WIP: Adding Likelihood Q Ratio Test (LQRT)

[adam2392]

Addresses: #83

Description: This is a PR that adds the LQRT robust statistical test, which is proven in [1] to be more robust in terms of power for fixed alpha level compared to both the t-test and the Wilcoxon rank-sum test. The python package at: https://github.com/alyakin314/lqrt implements this in Python in both 1-sample and 2-sample case with the possibility of unequal variances to be as general as possible.

Currently work in progress.

References:
[1]: Qin, Yichen & E. Priebe, Carey. (2017). Robust Hypothesis Testing via Lq-Likelihood. Statistica Sinica. 27. 10.5705/ss.202015.0441.

[adam2392]

Let me know what else would be necessary at this point to move this into the release stage. I would love to get this working to use for myself :p.

[josesho]

Hi @adam2392 ,

Tests are passing, great! Here's what's left to do.

TODO:

• Update release-notes.rst with a short description of the changes.
• Update version number (0.2.8) in setup.py and __init__.py.
• Add tests in test_01_effsizes_pvals.py to check that LqRT results (for simple 2-group paired and unpaired cases)