Building tools on the shoulders of Mosek.
We created this package to support the experiments given in the paper by Schmelzer, Hauser, Dahl and Andersen.
A first glimpse of the notebooks we provide is given here: http://nbviewer.jupyter.org/github/tschm/MosekRegression/tree/master/books/
For reproducing our results in exactly the environment we have used clone the underlying repository
git clone [email protected]:tschm/MosekRegression.git
and build the pymosek docker container with
docker-compose build pymosek
Docker will download all the required dependencies. Start the container with
docker-compose run -d -p 2016:9999 pymosek
If you have a valid Mosek license you may want to edit the Dockerfile and adjust the definition of the environment variable in there.
You can solve various (un)constrained regression and Markowitz problems.