Numerical-statistical study of the prognostic efficiency of the SEIR model

Результат исследования: Научные публикации в периодических изданияхстатьярецензирование


A comparative analysis of the differential and the corresponding stochastic Poisson SEIR-models is performed for the test problem of COVID-19 epidemic in Novosibirsk modelling the period from March 23, 2020 to June 21, 2020 with the initial population N = 2 798 170. Varying the initial population in the form N = n m with m ≥ 2, we show that the average numbers of identified sick patients is less (beginning from April 7, 2020) than the corresponding differential values by the quantity that does not differ statistically from C(t)/m, with C ≈ 27.3 on June 21, 2020. This relationship allows us to use the stochastic model for big population N. The practically useful 'two sigma' confidential interval for the time interval from June 1, 2020 to June 21, 2020 is about 108% (as to the statistical average) and involves the corresponding real statistical estimates. The influence of the introduction of delay on the prognosis, i.e., the incubation period corresponding to Poisson model is also studied.

Язык оригиналаанглийский
Страницы (с-по)337-345
Число страниц9
ЖурналRussian Journal of Numerical Analysis and Mathematical Modelling
Номер выпуска6
СостояниеОпубликовано - 1 дек 2021

Предметные области OECD FOS+WOS





Подробные сведения о темах исследования «Numerical-statistical study of the prognostic efficiency of the SEIR model». Вместе они формируют уникальный семантический отпечаток (fingerprint).