Table 5.
Performance parameters of PLS regression models for the prediction of rosmarinic acid content CRA in Rosmarinus officinalis, folium. Reproduced with permission from Royal Society of Chemistry, Ref. [75].
| Spectrometer | NIRFlex N-500 | microPHAZIR | MicroNIR 2200 | ||||
|---|---|---|---|---|---|---|---|
| samples | 60 | 60 | 60 | ||||
| outliers | 6 | 8 | 4 | ||||
| CRA (w/w) range/% | 1.138–2.425 | 1.138–2.425 | 1.138–2.425 | ||||
| validation method | CV | TSV | CV | TSV | CV | TSV | |
| R 2 | 0.91 | 0.91 | 0.73 | 0.73 | 0.84 | 0.85 | |
| SECV/% | SEP/% | 0.072 | 0.069 | 0.12 | 0.11 | 0.091 | 0.11 |
| SECV/SEC | SEP/SEC | 1.46 | 1.43 | 1.28 | 1.24 | 1.55 | 2.09 |
| factors | 8 | 8 | 5 | 5 | 11 | 12 | |
| RPD | 3.27 | 3.41 | 1.88 | 2.06 | 2.46 | 2.14 | |
Abbreviations: CV—cross validations; TSV—test-set validation; SECV—standard error of cross validation; SEP—standard errors of prediction; SEC—standard error of calibration; CRA —rosmarinic acid content (w/w).