Novel forecasting techniques using big data, network science and economics

Abstract

The combination of theoretical network approach with recently available abundant economic data leads to the development of novel analytic and computational tools for modeling and forecasting key economic indicators. The main idea of this study is to introduce a topological component into economic analysis, consistently taking into account higher-order interactions in the economic network. We present a multiple linear regression optimization algorithm to generate a relational network between individual components of national balance of payment accounts. Our model describes well annual country statistics using the explanatory power and best fits of related global financial and trade indicators. The proposed algorithm delivers good forecasts with high accuracy for the majority of the indicators.