Table 2.
TPs, FNs, FPs, TNs, accuracy, precision, recall, and F1 computed for the five RF classifiers tested so far, including morphology and variability features (RF1) and optical and NIR colors (two to five from RF2 to RF5, respectively), and for the RFcol classifier, where only morphology and color features are used.
| TPs | FNs | FPs | TNs | Accuracy | Precision | Recall | F1 | |
|---|---|---|---|---|---|---|---|---|
| (/225) | (/225) | (/2000) | (/2000) | |||||
| RF1 | 208 ± 0 | 17 ± 0 | 3 ± 0 | 1997 ± 0 | 0.9909 ± 0.0002 | 0.98576 ± 0.00003 | 0.923 ± 0.002 | 0.953 ± 0.001 |
| RF2 | 208 ± 1 | 17 ± 1 | 3 ± 0 | 1997 ± 0 | 0.9908 ± 0.0005 | 0.996 ± 0.002 | 0.922 ± 0.004 | 0.953 ± 0.002 |
| RF3 | 208 ± 1 | 17 ± 1 | 3 ± 1 | 1997 ± 1 | 0.9912 ± 0.0005 | 0.988 ± 0.002 | 0.924 ± 0.003 | 0.955 ± 0.002 |
| RF4 | 211 ± 1 | 14 ± 1 | 1 ± 0 | 1999 ± 0 | 0.9929 ± 0.0005 | 0.99526 ± 0.00003 | 0.934 ± 0.005 | 0.964 ± 0.003 |
| RF5 | 212 ± 1 | 13 ± 1 | 1 ± 0 | 1999 ± 0 | 0.9935 ± 0.0002 | 0.99529 ± 0.00001 | 0.940 ± 0.002 | 0.967 ± 0.001 |
| RFcol | 201 ± 1 | 24 ± 1 | 11 ± 0 | 1989 ± 0 | 0.9842 ± 0.0004 | 0.948 ± 0.002 | 0.893 ± 0.004 | 0.919 ± 0.002 |
Notes. Each error is the standard deviation from the average value obtained from a set of ten simulations per classifier, where each classifier builds each time 500 trees to return the final classification for each source.
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.