In
this
paper,
we
conducted
an
in-depth
study
on
the
diagnosis
and
prognosis
prediction
of
hematoma
dilatation
peripheral
edema
after
hemorrhagic
stroke.
First,
using
data
extracted
from
table
such
as
flow
number
imaging
examination,
a
discriminative
model
expansion
was
established,
which
discriminated
against
based
explicit
conditions
successfully
realized
probability
with
accuracy
rate
nearly
74%.
Secondly,
paper
analyzed
relationship
between
volume
time,
used
K-means
clustering
to
classify
volume,
explored
effect
different
treatment
methods
progression
volume.
Finally,
LSTM
neural
network
construct
for
MRS
score,
further
optimized
treatment.
Journal of Computer Assisted Learning,
Год журнала:
2023,
Номер
39(4), С. 1116 - 1131
Опубликована: Июнь 7, 2023
Abstract
Background
It
is
vital
to
understand
students'
Self‐Regulatory
Learning
(SRL)
processes,
especially
in
Blended
(BL),
when
students
need
be
more
autonomous
their
learning
process.
In
studying
SRL,
most
researchers
have
followed
a
variable‐oriented
approach.
Moreover,
little
has
been
known
about
the
unfolding
process
of
SRL
profiles.
Objectives
We
present
insights
derived
from
study
that
measured
motivation
and
strategies
used
by
198
university
entry‐level,
business
school,
BL
course
develop
an
understanding
processes.
Methods
The
Strategies
for
Questionnaire
(MSLQ)
was
survey
three
times
during
semester
investigate
profiles
how
they
unfolded
as
progressed
using
person‐oriented
Through
clustering
approach,
we
focus
on
MSLQ's
aspects
its
importance
emphasised
different
theories,
extant
research
into
analytics
(LA)
still
lacking.
Results
Conclusions
longitudinal
identified
minimally,
average,
highly
acknowledged
might
change
result
feedback
received.
What
are
1
or
2
Major
Takeaways
Study?
This
contributes
theory
examining
adaptation
longitudinally
(addressing
challenge
regarding
cyclical
nature
SRL).
LA
investigating
motivational
constructs
currently
lacking
field
bringing
forward
based
empirical
evidence
inform
practice.
Journal of Computational Methods in Sciences and Engineering,
Год журнала:
2024,
Номер
24(3), С. 1341 - 1353
Опубликована: Июнь 17, 2024
Blended
learning
is
the
latest
and
inevitable
trend
in
development
of
education.
Although
blended
research
on
rise,
fewer
studies
examine
behaviour
college
students
environments.
This
study
aimed
to
investigate
behaviours
field
computer
science
these
using
data
mining
algorithms,
taking
teaching
practice
Digital
Signal
Processing
course
as
a
case
study.
A
total
18
behavioural
indicators
were
extracted
divided
into
three
categories:
basic
behaviours,
self-regulated
extended
behaviours.
Data
analysis
yielded
following
conclusions:
(1)
Students
did
not
have
habit
watching
playback
less
receptive
multiple
online
platforms;
(2)
Students’
midterm
performance
duration
livestream
directly
affected
their
with
all
showing
significant
correlations;
(3)
The
clustering
four
different
learner
patterns,
which
calls
for
personalised
strategies;
(4)
random
forest
algorithm
had
an
accuracy
95.4%
predicting
types
learners.
Sustainability,
Год журнала:
2023,
Номер
15(18), С. 14032 - 14032
Опубликована: Сен. 21, 2023
The
existing
flood
stochastic
simulation
methods
are
mostly
applied
to
the
of
intensity
characteristics,
with
less
consideration
for
randomness
hydrograph
shape
and
its
correlation
characteristics.
In
view
this,
this
paper
proposes
a
method
that
combines
morphological
indicators.
Using
Foziling
Xianghongdian
reservoirs
in
Pi
River
basin
China
as
examples,
utilizes
three-dimensional
asymmetric
Archimedean
M6
Copula
construct
models
peak
flow,
volume,
duration.
Based
on
K-means
clustering,
multivariate
Gaussian
is
employed
dimensionless
model.
Furthermore,
separate
two-dimensional
symmetric
established
capture
correlations
between
characteristics
variables
such
coefficient,
occurrence
time,
rising
inflection
point
angle,
coefficient
variation.
By
evaluating
fit
simulated
hydrograph,
complete
synthesized,
which
can
be
control
dispatch
simulations
other
related
fields.
feasibility
practicality
proposed
model
analyzed
demonstrated.
results
indicate
floods
closely
resemble
natural
floods,
making
outcomes
crucial
reservoir
scheduling,
risk
assessment,
decision-making
processes.
Research Square (Research Square),
Год журнала:
2024,
Номер
unknown
Опубликована: Июль 4, 2024
Abstract
This
study
investigates
the
application
of
a
deep
learning-based
predictive
model
to
predict
student
performance.
