A necessary competence in the present-day reality is the ability to analyze big data in the economy, and therefore one of the key issues is the choice of tools for such analysis. One of the most promising tools is nowcasting, which allows you to accurately determine economic changes in very short time periods. The aim of the study is to analyze successful modern practices of using nowcasting for statistical forecasting of socio-economic indicators. The hypothesis of the research lies in the assumption that nowcasting as a method of macroeconomic analysis can in the near future become a worthy alternative to traditional methods of analysis and statistical forecasting of indicators of socio-economic development, increasing the accuracy of their forecasting. The methodological basis of the study was the scientific works and applied developments of leading domestic and foreign scientists in the field of economic forecasting using statistics of search queries, as well as methods of comparative and statistical analysis, and the systematic approach. The novelty of the results obtained lies in the systematization and description of successful practices of using nowcasting and forecasting indicators using query statistics. The study highlights the basic principle of nowcasting, which is to obtain a more accurate assessment of the state of the economy as new data becomes available. It also describes the key statistical models used as tools for testing in foreign countries. As a result of the study, we highlight the importance of the analysis of statistical search queries, especially in the context of their correlation with classical survey metrics and general statistics. It is in an active phase of development, especially within the framework of the domestic forecasting practice. The results obtained can be applied both in a corporate environment and in the public sector to build macroeconomic forecasts.
Connecting reservoir computing with statistical forecasting and deep neural networks
