Table 4.
Confusion matrix for classification of images from the 3DHeLa dataset with a BPNN with 20 hidden units using SLF17. The results shown are averages over 10 cross-validation trials.
| True Class | Predicted Class | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| DNA | ER | Gia | Gpp | LA | Mit | Nuc | Act | TfR | Tub | |
| DNA | 98 | 2 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| ER | 0 | 100 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Giantin | 0 | 0 | 100 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| Gpp130 | 0 | 0 | 0 | 96 | 4 | 0 | 0 | 0 | 0 | 0 |
| LAMP2 | 0 | 0 | 0 | 4 | 95 | 0 | 0 | 0 | 0 | 2 |
| Mitochon-dria | 0 | 0 | 2 | 0 | 0 | 96 | 0 | 2 | 0 | 0 |
| Nucleolin | 0 | 0 | 0 | 0 | 0 | 0 | 100 | 0 | 0 | 0 |
| Actin | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 100 | 0 | 0 |
| TfR | 0 | 0 | 0 | 0 | 2 | 0 | 0 | 0 | 96 | 2 |
| Tubulin | 0 | 2 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 98 |
The average correct classification rate across all classes was 98% Results from reference (24).