FORMALIZATION OF NON-QUANTITATIVE MEDICAL DATA

Аннотация

Since different recognition algorithms perform differently on the same training sample, medical of information A synthetic decision rule that flexibly uses the strengths of algorithms is relevant from issues is considered.

Тип источника: Журналы
Годы охвата с 2021
inLibrary
Google Scholar
f
17
0

Скачивания

Данные скачивания пока недоступны.
Поделиться
Narzullayev , D., Tursunov, A., & Rahmonov, E. (2025). FORMALIZATION OF NON-QUANTITATIVE MEDICAL DATA. Евразийский журнал академических исследований, 5(10(MPHAPP), 17. извлечено от https://inlibrary.uz/index.php/ejar/article/view/138138
0
Цитаты
Crossref
Сrossref
Scopus
Scopus

Аннотация

Since different recognition algorithms perform differently on the same training sample, medical of information A synthetic decision rule that flexibly uses the strengths of algorithms is relevant from issues is considered.


background image

17

Volume 5, Issue 10: Special Issue
(EJAR)

ISSN: 2181-2020

MPHAPP

THE 6TH INTERNATIONAL SCIENTIFIC AND PRACTICAL

CONFERENCE

MODERN PHARMACEUTICS: ACTUAL

PROBLEMS AND PROSPECTS

TASHKENT, OCTOBER 17, 2025

in-academy.uz

FORMALIZATION OF NON-QUANTITATIVE MEDICAL DATA

Narzullayev D.Z.

1

Tursunov A.T.

2

Rahmonov E.D.

3

Tashkent Pharmaceutical Institute, Tashkent city, Republic of Uzbekistan

e-mail: davr1960@mail.ru

https://doi.org/10.5281/zenodo.17310339

Relevance:

Since different recognition algorithms perform differently on the same training

sample, medical of information A synthetic decision rule that flexibly uses the strengths of algorithms
is relevant from issues is considered.

Purpose of the study:

Other nonparametric methods occupy a special place in cases where the

shape of the distribution density curve is unknown and there are no assumptions about its nature.

Methods and techniques:

These include the multidimensional histogram method, the k-nearest

neighbors method, the Euclidean distance method, the potential function method, and a generalization
of these methods is the so-called Parzen estimates. These methods formally work with objects as
whole structures, but depending on the type of recognition problem, they can be either intensional or
extensional forms.

Nonparametric methods are used to analyze medical data in multidimensional spaces.

quantitative It analyzes the relative number of objects and uses various distance functions between
the objects in the training sample and the objects being recognized. For quantitative features, when
their number is much smaller than the sample size, operations with objects play an intermediate role
in estimating the local density of conditional probabilities, and objects do not have the semantic
meaning of independent information units.

Results:

Thus, the same technological operations of nonparametric methods, depending on the

conditions of the problem, have the meaning of locally estimating the probability distribution density
of the values of the given signs in a quantitative medical form or of assessing the similarity and
difference of objects.

Methods based on assumptions about the class of decision functions . In

this group of methods,

the general form of the decision function is known and its quality functional is assumed to be given.
The most common are the expressions of decision functions in the form of linear and generalized
nonlinear polynomials.

In that case, the distribution relationship will look like this:

𝐽 =

Λ

𝑊ℓ+1,ℓ+1

Λ

ℓ+1,ℓ+1

𝑎̅

𝑇

𝑊𝑎̅

𝑎̅

𝑇

𝑆𝑎̅

(1)

Λ

𝑊

≥ 0

and (1) is valid, so

𝑎̅

𝑇

𝑊𝑎̅

the square form is non-negative.

𝑎̅

Since the first sign of the

last equality does not depend on the set, and the criterion for choosing empirical scales in solving
symbol recognition problems can be written as follows:

𝑎̅

𝑇

𝑊𝑎̅

𝑎̅

𝑇

𝑆𝑎̅

→ 𝑚𝑖𝑛

(2)

here

𝑎̅ ∈ 𝐵

𝑔

When it is necessary to find optimal integer numerical symbols according to the condition (2),

the problem looks like this. Solving such problems is often achieved using some gradient algorithms.

Conclusions:

The diversity of methods in this group is explained by the wide range of decision

rule quality functionals and extremum search algorithms used.