Figure 3

Prediction accuracy and comparison of predictive models for tea quality-related metabolites. Data and error bars are the means ± standard deviations (10 replicates of tenfold cross-validation). RR ridge kernel regression, GAUSS Gaussian kernel regression.