Skip to main content
. 2010 Dec 31;11:612. doi: 10.1186/1471-2105-11-612

Figure 3.

Figure 3

Target site and mRNA level classifiers generalize to microarray datasets not seen during training. The ROC graphs show the performance of four target site level classifiers (A-D) and four mRNA level classifiers (E-H) trained on three microarray datasets and tested on the remaining one dataset. Dotted lines illustrate random prediction. The ROC score is shown in the box. TPR indicates True positive rate, while FPR indicates False positive rate. Figure headings state the dataset used for testing.