Iryna V. Rozora
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Iryna V. Rozora has a total of 47 co-authors. Here are the top co-authors by number of shared publications:
- Yurii Vasilyevich Kozachenko
Taras Shevchenko National University of Kyiv
8 shared publications - Anastasiia Melnyk
Taras Shevchenko National University of Kyiv
4 shared publications - O. I. Vasylyk
National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Kyiv, Ukraine
4 shared publications - Volodymyr Morenets
Taras Shevchenko National University of Kyiv
4 shared publications - Hanna Livinska
Taras Shevchenko National University of Kyiv
3 shared publications - Mykhailo Sharapov
Taras Shevchenko National University of Kyiv
3 shared publications
Latest Publications
- 21 October 2026
- Published by State Scientific Institution - Ukrainian Institute of Scientific and Technical Expertise and Info
- p. 158
Forming Probabilistic-Statistical Thinking as the Basis of IT Specialist Training within the Course “Probability Theory and Mathematical Statistics”
The work is dedicated to the methodology of forming probabilistic-statistical thinking in future IT specialists within the course “Probability Theory and Mathematical Statistics”. The comprehensive approach of the Department of Applied Statistics, which combines fundamental mathematics with programming, is described. The interactive class format “Report”, aimed at developing analytical and communication skills, is considered. Emphasis is placed on the integration of modern programming languages into educational web resources and the use of the author’s applied projects (“Index”, “Steganeisodos”), which develop critical thinking, a culture of results verification, and an understanding of the role of statistics in cybersecurity.
- 21 October 2026
- Published by State Scientific Institution - Ukrainian Institute of Scientific and Technical Expertise and Info
- p. 158
Forming Probabilistic-Statistical Thinking as the Basis of IT Specialist Training within the Course “Probability Theory and Mathematical Statistics”
The work is dedicated to the methodology of forming probabilistic-statistical thinking in future IT specialists within the course “Probability Theory and Mathematical Statistics”. The comprehensive approach of the Department of Applied Statistics, which combines fundamental mathematics with programming, is described. The interactive class format “Report”, aimed at developing analytical and communication skills, is considered. Emphasis is placed on the integration of modern programming languages into educational web resources and the use of the author’s applied projects (“Index”, “Steganeisodos”), which develop critical thinking, a culture of results verification, and an understanding of the role of statistics in cybersecurity.
- 21 October 2026
- Published by State Scientific Institution - Ukrainian Institute of Scientific and Technical Expertise and Info
- p. 143
A Systemic Methodology for the Statistical Analysis of Complex Stochastic Systems and the Verification of Machine Learning Algorithms Based on Simulation Modeling and Resampling Methods
The article proposes a systemic methodology for the verification of machine learning algorithms operating within complex stochastic environments, based on simulation modeling and resampling methods. It is demonstrated that classical asymptotic statistical criteria often prove inadequate due to the violation of fundamental assumptions regarding normality and homoscedasticity in real-world systems. Based on Monte Carlo simulation results, a comparative analysis of statistical power was conducted, establishing the advantages of resampling-based verification for systems characterized by “heavy-tailed” distributions, high heteroscedasticity, and limited sample sizes. Particular attention is paid to the conceptual risks of decision-making associated with low statistical power and latent Type II errors, which can lead to the false verification of inadequate models. The proposed methodology is recommended for use in intelligent decision support systems and medical diagnostics, where the reliability of conclusions is paramount.
- 21 October 2026
- Published by State Scientific Institution - Ukrainian Institute of Scientific and Technical Expertise and Info
- p. 143
A Systemic Methodology for the Statistical Analysis of Complex Stochastic Systems and the Verification of Machine Learning Algorithms Based on Simulation Modeling and Resampling Methods
The article proposes a systemic methodology for the verification of machine learning algorithms operating within complex stochastic environments, based on simulation modeling and resampling methods. It is demonstrated that classical asymptotic statistical criteria often prove inadequate due to the violation of fundamental assumptions regarding normality and homoscedasticity in real-world systems. Based on Monte Carlo simulation results, a comparative analysis of statistical power was conducted, establishing the advantages of resampling-based verification for systems characterized by “heavy-tailed” distributions, high heteroscedasticity, and limited sample sizes. Particular attention is paid to the conceptual risks of decision-making associated with low statistical power and latent Type II errors, which can lead to the false verification of inadequate models. The proposed methodology is recommended for use in intelligent decision support systems and medical diagnostics, where the reliability of conclusions is paramount.
