Fig. 4: Receiver operating characteristic and precision-recall curve comparison across models and datasets.
From: A deep learning model to predict glioma recurrence using integrated genomic and clinical data
Receiver operating characteristic (ROC) and precision-recall (PR) curves comparing LUNAR and the traditional baseline models’ performance on the TCGA (a, b) and GLASS (c, d) datasets. NAtt no attention, CAtt cross-attention only, SAtt self-attention only. *Indicates that a given model predicted one class only (no discriminative power).
