Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Viktoriya Petrakova has a total of 19 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    SSA post-analysis of COVID-19 incidence magnitude in Novosibirsk Region

    by Viktoriya S. Petrakova, Петракова Виктория Сергеевна

    The article is aimed at conducting post-analysis of COVID-19 epidemiological data in Novosibirsk for the period from 2020 to 2023. The study emphasizes the importance of data post-analysis for understanding the dynamics of SARS-CoV-2 spread and the characteristics of its impact on public health. The study results make possible to assess of how population susceptibility to different virus strains has changed, what is the difference between an outbreak of diseases associated with the emergence of a new virus strain and its seasonal infection spread. There was applied the SSA method for analyzing time series to separate them into components as well as studying key indicators such as the number of new infections, deaths, critical cases, hospitalizations, and ventilator-dependent patients in the Novosibirsk region. Three main components have been identified for the described data sets: a general trend that reflects changes in the rate of virus spread related to spread of new strains, as well as periodic phenomena associated with virus strains and seasonality. The results show that a significant part of the changes in disease dynamics is accounted for by the emergence of new strains, but also due to «chronicity» epidemic with seasonal fluctuations. The observed relationships and time lags between the number of critically-ill patients and the number of recorded deaths due to COVID-19, as well as between the number of hospitalized patients and ventilator-dependent patients are shown. Thus, it is concluded that the identified trend depicting a change between the number of infected people and development of virus strains can be useful for refining the parameters of mathematical models for COVID-19 spread. The SEIR-HCD differential model, which was previously used to simulate the disease spread in the Novosibirsk region, was chosen as an illustrative example. It is shown that the parameter of the virus spread rate, restored through the selected trend, when introduced into the model, provides a smaller modeling error than the if it was generated using the solution of the inverse issue.

  • Optimal Control Model for Assessing the Economic Consequences of the Epidemic Process

    by V. Petrakova

    This study proposes a mathematical model for estimating the economic consequences of epidemic processes, and tested on COVID-19 pandemic data. The proposed model combines a modified epidemiological compartmental model of the SIR type, including vaccinated and deceased population groups, with a lagged Solow–Swan economic growth model. The formulation of this coupled system as an optimal control problem is also proposed, where the control function represents time-dependent interventions (e.g., social distancing measures) that affect the rate of virus transmission. The cost functional of the optimal control model balances epidemiological outcomes (the number of infections and deaths) with the socioeconomic costs of introducing restrictions. Approaches to estimating the model parameters for both the epidemiological and economic components are proposed using real data for 2020–2023 time period. Computational experiments show that the optimal control model provides a more accurate forecast of the spread of the epidemic compared to the baseline compartmental model. The results quantitatively demonstrate the significant negative impact of epidemic outbreaks on human capital and gross regional product compared to a scenario without the pandemic. Modifying the model after introducing of control allows for scenario analysis, assessing the economic impact of various intervention strategies. The study concludes that, although accurate epidemiological forecasting is not always necessary for economic evaluation, an integrated optimal control approach provides a valuable tool for developing balanced strategies that will mitigate both public health and economic losses during future pandemics.

  • About Primary Analysis of Time Series of Field Measurements and Non-Stationarity Concept

    by Viktoriya Petrakova

    The paper is aimed at analyzing the difficulties that arise when checking time series of field measurements for stationarity and identifying the main characteristics of the series. This paper provides a comprehensive review of methods for assessing stationarity in time series, highlighting the inherent challenges in primary analysis. The author emphasizes the limitations of graphical methods (including (P)ACF), statistical tests (such as ADF, PP, KPSS, ZA tests and CUSUM), and procedures like the variance ratio test, particularly when dealing with complex or real-world data. A novel, simple procedure is proposed to detect changes in the mean and variance of a series by analyzing the symmetry and distribution of aggregated samples of mean and variances of time series subsamples. The method is tested on synthetic data, including white noise, trend, seasonality, and heteroscedastic series, demonstrating its ability to visually identify non-stationary behaviors through constructed scatterplots and histograms. While not a formal stationarity test, the procedure offers a preliminary tool for characterizing time series behavior, complementing existing methodologies by focusing on fundamental statistical properties.

