The moderating influence of safety on green space’s health benefits in Chinese urban communities
Jia Chen,
No information about this author
Longfeng Wu,
No information about this author
Han Ma
No information about this author
et al.
Journal of Environmental Management,
Journal Year:
2025,
Volume and Issue:
375, P. 124232 - 124232
Published: Jan. 25, 2025
Language: Английский
Comparing XAI techniques for interpreting short-term burglary predictions at micro-places
Robin Khalfa,
No information about this author
Naomi Theinert,
No information about this author
Wim Hardyns
No information about this author
et al.
Computational Urban Science,
Journal Year:
2025,
Volume and Issue:
5(1)
Published: May 9, 2025
Abstract
This
study
empirically
compares
multiple
eXplainable
Artificial
Intelligence
(XAI)
techniques
to
interpret
short-term
(weekly)
machine
learning-based
burglary
predictions
at
the
micro-place
level
in
Ghent,
Belgium.
While
previous
research
predominantly
relies
on
SHAP
spatiotemporal
crime
predictions,
this
is
first
systematically
evaluate
alongside
other
XAI
techniques,
offering
both
global
and
local
model
interpretability
within
context
of
prediction.
Using
data
from
2014
2018
residential
burglary,
repeat
near-repeat
victimization,
environmental
features,
socio-demographic
indicators,
seasonal
effects,
we
trained
an
XGBoost
with
76
features
predict
weekly
hot
spots.
serves
as
a
basis
for
comparing
interpretative
power
different
techniques.
Our
results
show
that
built
environment
land
use
characteristics
are
most
consistent
predictors
risk.
However,
their
influence
varies
substantially
level,
revealing
importance
spatial
context.
feature
rankings
broadly
aligned
across
explanations,
especially
between
LIME,
often
diverge.
These
discrepancies
highlight
need
careful
method
selection
when
translating
into
prevention
strategies.
In
addition,
demonstrates
risks
influenced
by
complex
interactions
threshold
effects
social
disorganization
features.
We
these
findings
through
lens
criminological
theory,
argue
more
integrated
approaches
go
beyond
examining
isolated
specific
predictors.
Finally,
call
greater
attention
methodological
implications
arise
applying
particularly
learning
outputs
used
inform
policy
decisions.
Language: Английский
Neighborhood social environment and mental health of older adults in China: the mediating role of subjective well-being and the moderating role of green space
Tzu-Ming Lin,
No information about this author
Qianhui Wang,
No information about this author
Zixuan Tan
No information about this author
et al.
Frontiers in Public Health,
Journal Year:
2024,
Volume and Issue:
12
Published: Dec. 6, 2024
With
the
continuous
development
of
global
aging
trend,
mental
health
older
adults
has
been
a
concern
by
world.
The
living
space
is
limited
due
to
decline
their
activity
function.
Neighborhood
environment,
especially
neighborhood
social
become
an
important
factor
affecting
adults.
Therefore,
this
study
explores
mechanism
that
influences
environment
and
adults,
mediating
effect
subjective
well-being
(SWB),
moderating
green
space.
Language: Английский
Analyzing Urban Crime Through Street View Imagery: Insights from Urban Micro Built Environment and Perceptions
Urban Science,
Journal Year:
2024,
Volume and Issue:
8(4), P. 247 - 247
Published: Dec. 7, 2024
Understanding
the
relationship
between
urban
crime
and
built
environment
is
crucial
for
developing
effective
prevention
strategies,
particularly
in
context
of
rapid
development
city
planning.
As
cities
grow,
urbanization
leads
to
environments
that
either
promote
or
inhibit
criminal
activity,
making
it
essential
explore
interactions
design
crime.
This
study
investigates
impact
micro
(MBE)
elements
place
perceptions
on
occurrences
Toronto
using
street
view
imagery
(SVI)
data
machine
learning
models.
We
used
logistic
regression
models
an
XGBoost
(Version
1.7.5)
classifier
assess
significance
MBE
perception
variables
classifying
non-crime
intersections.
Our
findings
reveal
intersections
with
activity
tend
be
related
more
mobility-related
features,
such
as
roads
vehicles,
fewer
natural
elements,
vegetation.
The
“beautiful”
“depressing”
emerged
most
significant
explaining
events,
surpassing
commonly
studied
“safety”
perception.
model
achieved
86%
accuracy,
indicating
are
strong
predictors
risk.
These
suggest
enhancing
vegetation
improving
aesthetics
could
serve
measures
environments.
However,
limitations
include
general
nature
reliance
aggregated
data.
Future
research
should
incorporate
local
fine-scale
provide
tailored
insights
planning
Language: Английский