What method is used to calculate field importance for the AI Studio models and ensembles?
The importance is also known as relative error reduction. For each split in a model AI Studio has an estimate of the prediction error that the split helps reduce (AI Studio uses these same estimates when pruning). To get the importance AI Studio sums the error reduction for every split but grouped by the field. So, we get a sum of error reduction for each field. AI Studio normalizes those, so that they total to one, and that gives the importance values.