Classifying chronic kidney disease using selected machine learning techniques
International Journal of ADVANCED AND APPLIED SCIENCES,
Год журнала:
2025,
Номер
12(2), С. 72 - 79
Опубликована: Фев. 1, 2025
Chronic
kidney
disease
(CKD)
is
a
serious
global
health
problem
with
high
mortality
rates,
often
due
to
late
diagnosis.
Early
detection
and
classification
are
essential
improve
treatment
outcomes
slow
progression.
This
study
evaluates
the
performance
of
four
machine
learning
algorithms—linear
discriminant
analysis
(LDA),
Naïve
Bayes,
C4.5
decision
tree,
Random
Forest—in
classifying
CKD
using
Kaggle
dataset
containing
1,659
instances
52
features,
covering
demographic,
lifestyle,
clinical
data.
After
data
pre-processing,
accuracies
algorithms
were
assessed.
LDA
showed
highest
accuracy
at
92.8%,
followed
by
Bayes
(92.1%),
(92.0%),
Forest
(91.9%)
before
hyperparameter
tuning.
tuning,
achieved
92.5%,
(92.2%),
remaining
92.1%.
However,
even
after
remained
most
accurate,
demonstrating
superior
performance.
The
key
features
contributing
serum
creatinine,
glomerular
filtration
rate
(GFR),
muscle
cramps,
protein
in
urine,
fasting
blood
sugar,
itching,
systolic
pressure,
urea
nitrogen
(BUN),
HbA1c,
edema,
total
cholesterol,
body
mass
index
(BMI),
gender.
These
findings
confirm
that
outperforms
other
without
need
for
emphasizing
value
improving
early
diagnosis
management
CKD.
Язык: Английский
Sentiment Analysis on Marketplace in Indonesia using Support Vector Machine and Naïve Bayes Method
Jurnal Ilmiah Teknik Elektro Komputer dan Informatika,
Год журнала:
2024,
Номер
10(1), С. 39 - 39
Опубликована: Фев. 16, 2024
This
research
addresses
the
challenges
of
marketplace
customer
feedback,
which
is
an
important
aspect
in
today's
era
online
transactions.
Marketplaces
often
receive
many
unsatisfactory
comments
from
their
customers
through
social
media
platforms.
One
approach
that
can
be
used
to
address
this
sentiment
analysis.
contributes
new
insights
as
recommendations
for
marketplaces
based
on
opinions
available
services
and
delivery
.
The
analysis
methods
are
Naive
Bayes
Support
Vector
Machine
because
they
considered
best
training
text-based
classification
models.
Before
being
classified,
data
goes
preprocessing
stages
such
cleaning,
case
folding,
filtering,
stemming,
tokenizing,
well
feature
extraction
using
Term
Frequency
-
Inverse
Document
(TF-IDF).
objects
analyzed
divided
into
several
well-known
Indonesia
Tokopedia,
Lazada,
Shopee
discussing
goods.
study
comes
Twitter
(X)
accessed
August
27,
2023,
crawling
techniques
successfully
obtained
much
2057
Tweet
data.
accuracy
SVM
method
when
compared
method.
Words
service
talks
include
price,
service,
application
independence,
others.
As
goods,
common
words
COD,
delivery,
package,
courier,
cheap,
others
appear.
Both
have
good
recommended
use
similar
Язык: Английский
Impact of a Synthetic Data Vault for Imbalanced Class in Cross-Project Defect Prediction
Journal of Electronics Electromedical Engineering and Medical Informatics,
Год журнала:
2024,
Номер
6(2), С. 219 - 230
Опубликована: Апрель 20, 2024
Software
Defect
Prediction
(SDP)
is
crucial
for
ensuring
software
quality.
However,
class
imbalance
(CI)
poses
a
significant
challenge
in
predictive
modeling.
This
study
delves
into
the
effectiveness
of
Synthetic
Data
Vault
(SDV)
mitigating
CI
within
Cross-Project
(CPDP).
Methodologically,
addresses
across
ReLink,
MDP,
and
PROMISE
datasets
by
leveraging
SDV
to
augment
minority
classes.
Classification
utilizing
Decision
Tree
(DT),
Logistic
Regression
(LR),
K-Nearest
Neighbors
(KNN),
Naive
Bayes
(NB),
Random
Forest
(RF),
also
model
performance
evaluated
using
AUC
t-Test.
The
results
consistently
show
that
performs
better
than
SMOTE
other
techniques
various
projects.
superiority
evident
through
statistically
improvements.
KNN
dominance
average
results,
with
values
0.695,
0.704,
0.750.
On
16.06%
improvement
over
imbalanced
12.84%
SMOTE.
Similarly,
on
20.71%
10.16%
Moreover,
PROMISE,
13.55%
7.01%
RF
displays
moderate
performance,
closely
followed
LR
DT,
while
NB
lags
behind.
statistical
significance
these
findings
confirmed
t-Test,
all
below
0.05
threshold.
These
underscore
SDV's
potential
enhancing
CPDP
outcomes
tackling
challenges
SDV.
With
as
best
classification
algorithm.
Adoption
could
prove
be
promising
tool
defect
detection
mitigation
Язык: Английский
Research on the Promotion of Constraining Relationships by Artificial Intelligence Based on the Composite Weighting Model, Attribution Theory, and Naive Bayes Classifier
Опубликована: Окт. 18, 2024
Язык: Английский
Survey on the Applications of Differential Privacy
Опубликована: Дек. 13, 2024
Язык: Английский
Interplay of Marketing Strategies, Smart City Development, and Information Systems: A Comprehensive Review
Опубликована: Дек. 6, 2023
This
comprehensive
review
explores
the
intricate
interplay
between
marketing
strategies,
smart
city
development,
and
information
systems
in
contemporary
urban
contexts.
The
study
addresses
evolving
landscape
of
urbanization,
where
traditional
approaches
integrate
with
cutting-edge
technologies
within
initiatives.
primary
aim
is
to
provide
a
understanding
relationships
among
these
three
dimensions
their
impact
on
sustainable
development.
A
systematic
literature
methodology
employed,
encompassing
databases
such
as
PubMed,
IEEE
Xplore,
ScienceDirect,
JSTOR.
results
highlight
dynamic
evolution
strategies
cities,
role
catalysts
for
innovation,
challenges
opportunities
associated
this
interplay.
contributes
novel
insights
by
identifying
gaps
current
knowledge,
emphasizing
importance
stakeholder
collaboration,
ethical
considerations,
need
inclusive
culturally
sensitive
realm
Язык: Английский