In this paper we suggest using multilayer perceptron articial neural network to model andforecast the United States regional economic growth in 1997 - 2012. Based on the data onthe US regional domestic product we estimate a number of ANNs and show that the basicMLPspecications based as inputs on only previous values of dependent variables deliversignicantly spatially correlated prediction error terms. Hence we use the neighborhoodmatrix to construct spatially lagged dependent variable and include them into the set ofANNs' inputs, which appears to signicantly decrease the prediction error spatial correlationand improve the prediction quality of the ANNs.
временные ряды, нелинейная модель, искусственная нейронная сеть, ошибки прогнозирования, пространственная корреляция, ВРП, США, temporary ranks, nonlinear model, artificial neural network, forecasting errors, spatial correlation, GDP, USA
1. Timo Terasvirta, Dick van Dijk, Marcelo C. Medeiros, Linear models, smooth transition autoregressions, and neural networks for forecasting macroeconomic time series: A re-examination (2005).
2. N. Schanne, R. Wapler, A. Weyh, Regional unemployment forecasts with spatial interdependencies.(2010).
3. B. Kock, Timo Terasvirta Forecasting performances of three automated modelling techniques during the economic crisis 2007-2009 (2013).
4. Ping-Feng Pai, Wei-Chiang Hong, Forecasting regional electricity load based on recurrent support vector machines with genetic algorithms (2005).
5. Nikolas Kourentzes, Devon K. Barrow, Sven F. Crone, Neural network ensemble operators for time series forecasting (2014).
6. Che-Chiang Hsu, Chia-Yon Chen, Regional load forecasting in Taiwan-applications of artificial neural networks (2002).
7. Massimiliano Marcellinoa, James H. Stockb, Mark W. WatsoncA comparison of direct and iterated multistep AR methods for forecasting macroeconomic time series (2006).
8. Simon Haykin, Neural Networks: A Comprehensive Foundation (2nd Edition), Izdatel'stvo: Pearson Education, Inc., 2006.
9. Uossermen F., Neyrokomp'yuternaya tehnika: Teoriya i praktika - M.: Mir, 1992.
10. Hinton, G. E.; Osindero, S.; Teh, Y. (2006). "A fast learning algorithm for deep belief nets". Neural Computation 18 (7).
