MOESM9 of Infrared spectroscopy coupled to cloud-based data management as a tool to diagnose malaria: a pilot study in a malaria-endemic country

Additional file 9: Fig. S6. Using field data from the pilot trial the effect of sample size on classification accuracy was studied. Cross validation classification performance was monitored by cross validation using successive PLS-DA models with increasing number of samples in the calibration data sets (up to n = 200; using 20 replicates in each case). Classification error decreases exponentially with sample size (n). Extrapolation of the trend line predicts very low error rate at n = 500.