Table 3.
Validation mean accuracy and standard error for all the relevant combinations, separated into five blocks.
| Input space | Accuracy |
|---|---|
| Model B + model E | 67.14 ± 0.31% |
| Model C + model E | 67.24 ± 0.64% |
| Model BC + model E | 69.12 ± 0.50% |
| Model B + fibre E | 77.08 ± 0.73% |
| Model C + fibre E | 88.30 ± 0.45% |
| Model BC + fibre E | 87.40 ± 0.72% |
| Fibre B + model E | 62.56 ± 0.51% |
| Fibre C + model E | 62.40 ± 0.45% |
| Fibre BC + model E | 62.80 ± 0.55% |
| Fibre B + fibre E | 77.14 ± 0.60% |
| Fibre C + fibre E | 88.66 ± 0.37% |
| Fibre BC + fibre E | 87.84 ± 0.42% |
| Model E | 59.06 ± 0.76% |
| Fibre E | 83.76 ± 0.32% |
Notes. Each input space undergoes a min-max normalization. The table’s first two blocks combine the model B, C, and BC cases with either model E or fibre E, respectively; the third and fourth blocks do the same but for the fibre B, C, or BC cases; and the last block shows what happens for model E or fibre E alone. As in Fig. 3, the source distribution among the five validation folds is fixed to be the same.
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.