Fig. 5
From: An individualized predictor of health and disease using paired reference and target samples
Heatmaps of predictor error rates for H3N2 DEE2 dataset. The heatmaps of the sample-specific predictor error rate for classifying states 2,3,4 and 5, defined in Fig. 2, for samples in the H3N2 DEE2 dataset. The figure shows the difference between the error rates of the standard predictor (no reference) and the error rates of the reference-aided predictor (with reference) for H3N2 DEE2. The entries in red have higher error rate using the standard predictor than the reference-aided predictor, and vice versa