Advancing Water Quality Assessment and Prediction Using Machine Learning Models, Coupled with Explainable Artificial Intelligence (XAI) Techniques Like Shapley Additive Explanations (SHAP) For Interpreting the Black-Box Nature
Results in Engineering,
Год журнала:
2024,
Номер
23, С. 102831 - 102831
Опубликована: Сен. 1, 2024
Water
quality
assessment
and
prediction
play
crucial
roles
in
ensuring
the
sustainability
safety
of
freshwater
resources.
This
study
aims
to
enhance
water
by
integrating
advanced
machine
learning
models
with
XAI
techniques.
Traditional
methods,
such
as
index,
often
require
extensive
data
collection
laboratory
analysis,
making
them
resource-intensive.
The
weighted
arithmetic
index
is
employed
alongside
models,
specifically
RF,
LightGBM,
XGBoost,
predict
quality.
models'
performance
was
evaluated
using
metrics
MAE,
RMSE,
R2,
R.
results
demonstrated
high
predictive
accuracy,
XGBoost
showing
best
(R2
=
0.992,
R
0.996,
MAE
0.825,
RMSE
1.381).
Additionally,
SHAP
were
used
interpret
model's
predictions,
revealing
that
COD
BOD
are
most
influential
factors
determining
quality,
while
electrical
conductivity,
chloride,
nitrate
had
minimal
impact.
High
dissolved
oxygen
levels
associated
lower
indicative
excellent
pH
consistently
influenced
predictions.
findings
suggest
proposed
approach
offers
a
reliable
interpretable
method
for
prediction,
which
can
significantly
benefit
specialists
decision-makers.
Язык: Английский
Current methods in explainable artificial intelligence and future prospects for integrative physiology
Pflügers Archiv - European Journal of Physiology,
Год журнала:
2025,
Номер
unknown
Опубликована: Фев. 25, 2025
Abstract
Explainable
artificial
intelligence
(XAI)
is
gaining
importance
in
physiological
research,
where
now
used
as
an
analytical
and
predictive
tool
for
many
medical
research
questions.
The
primary
goal
of
XAI
to
make
AI
models
understandable
human
decision-makers.
This
can
be
achieved
particular
through
providing
inherently
interpretable
methods
or
by
making
opaque
their
outputs
transparent
using
post
hoc
explanations.
review
introduces
core
topics
provides
a
selective
overview
current
physiology.
It
further
illustrates
solved
discusses
open
challenges
existing
practical
examples
from
the
field.
article
gives
outlook
on
two
possible
future
prospects:
(1)
provide
trustworthy
integrative
(2)
integrating
expertise
about
explanation
into
method
development
useful
beneficial
human-AI
partnerships.
Язык: Английский
Ensemble deep learning model for protein secondary structure prediction using NLP metrics and explainable AI
Results in Engineering,
Год журнала:
2024,
Номер
unknown, С. 103435 - 103435
Опубликована: Ноя. 1, 2024
Язык: Английский
Harnessing explainable artificial intelligence (XAI) for enhanced geopolymer concrete mix optimization
Results in Engineering,
Год журнала:
2024,
Номер
unknown, С. 103036 - 103036
Опубликована: Сен. 1, 2024
Язык: Английский
Machine Learning Model Construction and Practice for Personalized Training Programs in Athletics Training
Applied Mathematics and Nonlinear Sciences,
Год журнала:
2025,
Номер
10(1)
Опубликована: Янв. 1, 2025
Abstract
Big
data
science
is
a
complexity
produced
in
the
new
era,
and
machine
learning
models
belong
to
its
main
branch,
which
has
characteristic
methodological
features
provides
ideas
scientifically
solve
personalized
formulation
of
training
programs
track
field
training.
In
this
paper,
firstly,
athletes’
sports
are
collected
by
installing
sensors
key
parts
athletes,
then
real-time
state
estimation
given
Kalman
filtering,
optimized
microelectromechanical
technology.
The
obtained
solution
results
inputted
into
important
movement
joint
model
human
body
so
as
realize
motion
capture
athletes.
Based
on
this,
for
been
constructed
using
an
ant
colony
algorithm.
generation
plan
varied
optimization
problem
with
constraints,
containing
discrete
continuous
variables.
Then,
method
adaptation
evaluation
constraints
updating
related
solutions
were
proposed,
thus
completing
construction
model.
experimental
group
improved
much
more
events
than
control
group,
24.96%
girls’
shot
put.
It
shows
that
program
developed
through
based
line
different
students’
own
needs,
generated
can
provide
athletes
efficient
guidance,
verifies
effectiveness
paper
practice
design
experiment.
