Figure 3

Chlorophyll content prediction model (one of the four partitions of training/validation data sets). Training (●) (n = 57) and validation (▲) (n = 19) data are represented in log10 mg/cell. Model performance parameters: variance captured (Variance); root mean square error of cross-validation (RMSECV); root mean square error of prediction (RMSEP); coefficients of determination (R2) and slopes of linear regression between observed and predicted data obtained respectively for the training and validation data sets; number of LVs used by the model.