A novel neural network for prediction of wood density time series

The use of artificial nueral networks (ANNs) has received increasing attention in the analysis and prediction of time series problems. Recent researches in prediction with ANNs suggest that ANNs can be a promising alternative to the traditional linear methods, such as autoregressive integrated moving average (ARIMA) modeling. This paper presents a novel neural network, the functional link neural network (FLNN), to solve the problem of optimum time series prediction for wood density. Results show that the neural network predictions are considerably more accurate than those of the traditional ARIMA models. The mean squared error, absolute mean error, and mean absolute percent error are all lower on average for the FLNN forecast than for the ARIMA.