Skip to main content
. 2005 May;15(5):724–736. doi: 10.1101/gr.2807605

Figure 4.

Figure 4.

Test performance of SPLP (red) and t-rank (blue) algorithms classifying real and randomized label sets. Test logodds ratio (LOR) is plotted in 20 cross-validation runs against fibrates, statins, azoles, and toxicant label sets and in 100 cross-validation runs against 100 different random label sets. An LOR of zero, which is expected from random chance, is marked with a dotted line.