Question 6/10 v4 lecture 7

What is a partial dependence plot?


It is a plot, constructed from a Random Forest classifier, that describes how a dependent variable varies as we alter one of the independent variables. It aims to hold all other things constant. For a description of how it is constructed please consult the referenced part of the lecture (essentially, we iteratively set the column of interest to all of its unique values, storing results across runs).

Relevant part of lecture