Here we present EdgeSHAPer, a workflow for explaining graph neural networks by approximating Shapley values using Monte Carlo sampling. In this protocol, we describe steps to execute Python scripts for a chemical dataset from the original publication; however, this approach is also applicable to any user-provided dataset. We also detail steps encompassing neural network training, an explanation phase, and analysis via feature mapping.
2022, STAR PROTOCOLS, Pages 101887- (volume: 3)
Protocol to explain graph neural network predictions using an edge-centric Shapley value-based approach (01a Articolo in rivista)
Mastropietro Andrea, Pasculli Giuseppe, Bajorath Jürgen
Gruppo di ricerca: Algorithms and Data Science