An Statistical Research on Feed Forward Neural Networks for Forecasting Time Series

摘要: In recent years, artificial neural networks have being successfully used in time series analysis. Using linear methods such as ARIMA and exponential smoothing for non cannot produce satisfactory results. Although there are various methods, these an important drawback that all of them require a specific model assumption. On the other hand, no restrictions linearity or assumptions. many applications within analysis, it has been seen more accurate results than those obtained from traditional methods. spite fact provide some advantages, re- searchers keep working on component selection problem method. The answer question which compo- nents method should be is vital issue terms forecasting performance. this study, effects number hidden layer length test set performance examined. Eight real implementation. analyzed by using statistical analysis interpreted.