TABLE 9.
Optimal features selected for classification in order of frequency used. Four classifications were tested for each variable (12 comparison), where four features are used in each case. This produces forty-eight total features selected when using the feature set composed of all available features.
| Number of classifications using each feature | ||||
|---|---|---|---|---|
| Modality | Feature | Age | Genotype | Treatment |
| 2PEF Oviduct | Difference Variance | 3 | 3 | 3 |
| 2PEF Ovary | ASM / Energy | 2 | 3 | 2 |
| 2PEF Oviduct | ASM / Energy | 3 | 2 | 1 |
| 2PEF Ovary | Difference Variance | 1 | 1 | 2 |
| 2PEF Oviduct | Sum Average | 1 | 1 | 1 |
| 2PEF Oviduct | IDM | 1 | 1 | 1 |
| SHG Oviduct | Difference Variance | 1 | 1 | 1 |
| 2PEF Ovary | Correlation | 1 | 1 | 1 |
| 2PEF Ovary | Sum Average | 1 | 0 | 1 |
| SHG Oviduct | Alpha | 1 | 0 | 1 |
| WFI | Ovary Intensity | 1 | 1 | 0 |
| 2PEF Ovary | IDM | 1 | 0 | 1 |
| SHG Oviduct | ASM | 0 | 1 | 0 |
| 2PEF Ovary | α | 0 | 1 | 0 |