Fig. 5.

Distribution of ranks for each of the 15 algorithms applied to the AML dataset only, based on eight different combinations of metrics used in the generation of the summary quality score, repeated in 10,000 randomized iterations. Each box in the boxplot thus represents 80,000 values of rank. Algorithms fall into three categories (indicated with orange, blue, purple) based on their performance on the non-malignant cells. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)