Search Results for author:
Found 6 papers, 5 papers with code
Model free variable importance for high dimensional data
Here we present some model-free methods that do not require access to the prediction function.
Variable importance without impossible data
The most popular methods for measuring importance of the variables in a black box prediction algorithm make use of synthetic inputs that combine predictor variables from multiple subjects.
Deletion and Insertion Tests in Regression Models
We find an expression for the expected value of the AUC under a random ordering of inputs to $f$ and propose an alternative area above a straight line for the regression setting.
Additive models
Explainable Artificial Intelligence (XAI)
+1
What makes you unique?
We use Shapley value to combine all of the reductions in log cardinality due to revealing a variable after some subset of the other variables has been revealed.
Cohort Shapley value for algorithmic fairness
Cohort Shapley value is a model-free method of variable importance grounded in game theory that does not use any unobserved and potentially impossible feature combinations.
Explaining black box decisions by Shapley cohort refinement
Instead of changing the value of a predictor we include or exclude subjects similar to the target subject on that predictor to form a similarity cohort.
