2.англ.
Ph.D. thesis: 203 p., 4 tables, 63 figures, 5 appendixes, 100 references.
INFORMATION TECHNOLOGY FOR FORECASTING TIME SERIES OF THE NUMBER OF PATIENTS WITH CORONAVIRUS USING MACHINE LEARNING METHODS
The goal of the work is to increase the accuracy of forecasting the number of coronavirus patients in the short-term using machine learning methods.
The object of the study is to forecast the number of coronavirus patients in different regions.
The subject of the research is information technology, methods, algorithms, and programs for forecasting the number of coronavirus patients.
Research methods include general scientific research methodology and the principles of a system approach, namely: analysis of literary sources, open data, experimental research, object-oriented programming, system analysis, time series theory methods, machine learning methods, and other intellectual methods.
The COVID-19 pandemic has made the task of forecasting the incidence of infectious diseases and the coronavirus urgent. But, considering the number of factors that affect the spread of the disease and the chaotic nature of their values, there is a need to optimize forecasting methods. There are already time series forecasting models created to solve such a typical problem, but this infectious disease creates new challenges, such as the optimal identification of forecasting model parameters, as well as the ability to characterize the multi-wave nature of the morbidity time series. In addition to actual disease forecasting in one or another region (or country), it is also worth paying attention to the analysis of the possible infectious impact of neighboring regions, because this directly affects the object of research.
The scientific provisions obtained during the writing of the Ph.D. thesis contributed to the development and creation of a set of Python programs that automate calculations using all the proposed methods and approaches and implement the construction and use of the developed models. In the Google platform Kaggle, 10 notebook programs co-authored with a scientific supervisor were published in open access, which were viewed more than 36 thousand times in 2020-2023 (as of 07.12.2023). The main scientific results and practical developments of the dissertation were approved at 4 scientific conferences, including at the 2nd international meetings at the National Academy of Sciences of Ukraine in Kyiv, and at meetings of the Working Group on Mathematical Modeling of Problems Related to the SARS-CoV-2 Coronavirus Epidemic in Ukraine: during 2020-2022, as a result of which they were included in 25 reports of this working group, which is confirmed by the act of implementation from the Institute of Problems of Mathematical Machines and Systems of the National Academy of Sciences of Ukraine.
