A neural network to predict multiple economic time series
Keshav P. Vishwakarma · 1994
Presents a fresh approach to the analysis of multiple economic time series. To model their common movement, the author resorts to the state space framework of dynamic systems theory and makes use of a neural network to account for nonlinear evolution of the system over time. This mathematical formulation allows not only the identification of business cycles in given data but it furnishes out-of-sample forecasts as well.>