Fig. 3
From: Machine learning approaches to predict the Plant-associated phenotype of Xanthomonas strains
Properties of the training data set and model prediction performances. Upper panel: Partial Least Squares Discriminant Analysis showing the discrimination between variables and phenotypic classes (a) Xanthomonas strains labelled by phenotype. (b) Xanthomonas strains labelled according to species classification. A standard PCA representation is available in supplementary file 8 Lower panel: Performance of the ML classifiers. (c) Classifier accuracy using a nested cross-validation scheme for each classification method. (d) Precision-Recall curves of each classification method calculated from the predicted median test-set probabilities