Time Series Long-term Forecasting Using Tensor Product Functional Link Neural Network

In this paper, a tensor product functional link neural network (TP-FLNN) was applied on the Mackey-Glass chaotic time series in the long-term forecasting. The forecasting performance of TP-FLNN was compared with simple forecasting methods namely naive, drift and average methods. The obtained results show the possibility to consider TP-FLNN in long-term forecasting.