- 21 October 2026
- Published by State Scientific Institution - Ukrainian Institute of Scientific and Technical Expertise and Info
- p. 110
Interdisciplinary Integration of Data Visualization Methods and Machine Learning Algorithms for the Verification of Classification Tasks
The work is devoted to the methodology of combining visual-analytical approaches and algorithmic solutions of machine learning. The main attention is paid to how the integrated use of visualization at each stage of developing a classification model allows to reveal hidden patterns, verify AI solutions and minimize the risks associated with erroneous predictions in critical areas (for example, diabetes diagnostics). The article considers a comprehensive approach to building a machine learning model for diabetes diagnostics. The main attention is paid not only to the accuracy of algorithms, but also to visual verification methods at each stage from data cleaning to assessing the risks of erroneous predictions. This approach allows to minimize the “black box” effect and increase trust in automated classification systems, in particular for medical diagnostics.
- 21 October 2026
- Published by State Scientific Institution - Ukrainian Institute of Scientific and Technical Expertise and Info
- p. 110
Interdisciplinary Integration of Data Visualization Methods and Machine Learning Algorithms for the Verification of Classification Tasks
The work is devoted to the methodology of combining visual-analytical approaches and algorithmic solutions of machine learning. The main attention is paid to how the integrated use of visualization at each stage of developing a classification model allows to reveal hidden patterns, verify AI solutions and minimize the risks associated with erroneous predictions in critical areas (for example, diabetes diagnostics). The article considers a comprehensive approach to building a machine learning model for diabetes diagnostics. The main attention is paid not only to the accuracy of algorithms, but also to visual verification methods at each stage from data cleaning to assessing the risks of erroneous predictions. This approach allows to minimize the “black box” effect and increase trust in automated classification systems, in particular for medical diagnostics.
- Open Access
- 1 January 2026
- Published by Taras Shevchenko National University of Kyiv in Journal of Numerical and Applied Mathematics
- No. 1,p. 33-44
On consistency for the estimator of impulse response function
The aim of the article is to construct and analyze a statistical estimator for the real-valued impulse response function (IRF) of a time-invariant continuous linear system, and to establish its properties of asymptotic unbiasedness and consistency. The estimation of the response function is performed using a sample inputoutput cross-correlogram approach. The input signal is modeled as a stationary zero-mean Gaussian stochastic process, represented as a trimmed Fourier series using a trigonometric orthonormal basis. The study employs functional analysis and probability theory to derive upper bounds for the mathematical expectation and variance of the proposed estimator. The study establishes that the proposed integral cross-correlogram estimator is asymptotically unbiased as the cutoff level N approaches infinity. Furthermore, by evaluating the upper bounds of the bias and variance, it is proven that the estimator is consistent in the mean square sense as both the cutoff level N and the averaging interval length T approach infinity. The proposed approach offers a rigorous mathematical framework for system identification applicable to signal processing, automatic control, econometrics, and oceanology. The use of a Fourier series representation makes this method particularly effective for analyzing linear systems driven by periodic or quasi-periodic input signals.
- 1 January 2026
- Published by Taras Shevchenko National University of Kyiv in Journal of Numerical and Applied Mathematics
- No. 1,p. 33-44
Open AccessOn consistency for the estimator of impulse response function
The aim of the article is to construct and analyze a statistical estimator for the real-valued impulse response function (IRF) of a time-invariant continuous linear system, and to establish its properties of asymptotic unbiasedness and consistency. The estimation of the response function is performed using a sample inputoutput cross-correlogram approach. The input signal is modeled as a stationary zero-mean Gaussian stochastic process, represented as a trimmed Fourier series using a trigonometric orthonormal basis. The study employs functional analysis and probability theory to derive upper bounds for the mathematical expectation and variance of the proposed estimator. The study establishes that the proposed integral cross-correlogram estimator is asymptotically unbiased as the cutoff level N approaches infinity. Furthermore, by evaluating the upper bounds of the bias and variance, it is proven that the estimator is consistent in the mean square sense as both the cutoff level N and the averaging interval length T approach infinity. The proposed approach offers a rigorous mathematical framework for system identification applicable to signal processing, automatic control, econometrics, and oceanology. The use of a Fourier series representation makes this method particularly effective for analyzing linear systems driven by periodic or quasi-periodic input signals.
- 9 October 2025
- Published by State Scientific Institution - Ukrainian Institute of Scientific and Technical Expertise and Info in Science Technologies Innovation
- No. 3(35),p. 61
Comparative analysis of imputation methods in machine learning models
Missing data is a prevalent issue in machine learning and data analysis that impacts the credibility and performance of predictive models. This article provides a comprehensive study of missing data, its types, consequences, and popular imputation methods. Using real datasets, we compare the performance of Mean/Median Imputation, K-Nearest Neighbors (KNN) Imputation, Multiple Imputation, Regression Imputation, and Hot Deck Imputation. Furthermore, we study how these imputation techniques affect machine learning models such as Random Forest, Gradient Boosting Machines (GBM), and Support Vector Machines (SVM). Our study emphasizes the need for careful experimentation and model-specific investigation when handling missing data, where an important part is played by the selection of suitable imputation techniques based on dataset attributes and machine learning models. Lastly, our findings underscore the importance of tailored imputation strategies in enhancing model fit and ensuring stable analytical findings.
