Fig. 5
From: LightCUD: a program for diagnosing IBD based on human gut microbiome data
Optimizing the feature sets for WGS-based modules. The cyan dots denoted the AUC values for a different round of five-fold cross validation with the different number of features, and the blue dots represented mean values of the cyan dots in the same column. (a) Illustrated the WGS-based health vs IBD case and (b) illustrated the WGS-based UC vs CD case. For both the cases, AUCs increased with more features at the beginning and decreased after reaching the top values. Forty-nine features for WGS-based health vs IBD case and 12 features for UC vs CD case were best