Artificial Intelligence Review,
Год журнала:
2024,
Номер
57(6)
Опубликована: Май 20, 2024
Abstract
Precise
bearing
capacity
prediction
of
circular
foundations
is
essential
in
civil
engineering
design
and
construction.
The
affected
by
factors
such
as
depth,
density
soil,
internal
angle
friction,
cohesion
foundation
radius.
In
this
paper,
an
innovative
perspective
on
a
fuzzy
inference
system
(FIS)
was
proposed
to
predict
capacity.
uncertainty
rules
eliminated
using
Z-number
theory.
effective
parameters,
i.e.,
radius
were
considered
inputs
the
model.
To
compare
regression
FIS
model
with
Z-based
FIS,
statistical
indices
coefficient
determination
(R
2
),
root
mean
square
error
(RMSE),
variance
account
for
(VAF)
employed.
For
training
testing
Z-FIS,
R
(0.977
0.971),
RMSE
(1.645
1.745),
VAF
(98.549%
98.138),
whereas
method,
values
(0.912
0.904),
(5.962
6.76),
(90.12%
88.49%).
It
should
be
mentioned
that
Z
theory
decreased
computational
time
89.28%
(174.04
s
18.65
s).
comparison
indicators
presented
models
revealed
superiority
Z-FIS
over
FIS.
Notably,
sensitivity
analysis
most
parameters
are
soil
density.
Buildings,
Год журнала:
2022,
Номер
12(2), С. 132 - 132
Опубликована: Янв. 27, 2022
Concrete
is
one
of
the
most
popular
materials
for
building
all
types
structures,
and
it
has
a
wide
range
applications
in
construction
industry.
Cement
production
use
have
significant
environmental
impact
due
to
emission
different
gases.
The
fly
ash
concrete
(FAC)
crucial
eliminating
this
defect.
However,
varied
features
cementitious
composites
exist,
understanding
their
mechanical
characteristics
critical
safety.
On
other
hand,
forecasting
concrete,
machine
learning
approaches
are
extensively
employed
algorithms.
goal
work
compare
ensemble
deep
neural
network
models,
i.e.,
super
learner
algorithm,
simple
averaging,
weighted
integrated
stacking,
as
well
separate
stacking
order
develop
an
accurate
approach
estimating
compressive
strength
FAC
reducing
high
variance
predictive
models.
Separate
with
random
forest
meta-learner
received
predictions
(97.6%)
highest
coefficient
determination
lowest
mean
square
error
variance.