- 9 October 2025
- Published by State Scientific Institution - Ukrainian Institute of Scientific and Technical Expertise and Info in Science Technologies Innovation
- No. 3(35),p. 61
Comparative analysis of imputation methods in machine learning models
Missing data is a prevalent issue in machine learning and data analysis that impacts the credibility and performance of predictive models. This article provides a comprehensive study of missing data, its types, consequences, and popular imputation methods. Using real datasets, we compare the performance of Mean/Median Imputation, K-Nearest Neighbors (KNN) Imputation, Multiple Imputation, Regression Imputation, and Hot Deck Imputation. Furthermore, we study how these imputation techniques affect machine learning models such as Random Forest, Gradient Boosting Machines (GBM), and Support Vector Machines (SVM). Our study emphasizes the need for careful experimentation and model-specific investigation when handling missing data, where an important part is played by the selection of suitable imputation techniques based on dataset attributes and machine learning models. Lastly, our findings underscore the importance of tailored imputation strategies in enhancing model fit and ensuring stable analytical findings.
- Open Access
- 5 January 2025
- Published by Austrian Statistical Society in Austrian Journal of Statistics
- Vol. 54 (1) , 200-213
Statistical Estimation and Hypothesis Testing on Impulse Response Function
In this paper a time-invariant continuous linear system is considered with a real-valued impulse response function (IRF) which is defined on a bounded domain. A sample input- output cross-correlogram is taken as an estimator of the response function. The input processes are supposed to be zero-mean stationary Gaussian process and can be repre- sented as a finite sum with uncorrelated terms. A rate of convergence of IRF estimator in the space L2([0,Λ]) is obtained that gives a possibility to propose a nonparametric goodness-of-fit testing on IRF.
- 5 January 2025
- Published by Austrian Statistical Society in Austrian Journal of Statistics
- Vol. 54 (1) , 200-213
Open AccessStatistical Estimation and Hypothesis Testing on Impulse Response Function
In this paper a time-invariant continuous linear system is considered with a real-valued impulse response function (IRF) which is defined on a bounded domain. A sample input- output cross-correlogram is taken as an estimator of the response function. The input processes are supposed to be zero-mean stationary Gaussian process and can be repre- sented as a finite sum with uncorrelated terms. A rate of convergence of IRF estimator in the space L2([0,Λ]) is obtained that gives a possibility to propose a nonparametric goodness-of-fit testing on IRF.
- Open Access
- 1 January 2025
- Published by Taras Shevchenko National University of Kyiv in Bulletin of Taras Shevchenko National University of Kyiv. Series: Physics and Mathematics
- Vol. 81 (2) , 7-11
Pavlo Knopov: The scientific path and the formation of the Kyiv school of stochastic optimization
The article is dedicated to the 85th anniversary of Pavlo Solomonovych Knopov (b. May 21, 1940, Kyiv) – Doctor of Physical and Mathematical Sciences, Corresponding Member of the National Academy of Sciences of Ukraine; a leading specialist in stochastic optimization, statistics of random processes and fields, risk theory and control; author of over 250 scientific works and 13 monographs. Since 1999, he has headed the Department of Mathematical Methods of Operations Research at the V. M. Glushkov Institute of Cybernetics of the NAS of Ukraine, and since 1987 he has been a professor at Taras Shevchenko National University of Kyiv. He is a laureate of the State Prize of Ukraine in Science and Technology (2009), the V. M. Glushkov Prize (1997), the V. S. Mikhakevich (2021), and was awarded the title of Honored Scientist and Engineer of Ukraine (2021). The Kyiv school of stochastic optimization, to the development of which P. S. Knopov made a significant contribution, represents a scientific phenomenon that combines probability theory, mathematical statistics, and control theory for solving decision-making problems under uncertainty. The article also outlines the difficult life and scientific path of the scholar, who went through the trials of his time and established himself as one of the leading mathematicians of his generation.
