A new perspective on Shapley values, part I: Intro to Shapley and SHAP by Edden Gerber

This post is the first in a series of two posts about explaining statistical models with Shapley values.

There are two main reasons you might want to read it:
1. To learn about Shapley values and the SHAP python library.
This is what this post is about after all. The explanations it provides are far from exhaustive, and contain nothing that cannot be gathered from other online sources, but it should still serve as a good quick intro or bonus reading on this subject.
2. As an introduction or refresher before reading the next post about Naive Shapley values.
The next post is my attempt at a novel contribution to the topic of Shapley values in machine learning. You may be already familiar with SHAP and Shapley and are just glancing over this post to make sure we’re on common ground, or you may be here to clear up something confusing from the next post.

Link: Post

Navigational hashtags: #armknowledgesharing #armsites
General hashtags: #shap #shapley #interpretation #ml #python

@data_science_weekly