Add Policy Tree class for learning policies based on IRM

[OliverSchacht]

With this PR I would like to add a Policy Tree implemenation.

Description

• Adding a DoubleMLPolicyTree class that fits a weighted classification of arbitrary covariates on an orthogonal basis.
• Adding a policy_tree() method to the IRM model, that fits and returns a Policy Tree based on the ATE score.
• Adding unit tests for both the new class and the new method.

Reference to Issues or PRs

Development of new features.

PR Checklist

• The title of the pull request summarizes the changes made.
• The PR contains a detailed description of all changes and additions.
• References to related issues or PRs are added.
• The code passes all (unit) tests.
• Enhancements or new feature are equipped with unit tests.
• The changes adhere to the PEP8 standards.

[SvenKlaassen]

• remove the sentence "However, they ..." from the depth documentation. This should be part of the docs.
• "learnt" to "learned" ? Maybe easier with US English
• Line 488. groups should be x_vars. Maybe rename x_vars to covariates or features.
• is min_samples_leaf=5 sufficient? I think, we should increase it slightly.
• Line 486. Change "Default is 2" to "Default is 2."

[SvenKlaassen]

• See documentation comments in double_ml_irm.py

[SvenKlaassen]

Pls remove

[SvenKlaassen]

What are you testing specifically. Would this not be better at test_doubleml_return_types ?

[OliverSchacht]

Thank you for your review @SvenKlaassen
I adapted the following changes:

• rename x_vars to features
• set min_samples_leaf to 8
• move policytree return type test from test_irm.py to test_doubleml_return_types.py
• fixed minor spelling / documentation issues you mentioned