Fig. B.3
Evolution of the SME for the training and validation samples during the training of the ANN. The SME decreases rapidly during the first iterations until convergence is reached. Since the training sample used (the DIAD models) do not have noise, no overfitting occurs (no increase is seen in the validation error with increasing iterations).
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.