Forecasting inflation using econometric and artificial neural network models
Artificial neural network modelling has recently attracted much attention as a new technique for estimation and forecasting in economics and finance. The chief advantages of this new approach are that such models are able to solve ery complex problems, and that they are free from the assumption of linearity that is often adopted to make the traditional methods tractable. In this research I compare the performance of the Artificial Neural Network (ANN) models with the traditional econometric approaches to forecasting the inflation rate. Of the ANN models I apply a back-propagation neural network (BPN) model, a radial basis function network (RBFN) model, and a Recurrent Network (RN) model. Of the traditional econometric models I use a structural reduced form model, an ARIMA model, a vector autoregressive model, and a Bayesian vector autoregression model. I use the econometric models as a guide to design the ANN models and compare each econometric model with an ANN model which uses the same set of variables. Static and dynamic forecasts are compared for three different horizons: one, three and twelve months ahead. Root mean squared errors and mean absolute errors as well as the information test method are used to compare quality of forecasts. The results show the ANN models able to forecast as well as all the traditional econometric methods, and to outperform them in some cases.