Search papers, labs, and topics across Lattice.
The paper introduces asymmetric Shapley values as a method for quantifying feature importance in high-dimensional prediction models, particularly in clinical settings where collinearity and directional dependencies exist. They derive efficient algorithms for computing both local and global asymmetric Shapley values, focusing on scenarios where a disease state mediates genomic effects. The framework is demonstrated through the prediction of progression-free survival for colorectal cancer patients, showcasing the utility of local values for inference and global values for performance decomposition.
Asymmetric Shapley values offer a more robust and interpretable approach to feature importance in clinical prediction by accounting for collinearity and known directional dependencies, overcoming limitations of traditional methods.
In clinical prediction settings the importance of a high-dimensional feature like genomics is often assessed by evaluating the change in predictive performance when adding it to a set of traditional clinical variables. This approach is questionable, because it does not account for collinearity nor known directionality of dependencies between variables. We suggest to use asymmetric Shapley values as a more suitable alternative to quantify feature importance in the context of a mixed-dimensional prediction model. We focus on a setting that is particularly relevant in clinical prediction: disease state as a mediating variable for genomic effects, with additional confounders for which the direction of effects may be unknown. We derive efficient algorithms to compute local and global asymmetric Shapley values for this setting. The former are shown to be very useful for inference, whereas the latter provide interpretation by decomposing any predictive performance metric into contributions of the features. Throughout, we illustrate our framework by a leading example: the prediction of progression-free survival for colorectal cancer patients.
