Training feedforward networks with the Marquardt algorithm
IEEE Transactions on Neural Networks
| Hi-index | 0.01 |
In this paper we present the general methodology and main issues related to the application of neural networks to paleoclimatic reconstruction problems. We establish the basic methodological framework, data selection, organization and their relation to neural networks' features. We also describe a skill score to compare regressors' performance and finally the paleoclimatic variable's reconstruction. We show a case study focused on winter precipitation reconstruction in the Mediterranean back to 1700, using multi-layer perceptrons, and the comparison of the obtained results to that of the existing alternative methodologies.