How do I implement the MIC correlation algorithm in python?

Question

The algorithm for finding correlation in non-linear data has some python resources, namely minepy and its tool for MIC analysis mictools. However, it all seems very obscure and it's unclear how I could use it in a python program to find the MIC correlation between two lists x and y.

Can anyone help with this?

Answer 1

The documentation on GitHub is terrible, however it does contain the one useful thing: the pointer to the paper by Reshef et al., which gives a fair bit of background and explanation (no, it is not a model of lucidity either, but much better than the alternatives).