Make pip-installable

Open ivanistheone opened this issue 7 years ago • 0 comments

This is a very cool demo! It would be nice to build out more and make it pip-installable + add docs for how to reuse the base class.

I'm thinking this could be a great tool for teaching:

introduce general ML concepts via scikit learn API
show more advanced Bayesian topics using "pictures" (e.g. using tikz-bayesnet) and matching STAN description

For reference, here is the pymc3_models base class https://github.com/parsing-science/pymc3_models/blob/master/pymc3_models/models/init.py#L10 and an example Logistic Regression subclass https://github.com/parsing-science/pymc3_models/blob/master/pymc3_models/models/LinearRegression.py#L11

Dec 27 '18 16:12 ivanistheone