# Credal classification

The aim of the project is for you to learn more about machine learning with bounded probability. It consists of 3 tasks:
    1. Further explore the breast cancer dataset that we introduced in the lectures.
    2. Derive some theoretical results to improve the probability bounds that we used in the lectures. Use this theoretical result to improve the credal classifier from the lectures.
    3. Derive some theoretical results concerning robust Bayes maximality and robust Bayes admissibility for the credal classifier. Use this theoretical result to further improve the credal classifier from the lectures.

Each task is subdivided in very specific subtasks, to guide you along. Most subtasks require some coding in Python. However, the 2nd and 3rd task also have subtasks that concern purely theoretical questions to be solved on pen and paper.

You may do all three tasks, or only a selection of them; this is up to you.