I had been reading about SHAP (Paper), (Code). Is it possible to use this in case of tabular data where I have a number of samples, each having a fixed number of features and single label class labels? I wish to get a rank of all features based on their contribution in classification.
I have seen how to use SHAP with scikit learn models. I am wondering how to integrate it with Keras/PyTorch which are used to build deep learning models. If this is not possible, can anyone please suggest any other method to do so?