Training set and cross validation errors of various models inferred through the informative data obtained from synthetic ‘AFH’ and ‘aWR’ targets (Fig. 3).

Three independent trials are conducted to infer both the synthetic targets in Fig. 3. The mean training set and cross validation errors for the eight selected ‘AFH’ models are 3.0% and 4.9% respectively. These for eight chosen models representing the ‘aWR’ target are 2.9% and 5.6%.