Applied Sciences,
Год журнала:
2024,
Номер
14(24), С. 11628 - 11628
Опубликована: Дек. 12, 2024
Telecom
fraud
has
emerged
as
one
of
the
most
pressing
challenges
in
criminal
field.
With
advancements
artificial
intelligence,
telecom
texts
have
become
increasingly
covert
and
deceptive.
Existing
prevention
methods,
such
mobile
number
tracking,
detection,
traditional
machine-learning-based
text
recognition,
struggle
terms
their
real-time
performance
identifying
fraud.
Additionally,
scarcity
Chinese
data
limited
research
this
area.
In
paper,
we
propose
a
detection
model,
RoBERTa-MHARC,
which
combines
RoBERTa
with
multi-head
attention
mechanism
residual
connections.
First,
model
selects
categories
from
CCL2023
dataset
basic
samples
merges
them
collected
data,
creating
five-category
covering
impersonation
customer
service,
leadership
acquaintances,
loans,
public
security
fraud,
normal
text.
During
training,
integrates
enhances
its
training
efficiency
through
Finally,
improves
multi-class
classification
accuracy
by
incorporating
an
inconsistency
loss
function
alongside
cross-entropy
loss.
The
experimental
results
demonstrate
that
our
performs
well
on
multiple
benchmark
datasets,
achieving
F1
score
97.65
FBS
dataset,
98.10
own
93.69
news
dataset.
Technology and Health Care,
Год журнала:
2024,
Номер
32(6), С. 3801 - 3813
Опубликована: Авг. 2, 2024
Schwann
cell
sheaths
are
the
source
of
benign,
slowly
expanding
tumours
known
as
acoustic
neuromas
(AN).
The
diagnostic
and
treatment
approaches
for
AN
must
be
patient-centered,
taking
into
account
unique
factors
preferences.
Cybernetics and Computer Technologies,
Год журнала:
2024,
Номер
4, С. 110 - 120
Опубликована: Дек. 18, 2024
Introduction.
The
ability
to
automate
processes
is
a
key
aspect
of
modern
information
technology.
construction
and
use
the
conceptual
structure
knowledge
base
becoming
an
urgent
need
in
world,
where
amount
growing
exponentially.
processes,
including
ontologies,
which
requires
extraction
from
full-text
sources
their
automatic
structuring,
important.
Knowledge
bases
are
used
manage
complex
dynamic
systems
by
ensuring
storage,
organization,
access
large
that
allows
for
effective
analysis
prediction
behavior
such
systems.
purpose
paper.
paper
demonstrate
effectiveness
using
deep
learning
methods
formation
base.
study
also
aims
show
how
integration
with
can
improve
quality
forecasts
increase
efficiency
rehabilitation
trajectory
management.
Results.
algorithm
successfully
extracted
processed
symptom
medical
cases,
effectively
handling
duplicates
synonyms.
utilization
cosine
similarity
enabled
identification
synonymous
symptoms
within
established
base,
facilitating
seamless
new
while
preventing
redundancy.
system
demonstrated
its
capability
discern
should
be
incorporated
into
omitted
based
on
existing
entries.
outcomes
underscore
potential
this
automated
approach
enhance
contribute
refinement
predictive
models
healthcare
domain.
Conclusions.
automating
enhances
filling
comprehensiveness
crucial
building
patient
trajectories
improving
decision
support.
Keywords:
Knowledge-Oriented
Management
Systems,
Support
Vector
Machine,
Word2Vec,
Skip-Gram,
BioBERT.
BIO Web of Conferences,
Год журнала:
2024,
Номер
146, С. 01041 - 01041
Опубликована: Янв. 1, 2024
This
research
explores
the
impact
of
integrating
Bidirectional
Encoder
Representations
from
Transformers
(BERT)
into
Retrieval
Hadith
Information
(RoHI)
application
within
realm
religious
education
media.
Hadith,
sayings
and
actions
Prophet
Muhammad,
play
a
pivotal
role
in
Islamic
teachings,
requiring
accurate
contextually
relevant
retrieval
for
educational
purposes.
RoHI,
designed
to
enhance
access
comprehension
literature,
employs
BERT's
advanced
natural
language
processing
capabilities.
The
study
assesses
how
BERT-enhanced
RoHI
facilitates
efficient
interpretation
texts.
By
leveraging
ability
capture
intricate
patterns
semantics,
aims
precision
contextual
appropriateness
retrieved
information.
also
discusses
implications
digital
learning
platforms,
emphasizing
potential
NLP
technologies
foster
broader
knowledge
promote
inclusive
practices.
contributes
field
by
proposing
framework
that
integrates
AI
techniques
with
education,
ensuring
learners
receive
meaningful
information
tailored
their
needs.
findings
highlight
BERT
revolutionizing
processes
studies,
paving
way
more
effective
tools
resources
environments.
