2022 5th International Conference on Data Science and Information Technology (DSIT),
Год журнала:
2023,
Номер
unknown, С. 272 - 276
Опубликована: Июль 28, 2023
Using
rough
set
theory,
we
can
achieve
effective
prediction
of
online
public
opinion,
improve
the
scientific
and
timely
decision-making
provide
strong
bases
for
opinion
early
warning
management
relevant
departments.
In
this
paper,
taking
"Henan
rainstorm"
incident
as
an
example,
use
classical
theory
to
filter
mine
microblog
data,
build
a
decision
information
system,
establish
model
attribute
simplification,
obtain
difference
matrix,
then
required
(key
attributes
warning),
when
departments
monitor
entries
containing
key
warning,
they
should
attributes,
pay
attention
them
immediately
implement
corresponding
measures.
The
actual
case
shows
that
final
reduced
obtained
by
has
only
4%
from
expert
decisions
be
effectively
applied
opinion.
Purpose
The
metaverse,
which
is
now
revolutionizing
how
brands
strategize
their
business
needs,
necessitates
understanding
individual
opinions.
Sentiment
analysis
deciphers
emotions
and
uncovers
a
deeper
of
user
opinions
trends
within
this
digital
realm.
Further,
sentiments
signify
the
underlying
factor
that
triggers
one’s
intent
to
use
technology
like
metaverse.
Positive
often
correlate
with
positive
experiences,
while
negative
may
issues
or
frustrations.
Brands
consider
these
implement
them
on
metaverse
platforms
for
seamless
experience.
Design/methodology/approach
current
study
adopts
machine
learning
sentiment
techniques
using
Support
Vector
Machine,
Doc2Vec,
RNN,
CNN
explore
individuals
toward
in
user-generated
context.
topics
were
discovered
topic
modeling
method,
was
performed
subsequently.
Findings
results
revealed
users
had
notion
about
experience
orientation
having
attitude
towards
economy,
data,
cyber
security.
accuracy
each
model
has
been
analyzed,
it
concluded
provides
better
an
average
89%
compared
other
models.
Research
limitations/implications
Analyzing
can
reveal
general
public
perceives
suggest
enthusiasm
readiness
adoption,
might
indicate
skepticism
concerns.
Given
notions
metaverse’s
orientation,
developers
should
continue
focus
creating
innovative
immersive
virtual
environments.
At
same
time,
users'
concerns
cybersecurity
economy
are
critical.
suggests
need
innovation
economic
models
Also,
platform
operators
prioritize
robust
data
security
measures.
Implementing
strong
encryption
two-factor
authentication
educating
best
practices
address
enhance
trust.
Social
implications
In
terms
societal
dynamics,
could
revolutionize
communication
relationships
by
altering
traditional
proximity
presence
its
users.
economies
emerge,
assets
real-world
value,
presenting
both
opportunities
challenges
industries
regulators.
Originality/value
contributes
research
as
first
kind
deep
evaluate
Social
networks
accelerate
information
communication
in
public
health
emergencies.
Some
negative
may
cause
an
outbreak
of
opinion
crisis.
Accurately
predicting
online
trends
can
help
the
relevant
departments
take
timely
and
effective
measures
to
cope
with
risks.
Therefore,
this
research
proposes
a
prediction
model
incorporating
swarm
intelligence
optimization
algorithm
deep
learning
method.
In
model,
we
improve
Harris
Hawks
Optimization
(HHO)
by
introducing
Cauchy
distribution
function,
stochastic
contraction
exponential
adaptive
inertia
weight.
Then
utilize
improved
HHO
(IHHO)
optimize
hyperparameters
method
LSTM,
including
rate
number
neurons
hidden
layer.
Finally,
construct
IHHO-LSTM
make
predictions
three
The
experiments
verify
that
proposed
outperforms
other
single
hybrid
models.
MAPE
values
reduce
78.34%,
54.46%,
46.42%
relative
average
Compared
mean
two
models,
decrease
47.69%,
18.45%,
5.78%.
be
applied
early
warning
reversal
identification,
providing
reference
management.
Scientific Reports,
Год журнала:
2025,
Номер
15(1)
Опубликована: Март 6, 2025
Public
opinion
on
technological
disasters
is
influenced
by
unique
factors
and
characteristics.
Based
the
infectious
disease
model,
this
paper
develops
a
public
dissemination
model
for
disasters,
considering
such
as
disaster
severity,
government
response,
accountability,
impact
of
both
positive
negative
media
content.
Using
differential
equation
stability
theory,
we
analyze
existence
free
propagation
equilibrium
point
point.
The
next-generation
matrix
method
applied
to
calculate
threshold,
revealing
that
accountability
are
key
in
spread
opinion.
Sensitivity
analyses
examine
how
these
affect
dynamics.
A
case
study
Shiyan
gas
explosion
Hubei
Province
presented,
with
microblog
data
used
parameters.
proposed
compared
two
other
models,
demonstrating
viability
effectiveness
developed
model.
also
show
well-handled
responses
can
help
calm
opinion,
even
cases
where
lacking.
Finally,
policy
suggestions
offered
enhance
management
during
disasters.
Detecting
fake
news
and
missing
information
is
gaining
popularity,
especially
after
social
media
online
platforms
advancements.
Social
the
main
speediest
source
of
propagation,
whereas
websites
contribute
to
dissipation.
In
this
study,
we
propose
a
framework
detect
using
temporal
features
text
consider
user
feedback
determine
whether
or
not.
recent
studies,
in
documents
gain
valuable
consideration
from
Natural
Language
Processing
only
try
classify
textual
data
as
true.
This
research
article
indicates
impact
recurring
non-recurring
events
on
true
news.
We
use
different
models
such
LSTM,
BERT,
CNN-
BiLSTM
investigate,
it
concluded
that
get
better
results,
70%
recurring,
rest
30%
non-recurring.
