Aplication of artificial neural networks to macroeconomic model: A Turkey study

Artificial neural networks that mimic the human brain have a significantplace in artificial intelligence studies. In recent years, artificial neuralnetworks have found a lot of application areas and have been used widely inforecasting.In this thesis, inflation forecasting results obtained by employment ofartificial neural networks are studied. In the study, the feed forward neuralnetwork architecture and back propagation algorithms are employed. The 12month data for the interval 1987-2005 are separated as test data and traindata. The data is used after normalization as input for the network. Theforecasting results for the inflation obtained by the help of the artificial neuralnetwork techniques are compared with the results of VAR method. In thisway, a comparison of artificial neural technique with an econometricforecasting method is introduced. It is shown that the neural networks can beused as a good alternative for other forecasting modeling techniques.