Abstract: After COVID-19 pandemic, the epidemilogical data prediction had become of a great importance. Since that, numerous different prognostic models, including those involving neural-network based, have been developed, applied and verified. Shortterm models are capable to reproduce the oscillacion, but incapable to make a long term prognosis; long-term ones suffer from the noise in the data and require its reduction. In this paper, we propose a method of data prediction using values range discretization as an alternative to the smoothing to get rid of noise-borne problems and applying lag prediction. It is shown that the approach is capable to improve the prognosis quality even for the irregurlar data. Keywords: epidemiology, neural network, prognosis, discretization.

Cite: Elistratov S.A.
Data discretization in prognostic models for epidemiology
Cifra. Медико-биологические науки. 2024. N3. 2 :1-6. DOI: 10.60797/BMED.2024.3.2 РИНЦ

Dates:

Accepted:	Dec 12, 2024
Published print:	Dec 27, 2024
Published online:	Dec 27, 2024

Identifiers:

Elibrary:

77259727

Citing: Пока нет цитирований

Altmetrics: