Many
university
students
lack
clear
goals
at
the
start
of
their
academic
journey,
leading
to
a
motivation
and
active
engagement
in
classroom.
This
often
results
difficulties
understanding
lectures,
poor
exam
performance,
failure
acquire
essential
skills.
Despite
this,
finding
suitable
job
with
decent
salary
after
graduation,
particularly
data
science,
poses
significant
challenge
for
students.
To
enhance
science
students,
we
have
obtained
Data
Science
Salaries
2023
dataset
from
Kaggle
website
exploring
salaries
through
Exploratory
Analysis
(EDA).
Furthermore,
this
study
aims
uncover
hidden
patterns
extract
valuable
insights
predict
success
factors
higher
by
applying
Linear
Regression,
KNN
Decision
Tree
Regression
Models.
The
findings
could
assist
both
scientists
comprehending
contributing
salaries,
thereby
enriching
knowledge
future
career
endeavours.
International Journal of Computational Intelligence Systems,
Год журнала:
2024,
Номер
17(1)
Опубликована: Янв. 3, 2024
Abstract
The
continuous
progress
of
multimedia
technology
in
music
educational
institutions
has
led
to
the
recognition
its
importance
our
country
and
society.
traditional
approach
piano
teaching
limitations,
which
can
be
overcome
by
adopting
alternative
approaches
instrument,
using
advances
science
technology.
For
pianist,
expressing
emotions
thoughts
through
is
crucial,
teachers
now
use
tools
exemplify
their
musical
skills
students
effectively.
This
manuscript
proposes
Remote
Piano
Teaching
Based
on
Attention-Induced
Multi-Head
Convolutional
Neural
Network
Optimized
with
Hunter–Prey
Optimization
improve
piano-teaching
quality.
At
first,
input
data
taken
from
Triad
Wavset
dataset.
Afterward,
are
fed
preprocessing
stage.
stage
involve
cleaning
or
scrubbing
that
process
identifying
errors,
inconsistencies,
incorrectness
a
dataset
help
adaptive
distorted
Gaussian
matched
filter.
Then,
preprocessed
output
(AIMCNN)
for
effectively
predict
hunter–prey
optimization
(HPO)
algorithm
proposed
optimize
parameters
Network.
performance
technique
evaluated
under
metrics
like
accuracy,
computational
time,
learning
skill
analysis,
activity
behavior
analysis;
student
ratio
evaluation
analysis
evaluated.
RPT-AIMCNN-HPO
attains
better
prediction
accuracy
12.566%,
12.075%
15.993%,
higher
15.86%,
15.26%
16.25%
compared
existing
methods.
Case Studies in Thermal Engineering,
Год журнала:
2024,
Номер
57, С. 104326 - 104326
Опубликована: Март 31, 2024
This
paper
investigates
the
solubility
behavior
of
digitoxin
in
supercritical
carbon
dioxide
(CO2)
through
a
comprehensive
analysis
employing
ensemble
learning
techniques
and
various
regression
models.
The
dataset
consists
temperature
pressure
as
input
variables,
with
solvent
density
output
variables.
Utilizing
bagging
method,
Gaussian
process
(GPR),
Bayesian
Ridge
Regression
(BRR),
Orthogonal
Matching
Pursuit
(OMP),
Polynomial
(PR)
were
employed
Hyper-parameter
tuning
was
achieved
gradient-based
optimization.
Results
revealed
that
emerges
most
effective
model
for
predicting
both
solubility.
For
prediction,
BAG-PR
yields
an
R2
score
0.98527,
MSE
MAE
1.4290E-03
3.40547E-02,
respectively.
Concerning
density,
achieves
outstanding
0.99766,
along
7.4759E+01
7.33964E+00,
These
results
show
learning,
polynomial
can
accurately
predict
revealing
digitoxin's
CO2.
