This work is an application of Synaptic Delay Based Artificial Neural Networks to the prediction of sunspot activity, straight from the data, without any smoothing or preprocessing as other authors and techniques employ. The signal is simply introduced as it is to the network, sample by sample as time passes, and the network using trainable internal delay terms modeling the length of the synaptic connections, learns to perform all the temporal reasoning processes required for the prediction task through the application of Discrete Time Backpropagation. We test the validity of the approach with the real sunspot series where unpredictable noise is present and there is no explicit equation that determines the evolution in time.
展开▼
