Frontiers in Physics,
Journal Year:
2024,
Volume and Issue:
12
Published: Oct. 14, 2024
Introduction
As
5G
networks
become
widespread
and
their
application
scenarios
expand,
massive
amounts
of
traffic
data
are
continuously
generated.
Properly
analyzing
this
is
crucial
for
enhancing
services.
Methods
This
paper
uses
the
visibility
graph
method
to
convert
into
a
network,
conducting
feature
analysis
data.
Using
AfreecaTV
dataset
as
research
object,
constructs
at
different
scales
observes
evolution
degree
distribution
with
varying
volumes.
The
employs
Hurst
index
evaluate
network
community
detection
study
converted
from
applications.
Results
Experimental
results
reveal
significant
differences
in
node
topological
structures
across
scenarios,
such
star
multiple
subnetwork
structures.
It
found
that
exhibits
heterogeneity,
reflecting
uneven
growth
degrees
during
expansion.
discovers
retains
long-term
dependence
trends
original
Through
detection,
it
observed
applications
exhibit
diverse
structures,
high
centrality
nodes,
star-like
modularity,
multilayer
characteristics.
Discussion
These
findings
indicate
complex
heterogeneity
reflect
imbalance
connection
methods
show
inherits
data,
providing
basis
dynamic
characteristics
network.
inherent
modularity
hierarchy
which
helps
understand
performance
optimization
directions
Scientific Reports,
Journal Year:
2025,
Volume and Issue:
15(1)
Published: Feb. 10, 2025
Exploiting
complex
network
methods
to
describe
dynamical
behavior
based
on
speech
time
series
can
provide
fundamental
insights
into
the
function
of
underlying
processes
in
Alzheimer's
disease
(AD).
This
study
scrutinizes
dynamic
alterations
through
abstract
concepts
small-world
networks.
The
visibility
graph
(VG)
spontaneous
is
introduced
as
a
quantitative
method
differentiate
between
healthy
individuals
and
those
with
Alzheimer's.
patterns
across
three
AD
subjects
stages
are
analyzed
by
examining
feature
structure,
characterized
high
clustering
coefficient
(C)
short
average
path
length
(L)
VG.
These
characteristics
calculated
degree
K.
results
demonstrate
practical
utility
C
L
identifying
pathological
mechanisms
AD.
Furthermore,
all
exhibit
topology
VG,
changes
reflecting
brain
system's
pathology
that
impacts
individuals'
language
skills.
Gazi Üniversitesi Fen Bilimleri Dergisi Part C Tasarım ve Teknoloji,
Journal Year:
2024,
Volume and Issue:
12(1), P. 257 - 266
Published: March 11, 2024
Epilepsy
is
a
neurological
disorder
in
which
involuntary
contractions,
sensory
abnormalities,
and
changes
occur
as
result
of
abrupt
uncontrolled
discharges
the
neurons
brain,
disrupt
systems
regulated
by
brain.
In
epilepsy,
abnormal
electrical
impulses
from
cells
various
brain
areas
are
noticed.
The
accurate
interpretation
these
critical
illness
diagnosis.
This
study
aims
to
use
different
machine-learning
algorithms
diagnose
epileptic
seizures.
frequency
components
EEG
data
were
extracted
using
parametric
approaches.
feature
extraction
approach
was
fed
into
machine
learning
classification
algorithms,
including
Artificial
Neural
Network
(ANN),
Gradient
Boosting,
Random
Forest.
ANN
classifier
shown
have
most
significant
test
performance
this
investigation,
with
roughly
97%
accuracy
91%
F1
score
recognizing
episodes.
Boosting
classifier,
on
other
hand,
performed
similarly
ANN,
96%
93%
score.
IET Science Measurement & Technology,
Journal Year:
2025,
Volume and Issue:
19(1)
Published: Jan. 1, 2025
ABSTRACT
Arc
faults
in
low‐voltage
distribution
networks
significantly
threaten
power
system
safety
due
to
their
randomness
and
concealment.
Traditional
arc
fault
detection
methods,
which
rely
on
time‐domain
frequency‐domain
features,
often
struggle
with
accuracy
robustness
variable
load
environments.
To
address
these
challenges,
this
paper
introduces
Visibility
Graph
Convolutional
Learning
(VisGCL),
a
novel
approach
that
segments
current
signals
into
visibility
graphs
employs
hierarchical
graph
convolutional
for
analysis.
This
method
directly
learns
failure
modes
from
the
graphical
representation
of
signals,
simplifying
process
enhancing
both
robustness.
Experimental
results
demonstrate
proposed
achieves
an
98.58
±
0.14%,
precision,
recall,
F1‐score
reaching
98.05
0.25%,
98.36
0.47%,
98.16
0.23%,
respectively.
Extensive
experiments
validate
effectiveness
VisGCL,
confirming
its
superiority
detecting
under
diverse
conditions,
offering
new
efficient
reliable
solution
networks.
