Skip to main content
. 2023 Apr 18;14:1128326. doi: 10.3389/fimmu.2023.1128326

Figure 5.

Figure 5

Model performance on dbase using different splitting methods. (A) ROC curve and (B) PR curve for models using uniform splitting. (C) ROC curve and (D) PR curve for models using strict splitting. The dashed red lines indicate performance of random guessing. ROC and PR curve for DLpTCR looks “linear”, because DLpTCR outputs a binary and not a continuous probability.