SHAP doesn't work if LogNormal

[pkdmark]

Hello,

Problem can be reproduced with a slight modification to the notebook below:

https://github.com/stanfordmlgroup/ngboost/blob/master/examples/model_interpretation/SHAP_plots.ipynb

If the distribution is LogNormal (rather than normal) you will find that the Base Value + Shap Values of a single observation do not equal the prediction.

[Penna88]

I had a similar issue with Gamma dist and I commented as #291
Hope it helps.