This repository contains the codes and slides for my talk in NUS on 16th October, where I talked about various ways to dissect/interpret black box machine learning models. The PeekingIntoTheBlackBox (jupyter) notebook contains the main python codes for my demo. The slides (in html) are generated by a jupyter notebook Reveal.js slideshow extension and the pdf is generated from the html.
The pdp_wine.R is a R script where I generated the Partial Dependence Plot (PDP) using the pdp R library.
Feel free to raise an issue if you encounter any difficulty in reproducing the results or have any questions in general. Thanks.