Comparing NN and GMDH methods for prediction of socio-economic processes. Bulgakova Oleksandra, Samoilenko Oleksandr. IWIM, Prague, 2007.
Article (in pdf)
Abstract. There are considered two different methods for prediction of socio-economic processes: the combinatorial GMDH algorithm and an artificial neural network. These methods are analyzed in the task of modeling of the Ukrainian gross domestic product (GDP) as dependent from input arguments (investments). An analysis and comparison of these methods showed interesting results that gives preconditions to use capabilities of neural networks jointly with the GMDH algorithms.
Keywords. Artificial neural networks, Group Method of Data Handling (GMDH), linear neural networks, method of least squares.
References.
- Dyakonov V.P., Kruglov V.V. – MATLAB 6.5 SP1 7/7 SP1/7 SP1 + Simulink 5/6. The instruments of artificial intelligence and bioinformatics. – Moskow.: SOLON - PRESS, 2006. – 456 p. (In Russian)
- Korotky S. Neural networks: fundamental regulations.
- Zaychenko Y.P. Foundations of the design of intellectual systems. – Kyiv, 2003. (In Ukrainian)
