
operations.
To
enhance
the
level
of
intelligence
in
networks,
we
propose
visual
scene
feature
clustering
analysis
method
based
3D
sensors
adaptive
fuzzy
control
to
address
various
encountered.
Firstly,
construct
extraction
framework
for
scenes
using
employ
series
processing
operators
repair
cracks
noise
images.
Subsequently,
introduce
aggregation
approach
an
algorithm
carefully
screen
preprocessed
features.
Finally,
by
designing
similarity
matrix
network
environment,
obtain
recognition
results
current
environment
state.
Experimental
demonstrate
that
our
outperforms
competitive
approaches
with
mean
average
precision
(mAP)
value
0.776,
serving
as
theoretical
foundation
perception
enhancing
intelligence.
Engineering Research Express,
Journal Year:
2024,
Volume and Issue:
6(3), P. 035209 - 035209
Published: June 28, 2024
Abstract
Arrhythmia,
a
common
cardiovascular
disorder,
refers
to
the
abnormal
electrical
activity
within
heart,
leading
irregular
heart
rhythms.
This
condition
affects
millions
of
people
worldwide,
with
severe
implications
on
cardiac
function
and
overall
health.
Arrhythmias
can
strike
anyone
at
any
age
which
is
significant
cause
morbidity
mortality
global
scale.
About
80%
deaths
related
disease
are
caused
by
ventricular
arrhythmias.
research
investigated
application
an
optimized
multi-objectives
supervised
Machine
Learning
(ML)
models
for
early
arrhythmia
diagnosis.
The
authors
evaluated
model’s
performance
dataset
from
UCI
ML
repository
varying
train-test
splits
(70:30,
80:20,
90:10).
Standard
preprocessing
techniques
such
as
handling
missing
values,
formatting,
balancing,
directory
analysis
were
applied
along
Pearson
correlation
feature
selection,
all
aimed
enhancing
model
performance.
proposed
RF
achieved
impressive
metrics,
including
accuracy
(95.24%),
precision
(100%),
sensitivity
(89.47%),
specificity
(100%).
Furthermore,
study
compared
approach
existing
models,
demonstrating
improvements
across
various
measures.
Scientific Reports,
Journal Year:
2024,
Volume and Issue:
14(1)
Published: May 10, 2024
Abstract
The
challenge
of
developing
an
Android
malware
detection
framework
that
can
identify
in
real-world
apps
is
difficult
for
academicians
and
researchers.
vulnerability
lies
the
permission
model
Android.
Therefore,
it
has
attracted
attention
various
researchers
to
develop
using
or
a
set
permissions.
Academicians
have
used
all
extracted
features
previous
studies,
resulting
overburdening
while
creating
models.
But,
effectiveness
machine
learning
depends
on
relevant
features,
which
help
reducing
value
misclassification
errors
excellent
discriminative
power.
A
feature
selection
proposed
this
research
paper
helps
selecting
features.
In
first
stage
framework,
t
-test,
univariate
logistic
regression
are
implemented
our
collected
data
classify
their
capacity
detecting
malware.
Multivariate
linear
stepwise
forward
correlation
analysis
second
evaluate
correctness
selected
stage.
Furthermore,
as
input
development
models
three
ensemble
methods
neural
network
with
six
different
machine-learning
algorithms.
developed
models’
performance
compared
two
parameters:
F-measure
Accuracy.
experiment
performed
by
half
million
apps.
empirical
findings
reveal
implementing
achieved
higher
rate
set.
Further,
when
previously
frameworks
methodologies,
experimental
results
indicates
study
accuracy
98.8%.
PLoS ONE,
Journal Year:
2025,
Volume and Issue:
20(3), P. e0315545 - e0315545
Published: March 4, 2025
The
multi-objective
supply
chain
needs
a
full
look
at
enterprise
costs,
coordinated
delivery
of
different
products,
and
more
fluidity
efficiency
within
the
network
chain.
However,
existing
methodologies
rarely
delve
into
intricacies
industrial
Therefore,
in
emerging
network,
model
for
problem
was
made
using
meta-heuristic
approach,
specifically
improved
genetic
algorithm,
which
is
type
soft
computing.
To
create
initial
population,
hybrid
approach
that
combines
topology
theory
random
search
method
adopted,
resulted
modification
conventional
single
roulette
wheel
selection
procedure.
Additionally,
crossover
mutation
operations
were
enhanced,
with
determining
their
respective
probabilities
determined
through
fusion
elite
method.
simulation
results
indicate
algorithm
reduced
load
from
0.678
to
0.535,
labor
costs
1832
yuan
1790
yuan,
operational
time
by
approximately
39.5%,
48
seconds
29.5
seconds.
variation
node
utilization
rates
significantly
decreased
30.1%
12.25%,
markedly
enhancing
resource
scheduling
overall
balance