  • Open Access

    Temporal patterns of instrumentation failures in atmospheric boundary layer observational systems

    by Eugeniya Karepova, Yury Shan’kO, Alexei Tokarev, Viktoriya Petrakova

    His study examines characteristic patterns associated with equipment failures in time series data from urban environmental monitoring systems. While prolonged data gaps represent the most apparent failure indicator, equipment malfunctions manifest through more complex signatures that can compromise forecasting accuracy. Analysis of six-year datasets from two distributed air quality monitoring networks in Krasnoyarsk (comprising 30 monitoring stations) reveals multiple failure-induced patterns, including both localized and sequential outliers. The non-stationary nature of environmental time series necessitates careful discrimination between genuine equipment failures and natural anomalies, as not all deviations require data exclusion. Current methodologies for failure pattern detection are reviewed, with particular attention to their limitations when applied to real-world non-stationary measurement series. The findings highlight critical considerations for maintaining data quality in operational environmental monitoring systems.


  • Study of the Inverse SEIR-HCD Epidemiological Mean Field Problem

    by Viktoriya Petrakova

    The paper is devoted to the analysis of the inverse coefficient mean field problem for describing epidemiological processes. The idea of formulation of the inverse problem is to reconstruct the population behavior (for example, its tendency to choose a behavior strategy to avoid infection) with a known virus transmission scheme. As a virus transmission scheme, the compartmental model SEIR-HCD is considered, where exposed, hospitalized, and critically ill patients are taken into account in the population in addition to symptomatically infected, recovered, and dead. To solve the problem, the model functional determining the population behavior is represented as a linear combination of processes that force the population to choose one or another behavior strategy. The level of the influence of a single process is determined by the weight coefficient to be reconstructed. The article proposes computational experiments showing the realization of this approach. The hypothesis turned out to be correct, as approved through computational experiments, which allowed for the evaluation of the impact of the chosen functionality on the modeling results. An adjusted formulation of the direct problem is considered, allowing for direct consideration of population behavior in the epidemiological parameters of the model. The considered formulation showed a smaller forecast error relative to real data than the original approach.

  • Correction of Dynamic Model Parameters by Observed Data

    by V. Shaydurov, V. Petrakova

    The article presents a method for reconstructing the parameters of dynamic model. The models is described by a system of ordinary differential equations with the number of parameters in the right-hand side exceeding the number of sought functions. To determine the parameters, firstly the indefinite system of algebraic equations with a rectangular matrix is constructed which resulting from the approximation of the system of differential equations taking into account the known values of the functions given at two successive moments in time. In addition, secondly the smooth change of parameters over time is involved. Simultaneous minimization of the residual for the underdetermined system and the sum of squares of the parameter differences at two successive moments of time leads to a regularized system of linear algebraic equations with a positive definite matrix and with a unique solution. The presented method is compared with the method of solving the inverse problem of restoring parameters in the frame of the epidemiological model SEIR-HCD. The comparison showed a smaller error in the presented method compared to the traditional method of solving the inverse problem.

  • Open Access

    Inverse Coefficient Problem for Epidemiological Mean-Field Formulation

    by Viktoriya Petrakova

    The paper proposes an approach to solving the inverse epidemiological problem, written in terms of the “mean-field” theory. Finding the coefficients of an epidemiological SIR mean-field model is reduced to solving an optimization problem, for the solution of which only zero-order methods can be used. An algorithm for the solution of the inverse coefficient problem is proposed. Computational experiments were carried out to compare the obtained solutions with respect to synthetic and real data. The results of computational experiments have shown the efficiency of this approach. Ways to further improve the approach have also been determined.