Table S3.
Error rate (ER) for external validation of various classification methods for AMA vs MMA forms
| Observation | True label | PCA-LR | PCA-LDA | FPCA-LR | FPCA-LDA |
|---|---|---|---|---|---|
| 1 | 0 | 0 | 0 | 0 | 0 |
| 2 | 0 | 0 | 0 | 0 | 0 |
| 3 | 0 | 0 | 0 | 0 | 0 |
| 4 | 0 | 0 | 0 | 0 | 0 |
| 5 | 0 | 0 | 0 | 0 | 0 |
| 6 | 0 | 1 | 1 | 1 | 1 |
| 7 | 0 | 0 | 0 | 0 | 0 |
| 8 | 0 | 0 | 0 | 0 | 0 |
| 9 | 0 | 0 | 0 | 0 | 0 |
| 10 | 0 | 0 | 0 | 0 | 0 |
| ER (%) | – | 10 | 10 | 10 | 10 |
| Accuracy (%) | – | 90 | 90 | 90 | 90 |
Notes: True label = 0 for MMA observation and 1 for AMA observation. The 0/1 values in other columns represent the predicted labels for a given observation using the method for a specific column. For example, the value 0 in the first row under PCA-LR column means that by using the PCA-LR method we predicted that the first observation was MMA. For all the methods, the number of retained principal components (or functional principal components) was L=2. L=2 represents that only the first two principal components (or functional principal components) was retained.
Abbreviations: AMA, α-mycolic acid; FPCA, functional PCA; LDA, linear discriminant analysis LR, logistic regression; MMA, methoxy-MA; PCA, principal component analysis.