Materials,
Journal Year:
2022,
Volume and Issue:
15(12), P. 4164 - 4164
Published: June 12, 2022
Several
types
of
research
currently
use
machine
learning
(ML)
methods
to
estimate
the
mechanical
characteristics
concrete.
This
study
aimed
compare
capacities
four
ML
methods:
eXtreme
gradient
boosting
(XG
Boost),
(GB),
Cat
(CB),
and
extra
trees
regressor
(ETR),
predict
splitting
tensile
strength
28-day-old
self-compacting
concrete
(SCC)
made
from
recycled
aggregates
(RA),
using
data
obtained
literature.
A
database
381
samples
literature
published
in
scientific
journals
was
used
develop
models.
The
were
randomly
divided
into
three
sets:
training,
validation,
test,
with
each
having
267
(70%),
57
(15%),
(15%)
samples,
respectively.
coefficient
determination
(R2),
root
mean
square
error
(RMSE),
absolute
(MAE)
metrics
evaluate
For
training
set,
results
showed
that
all
models
could
SCC
RA
because
R2
values
for
model
had
significance
higher
than
0.75.
XG
Boost
best
performance,
showing
highest
value
=
0.8423,
as
well
lowest
RMSE
(=0.0581)
MAE
(=0.0443),
when
compared
GB,
CB,
ETR
Therefore,
considered
predicting
RA.
Sensitivity
analysis
revealed
variable
contributing
most
split
this
material
after
28
days
cement.
Journal of Materials Research and Technology,
Journal Year:
2023,
Volume and Issue:
24, P. 6348 - 6368
Published: April 27, 2023
The
significant
requirement
for
natural
resources,
specifically
as
ingredients
of
cement,
is
accelerating
due
to
the
considerable
growth
construction
sector.
Further,
cement
production
adversely
affects
climate
change
generation
bulk
CO2
emissions.
At
same
time,
a
quantum
ceramic
waste
generated
either
in
process
or
demolition
products
each
year.
unavailability
an
adequate
way
dispose
this
negatively
impacts
environment
and
landfills.
Numerous
researchers
have
explored
potential
utilizing
powder
partial
replacement
reduce
allied
issues.
Hence,
current
study,
supervised
machine
learning
(ML)
algorithms,
i.e.,
Decision
Tree
(DT),
AdaBoost
(AdB),
Bagging
(Bg),
Random
Forest
(RF),
Gradient
Boosting
(GB)
XGBoost
(XGB)
are
employed
predicting
Compressive
Strength
(CS)
concrete
(CWPC).
performance
models
also
assessed
by
using
coefficient
determination
(R2),
Mean
Absolute
Error
(MAE),
Root
Square
(RMSE),
Nash
Sutcliffe
efficiency
(NSE).
k-fold
cross-validation
technique
applied
afterwards
validate
model's
performance.
For
CS
CWPC,
RF
algorithm
most
effective
among
with
higher
R2
value
0.97
significantly
lesser
RMSE
MAE
values
1.40
1.13,
respectively.
SHAP
analysis
shows
that
curing
days
feature
has
highest
influence
on
CWPC.
As
per
quantitative
Environmental
Impact
Assessment
(EIA),
10%
CWP
content
can
6.78%,
8.68%,
7.18%,
7.19%
reduced
change,
ecosystem
quality,
human
health,
Moreover,
effects
non-renewable
energy
depletion
ozone
layer,
global
warming
primarily
be
maximum
7%,
6%,
9%,
application
ML
techniques
estimating
CWPC
would
benefit
field
civil
engineering
terms
conserving
effort,
time.
