Skip to main content
Fig. 7 | BMC Bioinformatics

Fig. 7

From: Interpretable prediction of necrotizing enterocolitis from machine learning analysis of premature infant stool microbiota

Fig. 7

Feature importance based on analysis of high-attention weight instances. For each patient, the highest attention weight instance was identified and used to assess feature importance using random forest analysis. The 15 features with the highest scores are shown for the taxonomic levels of phylum, class, and family. Features that appeared in analyses from both studies are displayed as mean importance, while features that appeared in only one study are indicated with a superscript w for Warner et al. or o for Olm et al. For complete feature importance lists from both studies see Additional file 7: Data S2

Back to article page