Differences

This shows you the differences between two versions of the page.

--- tips:model_selection_with_glmulti_and_mumin [2021/03/16 20:13] – Wolfgang Viechtbauer
+++ tips:model_selection_with_glmulti_and_mumin [2021/03/16 20:14] – Wolfgang Viechtbauer
@@ Line 167: / Line 167: @@
 {{ tips:model_selection_rel_imp.png?nolink }}
-The importance value for a particular predictor is equal to the sum of the weights/probabilities for the models in which the variable appears. So, a variable that shows up in lots of models with large weights will receive a high importance value. In that sense, these values can be regarded as the overall support for each variable across all models in the candidate set. The vertical red line is drawn at 0.8, which is sometimes used as a cutoff to differentiate between important and not so important variables, but this is again a more or less arbitrary division.
+The importance value for a particular predictor is equal to the sum of the weights/probabilities for the models in which the variable appears. So, a variable that shows up in lots of models with large weights will receive a high importance value. In that sense, these values can be regarded as the overall support for each variable across all models in the candidate set. The vertical red line is drawn at 0.8, which is sometimes used as a cutoff to differentiate between important and not so important variables, but this is again a more or less arbitrary division (and a cutoff of 0.5 has also been at times used/suggested).
 ==== Multimodel Inference ====