“Dataset No”. indicates 7 dataset patterns of pinpoint data shown in Table 2. “Clustering” indicates the approximated interval of prediction in side effect incidence of drug. “Correct rate” indicates to what extent each example of learning model can meet each correct interval in the prediction