This repository contains well structured and easy to understand code written in Python programming language for real-world applications of machine learning (ML) explainability (also known as Explainable AI(XAI)). We will be uploading state-of-the-art topics and techniques in XAI with easy to understand coding examples. The purpose of this repository is show how explainability techniques can be employed in every stage of the MLOps lifecycle, for easy adoption and adaptation to real-world problems encountered by data scientists and machine learning engineers.
Our audience will consist of:
- Junior to senior data scientists
- Junior to senior machine learning/data engineers
- Data analysts
- End-users of machine learning solutions
- Domain experts and decision makers
The current version of our repository contains code for:
- Permutation Importance (PI): A very simple explainability technique used to obtain the importance of features based on their impact on a trained ML model’s prediction. In the directory for PI, we have demonstrated the use of PI in classification and regression to understand the importance of features.