Salim Amoukou

Hi @samkodes @martinju In this paper [https://arxiv.org/pdf/2106.03820.pdf], you will find an empirical and theoretical analysis of the bias of the dependent TreeSHAP agorithm. It shows that the TreeSHAP algorithm is...

Hi, sorry for the late response, First of all, thank you for all the interesting references. The SHAFF approach seems very specific to RandomForest (not generalizable to boosting trees for...

Sorry for the late reply. Could you share the errors? Python version?

Sorry for the late reply. In general, "-1" means that there are no features, and here it was used to initialize the lists in the code in Cython. Just ignore...