Fig. 5: Random Forest Classifiers demonstrate high accuracy and precision when predicting ANXD and/or MDD diagnosis.
From: Medication use is associated with distinct microbial features in anxiety and depression
A Area under the receiver operating characteristic curve (AUC) is displayed for a diagnosis classifier distinguishing ANXD diagnosis, MDD diagnosis, or ANXD and MDD diagnosis from HC. Four types of feature selection were applied in addition to the raw data. B Random Forest Classifier was applied to only unmedicated participants. C Average precision (APR) scores were also calculated across the different classifications and feature selection methods. D APR scores in the unmedicated subset.