Case Studies in Thermal Engineering,
Год журнала:
2023,
Номер
49, С. 103236 - 103236
Опубликована: Июнь 29, 2023
Data-driven
models
were
employed
for
the
solubility
correlation,
while
focus
was
on
modeling
of
raloxifene
drug
and
density
carbon
dioxide
based
temperature
(T)
pressure
(P)
inputs.
Three
Machine
Learning
models,
namely
Multilayer
Perceptron
(MLP),
Bayesian
Ridge
Regression
(BRR),
LASSO
regression,
optimized
using
MPHPT
method
hyper-parameter
tuning.
The
dataset
consisted
experimental
measurements
(y),
CO2
density.
For
prediction
density,
MLP
model
exhibited
excellent
performance
with
an
R2
score
0.99726,
demonstrating
a
significant
level
association
between
anticipated
observed
values.
mean
squared
error
(MSE)
9.8721E+01,
absolute
percentage
(MAPE)
1.78565E-02,
maximum
1.86395E+01.
BRR
achieved
slightly
lower
accuracy,
scores
0.83317
0.83001,
respectively.
Regarding
drug,
demonstrated
strong
predictive
capability
0.99343.
MSE
3.0869E-02,
MAPE
4.02666E-02,
3.01133E-01.
also
provided
reasonable
predictions,
0.90955
0.8891,
However,
they
higher
MSEs
MAPEs
compared
to
model.
Forests,
Год журнала:
2023,
Номер
14(9), С. 1782 - 1782
Опубликована: Сен. 1, 2023
Tactical
planning
in
timber
harvesting
involves
aspects
related
to
forest
macro-planning
and,
particularly,
the
allocation
of
resources
and
sequencing
activities,
all
which
affect
yards
roads
productivity
machines.
Data-driven
approaches
encourage
use
information
obtained
from
data
enhance
decision-making
efficiency
support
development
short-term
strategies.
Therefore,
our
investigation
was
intended
determine
whether
a
data-driven
approach
can
generate
sufficient
input
for
modeling
forwarder
forwarding
Pinus
Eucalyptus
planted
forests,
tactical
planning.
We
utilized
3812
instances
raw
that
were
generated
over
36-month
period.
The
collected
23
loggers
who
operated
forests.
applied
22
regression
algorithms
supervised
learning
method
an
experimental
machine
instances.
evaluated
fitted
models
using
three
performance
metrics.
Out
tested
algorithms,
default
mode
light
gradient
boosting
produced
root
mean
squared
error
14.80
m3
h−1,
absolute
2.70,
coefficient
determination
0.77.
methods
adequately
forests
help
managers
with
Case Studies in Thermal Engineering,
Год журнала:
2023,
Номер
53, С. 103767 - 103767
Опубликована: Ноя. 29, 2023
This
research
paper
presents
a
comprehensive
thermodynamic
and
heat
transfer
study
on
predicting
the
ternary
solubility
of
Nystatin
in
SC-CO2-Ethanol
(supercritical
CO2
ethanol).
The
employed
process
is
thermal-based
green
processing
for
preparation
solid
nanoparticles.
data
collection,
consisting
temperature
pressure
as
input
features
target
variable,
was
used
to
train
evaluate
four
different
machine
learning
algorithms:
Random
Forest
(RF),
Extra
Trees
(ET),
NU-SVR,
EPSILON-SVR.
hyper-parameter
tuning
Bat
Optimization
Algorithm
(BA),
nature-inspired
optimization
technique
fine-tune
models
enhance
their
predictive
capabilities.
ET
model
had
notable
R2
score
0.98526,
RMSE
2.48774E-02,
MAE
2.13417E-02.
RF
also
yielded
strong
performance,
achieving
an
0.98436,
2.55130E-02,
2.06314E-02.
However,
NU-SVR
exhibited
superior
performance
compared
other
models,
evidenced
by
its
remarkable
0.99943,
thereby
showcasing
exceptional
precision.
were
4.92372E-03
3.94943E-03,
respectively,
underscoring
precision
solubility.
EPSILON-SVR
model,
while
still
respectable,
obtained
0.93574
terms
R2,
4.37434E-02,
3.79800E-02.
