Figure 1

Schematic overview of the systematic approach to assess the classification performance across independent data sets. The role of each data set is exhaustively alternated between training and testing. N equals 4 in case of the estrogen receptor status and equals 3 in case of the histological grade, see Table 1.