AN INTEGRATED APPROACH TO FORECASTING GOVERNMENT BOND PRICES BASED ON EVENT ANALYSIS AND MACHINE LEARNING

The article considers an integrated approach to forecasting the price of Russian government bonds based on a combination of event analysis (assessing the impact of sanctions) and machine learning (forecasting based on economic indicators). The methods of the study are based on event analysis and machine learning, namely random forest, and alternative approaches are also considered: regression and LTSM. The first two methods turned out to be statistically insignificant for the presented data, so it was decided to use the third method, random forest, as the main method. As a result of the research, the authors present an integrated approach linking economic events and parameters, which will allow investors and financial regulators to reduce market uncertainty and competently optimize their portfolio.