Fig. 4
From: A multimodal graph neural network framework for cancer molecular subtype classification
Performance of the Proposed Models and Baseline Models with Different Numbers of Genes on BRCA Dataset. (a) The accuracy of the proposed model with GAT (blue solid line) or GCN (orange solid line) and baseline models (dashed line) are plotted against different numbers of genes (300, 500, 700, 1000, 2000, and 5000) for BRCA subtype classification. (b) The F1 scores of the proposed model with GAT (blue solid line) or GCN (orange solid line) and baseline models (dashed line) are plotted against different numbers of genes (300, 500, 700, 1000, 2000, and 5000) for BRCA subtype classification