A Feedback-oriented Data Delay Modeling in a Dynamic Neural Network for Time Series Forecasting

Abstract

Abstract. In this study, we develop a neural network with a time shifting approach to forecast time series patterns. We investigate the impact of di erent layer-weight con gurations to capture the trends in seasonal, chaotic, etc. forms. We also hypothesize the combined e ect of the delayed inputs and the forward connections to introduce a dynamical structure. The e ect of over tting issue is procedurally monitored to gain the resistance property from the early stoppage of training process and to reduce the error of predictions. Finally, the performance of the proposed network is challenged by six well-known deterministic and non-deterministic time series and compared by the autoregression (AR), Arti cial Neural Network (ANN), Adaptive K-nearest Neighbors (AKN), and adaptive neural network (ADNN) models. The results show that the proposed network outperforms the conventional models, particularly in forecasting the chaotic and seasonal time series.