Transportmetrica B Transport Dynamics,
Год журнала:
2024,
Номер
13(1)
Опубликована: Дек. 3, 2024
Existing
research
faces
challenges
in
accurately
predicting
crashes
due
to
the
unreasonable
selection
of
spatial
units,
biased
crash
data
collection,
and
insufficient
integration
multi-source
data.
To
address
these
issues,
Graph
Neural
Networks
(GNNs)
for
node
classification
are
employed
predict
at
macroscopic
road
level.
Crash
alarm
incorporated
as
a
supplement
official
archive
ensure
spatial–temporal
distribution's
authenticity
mitigate
sparsity.
Additionally,
traffic
violation
included
feature
enrich
risk
information.
Finally,
multi-graph
deep
learning
framework
(STCM-GCN)
with
spatial,
temporal,
modules
has
been
developed.
Data
from
Shenzhen,
China,
demonstrates
that
STCM-GCN
outperforms
baseline
models
reasonable
structure.
The
inclusion
violations
contributes
performance
improvement.
model
exhibits
robustness,
analysis
computational
efficiency
provides
comprehensive
insights
into
model's
capabilities.
Machine
learning
has
demonstrated
potential
in
addressing
complex
nonlinear
changes
risk
assessment.
However,
further
exploration
is
needed
to
enhance
model
interpretability
and
optimize
performance.
Therefore,
this
study
aims
develop
a
novel
workplace
assessment
framework.
By
utilizing
the
SHapley
Additive
exPlanations
(SHAP)
analysis
method
ensemble
algorithms,
framework
maps
characteristic
attributes
levels.
Reliability
validation
of
critical
attribute
components
are
conducted
using
accidents
Chinese
coal
enterprises
as
case
study,
which
represents
one
most
serious
occupational
hazards.
The
results
indicate
that
issues
algorithms
yields
capable
accurately
assessing
understanding
decision-making
processes.
Comparative
experiments
show
achieves
an
accuracy
up
98.3%,
confirming
its
robust
outcomes
SHAP
for
feature
importance
facilitate
identification
explain
causal
relationships
leading
risk-level
findings.
This
provides
valuable
accident
prevention
strategies
minimize
injuries
losses.
Buildings,
Год журнала:
2025,
Номер
15(6), С. 847 - 847
Опубликована: Март 7, 2025
Although
safety
technology
has
recently
been
shown
to
prevent
occupational
incidents,
a
systematic
approach
identifying
technological
opportunities
is
still
lacking.
Incident
report
documents,
containing
large
volumes
of
narrative
text,
are
considered
valuable
resources
for
predetermining
incident
factors.
Additionally,
patent
data,
as
form
big
data
from
sources,
widely
utilized
explore
potential
solutions.
In
this
context,
study
aims
identify
by
integrating
two
types
textual
data:
documents
and
documents.
Text
mining
self-organizingmaps
employed
discover
applicable
technologies
prevention,
grouping
them
into
five
categories,
follows:
machine
tool
work,
high-place
vehicle-related
facilities,
hydraulic
machines,
miscellaneous
tools.
A
gap
analysis
between
incidents
patents
also
conducted
assess
feasibility
develop
strategy.
The
findings,
derived
both
provide
solutions
that
essential
improving
workplace
can
be
used
business
owners
managers.
Construction Innovation,
Год журнала:
2025,
Номер
unknown
Опубликована: Апрель 8, 2025
Purpose
This
study
aims
to
investigate
the
moderating
effects
of
internet
things
(IoT)
and
wearable
technologies
(WT)
on
relationship
between
traditional
safety
practices
(TSP)
management
(SM)
outcomes
in
Shanghai’s
construction
sector.
It
examines
how
these
enhance
performance
by
addressing
limitations
conventional
approaches.
Design/methodology/approach
A
survey
300
professionals,
including
project
managers,
site
managers
officers,
was
conducted
Shanghai.
Data
analysis
using
partial
least
squares
structural
equation
modelling
(PLS-SEM)
assessed
IoT
WT
SM
outcomes.
Findings
The
results
indicate
that
has
a
stronger
effect
(
ß
=
0.21,
p
<
0.01)
than
0.11,
0.07).
offers
immediate
benefits
through
real-time
worker
monitoring,
whereas
enhances
long-term
enabling
predictive
analytics
hazard
detection.
highlights
synergy
TSP
improving
Practical
implications
While
both
practices,
their
impacts
differ.
significantly
improves
safety,
making
it
essential
for
high-risk
zones,
contributes
risk
mitigation
data-driven
insights.
Construction
should
prioritise
adoption
improvements
while
integrating
IoT-driven
models
sustained
prevention.
Originality/value
provides
empirical
evidence
complementary
roles
enhancing
construction.
valuable
insights
into
digital
transformation’s
role
performance.
The Canadian Journal of Chemical Engineering,
Год журнала:
2024,
Номер
102(12), С. 4281 - 4296
Опубликована: Май 28, 2024
Abstract
The
evolution
of
hazardous
chemical
accidents
(HCAs)
is
characterized
by
uncertainty
and
complexity.
It
challenging
for
decision‐makers
to
expeditiously
adapt
emergency
response
plans
in
dynamically
changing
scenario
states.
This
study
proposes
a
data‐driven
methodology
constructing
accident
scenarios
develops
novel
hybrid
deep
learning
model
deduction
analysis.
aids
accurately
predicting
the
HCAs,
enabling
responders
prepare
implement
targeted
interventions
proactively.
First,
framework
an
database
presented,
based
on
time‐sequential
characteristics
progression.
employs
approach
describe
process
scenarios.
Second,
(CNN‐LSTM‐Attention)
that
integrates
convolutional
neural
network
(CNN),
long
short‐term
memory
(LSTM),
attention
mechanism
(AM)
developed
Finally,
illustrate
practical
application,
HCAs
established.
A
major
HCA
case
conducted
demonstrate
ability
this
analyze
various
scenarios,
thereby
improving
decision‐making
efficiency.
Compared
with
algorithms
such
as
CNN,
LSTM,
CNN‐LSTM,
prediction
accuracy
method
ranges
from
86%
93%,
signifying
improvement
over
7%.
work
provides
reliable
supporting
management.