- 1 January 2025
- Published by Taras Shevchenko National University of Kyiv in Bulletin of Taras Shevchenko National University of Kyiv. Series: Physics and Mathematics
- Vol. 81 (2) , 7-11
Open AccessPavlo Knopov: The scientific path and the formation of the Kyiv school of stochastic optimization
The article is dedicated to the 85th anniversary of Pavlo Solomonovych Knopov (b. May 21, 1940, Kyiv) – Doctor of Physical and Mathematical Sciences, Corresponding Member of the National Academy of Sciences of Ukraine; a leading specialist in stochastic optimization, statistics of random processes and fields, risk theory and control; author of over 250 scientific works and 13 monographs. Since 1999, he has headed the Department of Mathematical Methods of Operations Research at the V. M. Glushkov Institute of Cybernetics of the NAS of Ukraine, and since 1987 he has been a professor at Taras Shevchenko National University of Kyiv. He is a laureate of the State Prize of Ukraine in Science and Technology (2009), the V. M. Glushkov Prize (1997), the V. S. Mikhakevich (2021), and was awarded the title of Honored Scientist and Engineer of Ukraine (2021). The Kyiv school of stochastic optimization, to the development of which P. S. Knopov made a significant contribution, represents a scientific phenomenon that combines probability theory, mathematical statistics, and control theory for solving decision-making problems under uncertainty. The article also outlines the difficult life and scientific path of the scholar, who went through the trials of his time and established himself as one of the leading mathematicians of his generation.
- Open Access
- 1 January 2024
- Published by Taras Shevchenko National University of Kyiv in Bulletin of Taras Shevchenko National University of Kyiv. Series: Physics and Mathematics
- Vol. 79 (2) , 5-9
Mykhaylo Savchuk (19.05.1947–28.08.2024): a life dedicated to cryptography
Dedicated to the memory of the outstanding Ukrainian scientist, corresponding member of the NAS of Ukraine, laureate of the State Prize in Science and Technology, professor of the National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" Mykhailo Mykolayovych Savchuk. The article analyzes the scientific achievements of the scientist, and his significant contribution to the development of national science, in particular in the field of cryptography. A list of key scientific works and awards received is also provided.
- 1 January 2024
- Published by Taras Shevchenko National University of Kyiv in Bulletin of Taras Shevchenko National University of Kyiv. Series: Physics and Mathematics
- Vol. 79 (2) , 5-9
Open AccessMykhaylo Savchuk (19.05.1947–28.08.2024): a life dedicated to cryptography
Dedicated to the memory of the outstanding Ukrainian scientist, corresponding member of the NAS of Ukraine, laureate of the State Prize in Science and Technology, professor of the National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" Mykhailo Mykolayovych Savchuk. The article analyzes the scientific achievements of the scientist, and his significant contribution to the development of national science, in particular in the field of cryptography. A list of key scientific works and awards received is also provided.
- Open Access
- 1 January 2024
- Published by Taras Shevchenko National University of Kyiv in Bulletin of Taras Shevchenko National University of Kyiv. Series: Physics and Mathematics
- Vol. 78 (1) , 147-153
Simulation of the fractional Brownian process with given accuracy and reliability
Background. Random process theory is being used more and more in various fields of science due to the high computing power of modern computers. However, it's often important to know how much we can rely on the models we use. Methods. This paper examines the modelling of the fractional Brownian motion with given accuracy and reliability. The modelling is based on Dzhaparidze and van Zanten series representation of the fractional Brownian motion. We consider the fractional Brownian motion as an input process to a time-invariant linear system with a real-valued square-integrable impulse response function, which is defined on the finite domain. Results. We prove the theorem that gives the conditions, specifically the value of the upper limit of the summing in the model, under which the obtained model approximates fractional Brownian motion with given accuracy and reliability taking into account the response of the system. Conclusions. For the proof, we use the properties of square-Gaussian stochastic processes.
- 1 January 2024
- Published by Taras Shevchenko National University of Kyiv in Bulletin of Taras Shevchenko National University of Kyiv. Series: Physics and Mathematics
- Vol. 78 (1) , 147-153
Open AccessSimulation of the fractional Brownian process with given accuracy and reliability
Background. Random process theory is being used more and more in various fields of science due to the high computing power of modern computers. However, it's often important to know how much we can rely on the models we use. Methods. This paper examines the modelling of the fractional Brownian motion with given accuracy and reliability. The modelling is based on Dzhaparidze and van Zanten series representation of the fractional Brownian motion. We consider the fractional Brownian motion as an input process to a time-invariant linear system with a real-valued square-integrable impulse response function, which is defined on the finite domain. Results. We prove the theorem that gives the conditions, specifically the value of the upper limit of the summing in the model, under which the obtained model approximates fractional Brownian motion with given accuracy and reliability taking into account the response of the system. Conclusions. For the proof, we use the properties of square-Gaussian stochastic processes.
- 20 December 2023
- Published by Springer Nature in Journal of Theoretical Probability
- Vol. 37 (2) , 1627-1653
On Convergence of the Uniform Norm and Approximation for Stochastic Processes from the Space ${\textbf{F}}_\psi (\Omega )$- 20 December 2023
- Published by Springer Nature in Journal of Theoretical Probability
- Vol. 37 (2) , 1627-1653
On Convergence of the Uniform Norm and Approximation for Stochastic Processes from the Space ${\textbf{F}}_\psi (\Omega )$