Machine learning models for predicting the number of COVID-19 patients in Ukraine and India

Models for predicting the number of patients with COVID-19 using machine learning methods have been built. The data models are obtained from various official sources, including World Health Organization, beginning epidemic to present time. in Ukraine and India were selected teach COVID-19. Algorithms linear regression gradient boosting proved be that provided high accuracy forecast existing data. Data analysis was performed Python programming language Sklearn library which is based on SciPy (Scientific Python). In addition, XGboost boost algorithm used. To develop model, multifactor prediction time series delays as predictors chosen. It established such characteristics date event, day week, week number, month affect model. Model errors smallest estimated best values 0.83 0.75 India. built allow predict epidemiological situation future, coordinate actions different areas health care carry out reasonable preventive measures at state level.