Table 3.
Performance parameters of prediction of blood constituents by partial least squares (PLS) regression models developed for IR and NIR spectra of a 5-component model mixture in artificial dialysate solutions. Reproduced in agreement with CC BY 4.0 license, from Ref. [33].
| Model | Factor | R 2 | RMSECV in mg/dL | RMSEP in mg/dL | LODmin in mg/dL | LODmax in mg/dL | LOQmin in mg/dL | LOQmax in mg/dL | ||
|---|---|---|---|---|---|---|---|---|---|---|
| Urea | CV | NIR | 4 | 0.97 | 12 | - | 10 | 24 | 29 | 72 |
| IR | 4 | 0.99 | 7.9 | - | 10 | 18 | 31 | 55 | ||
| TV | NIR | 4 | 0.98 | - | 19 | - | - | - | - | |
| IR | 5 | 0.99 | - | 6.6 | - | - | - | - | ||
| Glucose | CV | NIR | 4 | 0.89 | 37 | - | 36 | 73 | 108 | 218 |
| IR | 3 | 0.96 | 22 | - | 47 | 142 | 140 | 428 | ||
| TV | NIR | 4 | 0.86 | - | 54 | - | - | - | - | |
| IR | 2 | 0.99 | - | 11 | - | - | - | - | ||
| Lactate | CV | NIR | - | - | - | - | - | - | - | - |
| IR | 5 | 0.95 | 8.2 | - | 28 | 90 | 84 | 271 | ||
| TV | NIR | - | - | - | - | - | - | - | - | |
| IR | 8 | 0.99 | 3.0 | - | - | - | - | - | ||
| Phosphate | CV | NIR | - | - | - | - | - | - | - | - |
| IR | 8 | 0.99 | 1.1 | - | 1.0 | 2.6 | 3.0 | 7.9 | ||
| TV | NIR | - | - | - | - | - | - | - | - | |
| IR | 8 | 0.95 | - | 2.0 | - | - | - | - | ||
| Creatinine | CV | NIR | - | - | - | - | - | - | - | - |
| IR | 5 | 0.98 | 1.8 | - | 2.6 | 4.5 | 7.9 | 13 | ||
| TV | NIR | - | - | - | - | - | - | - | - | |
| IR | 4 | 0.96 | - | 2.1 | - | - | - | - | ||
Abbreviations: RMSECV—root mean square error of cross-validation; RMSEP—root mean square error of prediction; LOD—limit of detection; LOQ—limit of quantification; CV—cross validation; TV—test-set validation; R2 —coefficient of determination.