ANALYTICAL REVIEW OF PUBLICATIONS ON MACHINE LEARNING METHODS IN ONCOLOGY AND APPROACH TO EVALUATING THEIR QUALITY
Computer Systems and Information Technologies,
Journal Year:
2024,
Volume and Issue:
1, P. 6 - 16
Published: March 28, 2024
The
work
includes
an
analytical
review
of
publications
on
machine
learning
methods
in
oncology
and
approach
to
evaluating
their
quality.
An
analysis
by
year
was
conducted
the
Web
Science
Scopus
bibliometric
databases.
highest
number
authors,
among
universities,
countries,
publication
categories
database
are
presented.
A
multifactor
regression
prediction
model
for
bone
tissue
density
oncological
pathology
predicting
four
severity
grades
studied
disease
course
proposed.
This
included
following
factors
with
corresponding
weights:
gender
(2.1),
age
(0.06),
stage
(0.9),
absence/presence
B-symptoms
(A/B)
international
prognostic
index
(IPI-NCCN)
(
1.1),
body
mass
(BMI)
(-0.2),
chemotherapy
courses
Charlson
comorbidity
(СCI)
(0.3),
mineral
after
completion
(HU
C)
(-0.08),
β-2-microglobulin
(B2M)
level
(0.0007),
lactate
dehydrogenase
(LDH)
(0.006),
surface
area
(BSA)
(-3.3).
To
assess
confidence
proposed
disorders
pathology,
ROC
performed
obtain
curves
under
them
estimated.
conclusion
made
about
quality
classification
sensitivity,
specificity,
value
positive
negative
results,
ratio
probability
as
well
accuracy
were
determined.
For
each
degrees
violations
(1C,
2C,
3C,
4C),
it
is
necessary
carry
out
appropriate
calculations,
matrices
inconsistencies
which
given
tables.
Sensitivity
calculated
1C
(98.8%),
2C
(97.5%),
3C
(95.2%)
4C
(98.5%);
specificity
(90.4%),
(83.3%),
(90.9%)
(95%);
predictive
a
result
(97.6%),
(95.2%),
(97.5%)
(97%);
(95%),
(90.9%),
(83.3%)
(97.4%);
(97.1%),
(97.1%).
According
results
curves,
high
(AUC=0.869),
(AUC=0.869)
established.
average
according
(AUC=0.758).
Language: Английский
OPTIMIZATION OF THE SCIENTIFIC SEARCH FOR LYME BORRELIOSIS SEVERITY DIAGNOSTICS
Eastern Ukrainian Medical Journal,
Journal Year:
2024,
Volume and Issue:
12(4), P. 767 - 776
Published: Jan. 1, 2024
The
article
deals
with
the
data
of
use
analytical
search
capabilities
for
optimization
literature
sources
analysis
in
Scopus
scientometric
database
when
studying
current
state
Lyme
borreliosis
diagnosis
and
treatment
children.
Methods
examination:
requests
assessment
were
used
to
investigate
scientific
Borreliosis
diagnosis.
gives
make
a
relevant
on
studied
issues.
In
particular,
assess
relevance
research
LB
infectious
diseases
field,
following
query
was
formulated
database:
(ALL("tick
bites")
OR
ALL("endemic
area
lyme
borreliosis")
ALL("questionnaire
parents
children
disease")
ALL("affected
by
ticks")
ALL("DNA
complex
B.
burgdorferi
s.l.
")
ALL("B.
miyamotoi").
Research
results.
optimizes
topic;
total
number
papers
comparisons
Cite
Score,
SIR,
SNIP;
results
authors,
subject
areas,
country,
among
leading
institutions,
types
publications,
largest
financial
sponsors
topic.
Conclusions:
review
publications
artificial
intelligence
medicine,
infectology
presented
article.
carried
out.
authors
papers,
taking
into
account
countries
presented.
Based
obtained
from
using
Scopus,
we
can
conclude
that
there
is
growing
interest
problem
causes
onset
progression
borreliosis.
Thus,
such
extremely
improving
maintaining
patient
health.
Language: Английский