Towards an Intelligent Neural System for Economic and Business Forecasting

Today's business and macroeconomic management needs precise forecasting of the near future with only a handful of observations, yet most econometric techniques demand dozens of historical records. Here we outline the design of an intelligent system built upon implicit knowledge from computational neuroscience. Its main element is the gated dipole, a neural circuit model with adaptive oscillating property and describing equally well the brain's dynamic rebalancing in a changing environment, and an economic system's fluctuations due to internal and external factors. We show how that model can, in principle, use only two or three points to predict the next, when applied to loosely oscillating socioeconomic processes.