Table 1 Summary of analytical performances of investigated calibration models.

From: A frugal implementation of Surface Enhanced Raman Scattering for sensing Zn2+ in freshwaters – In depth investigation of the analytical performances

Name

Type

Training set

Testing set

Recovery (%)

RMSEV (nM)

Average precision (%)

LOD (nM)

LOQ (nM)

U1

Univariate

10 titration series

10 titration series

89

397

11%

883

U2

Univariate

8 titration series

8 titration series

88

237

7%

575

M50

Multivariate (PLS)

10 titration series

10 titration series

96

138

4%

160

470

M40

Multivariate (PLS)

8 titration series

10 titration series

97

162

5%

253

532

M30

Multivariate (PLS)

6 titration series

10 titration series

98

246

7%

346

898

  1. Each titration series contains 16 spectra corresponding to 16 increasing Cj concentrations of Zn. For each model, the training and testing sets were built on data points with odd and even values of the concentration index j, respectively. The recovery rates, RMSEP, LOD and LOQ are determined from the \({C}_{predicted}\,vs.{C}_{reference}\) validation plots of each model. The average precision is estimated over 10 replicate measurements.