Sensors,
Год журнала:
2023,
Номер
23(15), С. 6664 - 6664
Опубликована: Июль 25, 2023
Contemporary
advancements
in
wearable
equipment
have
generated
interest
continuously
observing
stress
utilizing
various
physiological
indicators.
Early
detection
can
improve
healthcare
by
lessening
the
negative
effects
of
chronic
stress.
Machine
learning
(ML)
methodologies
been
modified
for
to
monitor
user
health
situations
sufficient
information.
Nevertheless,
more
data
are
needed
make
applying
Artificial
Intelligence
(AI)
medical
field
easier.
This
research
aimed
detect
using
a
stacking
model
based
on
machine
algorithms
chest-based
features
from
Wearable
Stress
and
Affect
Detection
(WESAD)
dataset.
We
converted
this
natural
dataset
into
convenient
format
suggested
performing
visualization
preprocessing
RESP
feature
analysis
Z-score,
SelectKBest
feature,
Synthetic
Minority
Over-Sampling
Technique
(SMOTE),
normalization.
The
efficiency
proposed
was
estimated
regarding
accuracy,
precision,
recall,
F1-score.
experimental
outcome
illustrated
efficacy
technique,
achieving
0.99%
accuracy.
results
revealed
that
methodology
performed
better
than
traditional
previous
studies.
Intelligent
Speech
Technology
(IST)
is
revolutionizing
healthcare
by
enhancing
transcription
accuracy,
disease
diagnosis,
and
medical
equipment
control
in
smart
hospital
environments.
This
study
introduces
an
innovative
approach
employing
federated
learning
with
Multi-Layer
Perceptron
(MLP)
Gated
Recurrent
Unit
(GRU)
neural
networks
to
improve
IST
performance.
Leveraging
the
“Medical
Speech,
Transcription,
Intent”
dataset
from
Kaggle,
comprising
a
variety
of
speech
recordings
corresponding
symptom
labels,
noise
reduction
was
applied
using
Wiener
filter
audio
quality.
Feature
extraction
through
MLP
sequence
classification
GRU
highlighted
model’s
robustness
capacity
for
detailed
understanding.
The
framework
enabled
collaborative
model
training
across
multiple
sites,
preserving
patient
privacy
avoiding
raw
data
exchange.
distributed
allowed
learn
diverse,
real-world
while
ensuring
compliance
strict
protection
standards.
Through
rigorous
five-fold
cross-validation,
proposed
Fed
MLP-GRU
demonstrated
accuracy
98.6%,
consistently
high
sensitivity
specificity,
highlighting
its
reliable
generalization
test
conditions.
In
real-time
applications,
effectively
performed
transcription,
provided
symptom-based
diagnostic
insights,
facilitated
hands-free
equipment,
reducing
contamination
risks
workflow
efficiency.
These
findings
indicate
that
IST,
powered
networks,
can
significantly
delivery,
operational
efficiency
clinical
settings.
research
underscores
transformative
potential
advanced
addressing
pressing
challenges
modern
setting
stage
future
innovations
intelligent
technology.
Sensors,
Год журнала:
2023,
Номер
23(15), С. 6664 - 6664
Опубликована: Июль 25, 2023
Contemporary
advancements
in
wearable
equipment
have
generated
interest
continuously
observing
stress
utilizing
various
physiological
indicators.
Early
detection
can
improve
healthcare
by
lessening
the
negative
effects
of
chronic
stress.
Machine
learning
(ML)
methodologies
been
modified
for
to
monitor
user
health
situations
sufficient
information.
Nevertheless,
more
data
are
needed
make
applying
Artificial
Intelligence
(AI)
medical
field
easier.
This
research
aimed
detect
using
a
stacking
model
based
on
machine
algorithms
chest-based
features
from
Wearable
Stress
and
Affect
Detection
(WESAD)
dataset.
We
converted
this
natural
dataset
into
convenient
format
suggested
performing
visualization
preprocessing
RESP
feature
analysis
Z-score,
SelectKBest
feature,
Synthetic
Minority
Over-Sampling
Technique
(SMOTE),
normalization.
The
efficiency
proposed
was
estimated
regarding
accuracy,
precision,
recall,
F1-score.
experimental
outcome
illustrated
efficacy
technique,
achieving
0.99%
accuracy.
results
revealed
that
methodology
performed
better
than
traditional
previous
studies.