Predicting Natural and Chaotic Time Series with a Swarm-Optimized Neural Network

Natural and chaotic time series are predicted using an artificial neural network (ANN) based on particle swarm optimization (PSO). Firstly, the hybrid ANN+PSO algorithm is applied on Mackey-Glass series in the short-term prediction x(t + 6), using the current value x(t) and the past values: x(t - 6), x(t - 12), x(t - 18). Then, this method is applied on solar radiation data using the values of the past years: x(t - 1), . . . , x(t- 4). The results show that the ANN+PSO method is a very powerful tool for making predictions of natural and chaotic time series.