The
objective
was
enhance
performance
by
predicting
and
monitoring
their
academic
activities,
including
attendance
at
synchronous
sessions,
interaction
with
digital
content,
participation
in
forums,
portfolio
creation
tasks
over
an
year.
applied
experimental
group
students.
Unlike
control
group,
which
did
not
receive
continuous
feedback,
received
personalized,
feedback
based
on
predictions
from
pre-trained
interpreted
OpenAI’s
GPT-4
language
model.
Significant
improvements
were
observed
compared
group.
average
score
quizzes
for
0.81,
notably
higher
than
group's
0.67.
Recorded
session
engagement
0.84,
0.65
Live
forum
activity
also
significantly
rates
0.61
0.62
respectively,
0.42
0.37.
However,
practice
slightly
mean
0.76
0.74
Portfolio
assessment
scores
0.73
0.69
These
results
support
hypothesis
that
using
models
complemented
provide
improves
learning
effectiveness.
Frontiers in Psychology,
Год журнала:
2024,
Номер
15
Опубликована: Окт. 23, 2024
Learning
in
asynchronous
online
settings
(AOSs)
is
challenging
for
university
students.
However,
the
construct
of
learning
engagement
(LE)
represents
a
possible
lever
to
identify
and
reduce
challenges
while
online,
especially,
AOSs.
analytics
provides
fruitful
framework
analyze
students'
processes
LE
via
trace
data.
The
study,
therefore,
addresses
questions
whether
can
be
modeled
with
sub-dimensions
effort,
attention,
content
interest
by
which
data,
derived
from
behavior
within
an
AOS,
these
facets
are
represented
self-reports.
Participants
were
764
students
attending
AOS.
results
best-subset
regression
analysis
show
that
model
combining
multiple
indicators
account
proportion
variance
(highly
significant
R
2
between
0.04
0.13).
identified
set
stable
over
time
supporting
transferability
similar
contexts.
this
study
contribute
both
research
on
AOSs
higher
education
application
teaching
(e.g.,
modeling
automated
feedback).
IEEE Access,
Год журнала:
2023,
Номер
11, С. 72485 - 72497
Опубликована: Янв. 1, 2023
In
the
application
scenarios
of
radio
frequency
identification
technology,
there
are
many
situations
where
a
large
number
labels
respond
to
reader
at
same
time,
resulting
in
not
being
able
be
identified
for
long
time.
order
address
label
collision
problem
identification,
this
paper
studies
impact
statistical
learning
method
on
resolution
and
decoding
labels,
proposes
novel
clustering
using
maximum
posteriori
probability
estimation
based
Monte-Carlo.
Unlike
traditional
algorithms,
proposed
does
require
prior
knowledge
clusters
need
constantly
iterate.
addition,
has
low
complexity
ensures
both
accuracy
robustness
while
quickly
finding
cluster
centroids.
Finally,
performance
is
evaluated
simulation
experiment
field
experiment,
resolved
signals
decoded
matched
filter
phase
jump.
Overall,
effectiveness
our
demonstrated
through
comparisons
with
different
metrics
benchmark
methods,
including
bit
error
rate,
efficiency,
throughput,
error,
time
complexity.
The
reopening
of
the
campus
marks
transition
from
virtual
learning
to
traditional
face-to-face
classroom
setups
in
Malaysian
higher
institutions,
due
influence
COVID-19
pandemic.
Upon
returning
campus,
both
students
and
lecturers
must
adapt
this
change
mode.
This
paper
focuses
on
perspectives
regarding
online
learning.
It
examines
students'
opinions
preferences,
flexibility,
interaction,
experience
modes
To
gather
data,
an
questionnaire
was
administered
Multimedia
University
(MMU)
Malaysia.
results
indicated
that
were
uncertain
about
their
preference
for
physical
versus
Furthermore,
study
highlighted
significance
attribute
flexibility
offered
by
However,
findings
revealed
perceived
importance
interaction
during
These
have
practical
implications
educators
policymakers
designing
post-pandemic
delivery
approaches
education
institutions.
Applied Mathematics and Nonlinear Sciences,
Год журнала:
2023,
Номер
9(1)
Опубликована: Сен. 25, 2023
Abstract
In
this
paper,
a
talent
training
model
based
on
big
data
analysis
is
designed
for
the
background
of
construction
Hainan
Free
Trade
Port.
A
learning
behavior
method
using
K-Means
clustering
algorithm
and
particle
swarm
optimization
algorithm,
which
can
accurately
mine
valuable
information
from
large
amount
user
provide
reference
exploration
Russian
model.
The
accuracy
rate
in
experimental
validation
reach
91.99%,
outstanding
important
support
establishing
systematic
context
