This add-in produces marginal model plots. Marginal model plots are used to assess the fit of a given model.
These are philosophically similar to the Marginal Model Plots produced under the Variable Importance outline in the Profiler, but they are calculated differently (based on the LOESS methodology), whereas those under Variable Importance are computed using a sobol-based decomposition method.
For more information about the LOESS methodology, consider the following references:
Jon Peltier (https://stats.stackexchange.com/users/17250/jon-peltier), What is the difference between simple linear model and loess model?, URL (version: 2015-02-21): https://stats.stackexchange.com/q/138665
NIST Engineering Statistics Handbook: Section 4.1.4.4 LOES (aka LOWESS)
https://www.itl.nist.gov/div898/handbook/pmd/section1/pmd144.htm
Practical Introduction to LOESS (or LOWESS), United States Environmental Protection Agency (EPA.gov)
https://www.epa.gov/sites/default/files/2016-07/documents/loess-lowess.pdf