Individual Conditional Expectation Curve

Just like the Partial Dependence Plot but without averaging over the entire dataset. This allows you to see heterogenous effects in the data. It yields a family of local explanations. By aligning them vertically at some value we can get somewhat global explanations.