Fig. 2: Cross-validation and feature selection analysis using the training data set of clinical samples (N_train = 209).

a The spot selection process. A heat-map (top left) is generated by plotting the cost function \(j_{{\mathrm{m}},\,{\mathrm{p}}}\) across the sensing membrane. The cross-validation performance, both MSLE and the coefficient of determination (R²), is then plotted against the number of spots selected based on \(j_{{\mathrm{m}},\,{\mathrm{p}}}\) (bottom). The optimal subset of spots (top right) is then selected based off the optimal quantification performance indicated by the solid red marker. b The condition selection process. Conditions are ranked based on an iterative elimination method (top left), and the cross-validation performance is plotted against the number of conditions input into the quantification network. The optimal subset of conditions (top right) is then selected based off the optimal quantification performance indicated by the solid red marker. c The cross validation results using the selected features, where the ground truth CRP concentration is plotted against the predicted CRP concentration. The marker color and shape represent the different reagent batch ID (RID) and the fabrication batch ID (FID), respectively. d Bland-Altman plot of the same cross-validation results, where the dashed red lines represent the ± standard deviation of the measurement difference from the tested VFAs.