Fig. 5: Predictive outcome of machine learning. | Nature Communications

Fig. 5: Predictive outcome of machine learning.

From: Identifying microbiota community patterns important for plant protection using synthetic communities and machine learning

Fig. 5: Predictive outcome of machine learning.The alternative text for this image may have been generated using AI.

a, b Density curves of pathogen colonization with global minima used for splitting into “protected” (positive) and “non-protected” (negative) classes presented as vertical dashed lines. c, d Performances of classification algorithms compared to a random classification (i.e., “No Model”) based on presence/absence of Mini5SynCom members. e, f Root mean square errors (RMSE) of the regression algorithms with dashed lines corresponding to predictions based on the global average of pathogen colonization (“No Model”) based on presence/absence of Mini5SynCom members or absolute abundance of Mini5SynCom members (“colonization”). a, c, e Results derived from algorithms trained on the median of pathogen colonization for each treatment. b, d, f Results derived from algorithms trained on pathogen colonization of individual plants. RF random forest, GMLNet elastic net regularized generalized linear model.

Back to article page