Skip to main content
. 2022 Apr 11;18(4):e1010050. doi: 10.1371/journal.pcbi.1010050

Fig 6. Comparison of different multimodal and single-modality models for colorectal cancer prediction.

Fig 6

Values are test ROC AUC, error bars are standard error obtained by repeating the experiment five times on different random train/test splits. RF: Random Forest. A: species-relative abundance profiles. M: strain-level marker profiles. Metabolic: metabolite profiles. For MVIB, the JMVIB objective has been adopted for the optimisation (Eq 5). Experiments are executed on the Colorectal-Metabolic dataset extracted from [37].