I've developped a glm-model and a random-forest on the same data with Tidymodels (classification).
Would it be useful to optimize these models by comparing the predictions-weights ("probs", I think in caret) where rf and glm divert? So, if glm-pos > rf-neg the outcome is glm, else rf?
And if so, would it be valid to optimize the outcome by correcting the probs with the accuratesse-factor?