Язык: Английский
A Deep Learning Approach Based on Interpretable Feature Importance for Predicting Sports Results
International Journal of Computer Science in Sport,
Год журнала:
2025,
Номер
24(1), С. 56 - 72
Опубликована: Фев. 1, 2025
Abstract
Football
match
result
prediction
is
a
challenging
task
that
has
been
the
subject
of
much
research.
Traditionally,
predictions
have
made
by
team
managers,
fans,
and
analysts
based
on
their
knowledge
experience.
However
recently
there
an
increased
interest
in
predicting
outcomes
using
statistical
techniques
machine
learning.
These
algorithms
can
learn
from
historical
data
to
identify
complex
relationships
between
different
variables,
then
make
about
outcome
future
matches.
Accordingly,
forecasting
plays
pivotal
role
assisting
managers
clubs
making
well-informed
decisions
geared
toward
securing
victories
leagues
tournaments.
In
this
paper,
we
presented
approach,
which
generally
applicable
all
areas
sports,
forecast
football
results
three
stages.
The
first
stage
involves
identifying
collecting
occurred
events
during
match.
As
multiclass
classification
problem
with
classes,
each
possible
outcomes.
Then,
applied
multiple
learning
compare
performance
those
models,
choose
one
performs
best.
final
step,
study
goes
through
critical
aspect
model
interpretability.
We
used
SHapley
Additive
exPlanations
(SHAP)
method
decipher
feature
importance
within
our
best
model,
focusing
factors
influence
predictions.
Experiment
indicate
Multilayer
Perceptron
(MLP),
neural
network
algorithm,
was
effective
when
compared
various
other
models
produced
competitive
prior
works.
MLP
achieved
0.8342
for
accuracy.
particular
significance
lies
use
SHAP
explain
model.
Specifically,
exploiting
its
graphical
representation
illustrate
dataset
Язык: Английский
Implementing machine learning algorithms to optimize sprint performance and biomechanical analysis of track and field athletes
Journal of Computational Methods in Sciences and Engineering,
Год журнала:
2025,
Номер
unknown
Опубликована: Май 9, 2025
Sprint
performance
is
a
crucial
component
of
athletic
performance,
especially
in
sports
like
track
and
field,
football,
rugby,
which
require
quick
bursts
peak
effort
over
short
durations.
Understanding
the
biomechanics
sprinting
essential
for
enhancing
preventing
injuries,
creating
effective
training
plans.
Traditional
research
on
sprint
evaluation
often
focuses
discrete
measures
while
neglecting
intricate
interactions
between
variables
that
evolve
throughout
sprint.
This
study
addresses
these
challenges
by
applying
machine
learning
(ML)
algorithm,
specifically
Polar
Bear-tuned
Multi-Source
Kernel
Support
Vector
Machine
(PB-MKSVM),
to
predict
optimize
field
athletes.
The
system
analyzes
biomechanical
characteristics
such
as
muscle
activation
patterns,
joint
angles,
ground
reaction
forces,
stride
length.
Data
were
collected
using
wearable
sensors
motion
capture
systems
during
standardized
trials,
various
parameters
recorded.
Standard
preprocessing
steps
including
noise
removal
outlier
detection
applied
data.
Power
Spectral
Density
(PSD)
was
employed
extract
features
from
preprocessed
results
demonstrate
proposed
method
outperforms
traditional
algorithms
predicting
efficiency
identifies
complex,
phase-specific
changes
movement
patterns.
model
effectively
sprinters’
movements
differentiate
skill
levels.
Using
Python
software,
achieved
impressive
metrics,
accuracy
(94.5%),
precision
(92.7%),
recall
(93.6%),
F1-score
(92.1%),
R
2
(0.92),
AUC
(0.91),
highlighting
its
robust
predictive
ability.
illustrates
how
models
can
advance
mechanics
provide
insightful
information
athletes
coaches
seeking
improve
performance.
Язык: Английский
AI-Based Player Fatigue and Workload Monitoring Systems
Advances in computational intelligence and robotics book series,
Год журнала:
2025,
Номер
unknown, С. 15 - 40
Опубликована: Май 13, 2025
The
Present
chapter
examines
AI-based
player
fatigue
and
workload
monitoring
systems
in
optimizing
the
performance
of
athletes
while
preventing
them
from
burning
out.
These
are
harnessed
advanced
models
machine
learning,
analyzing
physiological,
biomechanical,
mental
health
data
to
improve
training
recovery
strategies.
Advances
wearable
technologies
will
dramatically
change
way
coaches
approach
world
management
by
real-time
predictive
analytics.
This
also
deals
with
some
difficulties,
including
accuracy,
privacy
ethical
issues,
associated
athlete
autonomy.
Other
possible
future
trends
relate
cognitive
metrics,
multisport
sharing,
involvement
virtual
or
augmented
reality
environments.
goal
at
end
road
for
is
optimal
athletic
performance,
but
all
this
should
be
done
aim
supporting
long-term
well-being.
Язык: Английский