This is my first Haskell package. It is not yet even remotely complete or positively correct. This package contains an implementation of a manifold learning algorithm discovered by Andres Brun et. al in their 2005 published academic work "Fast Manifold Learning Based on Riemannian Normal Coordinates". I still have to tweak the interpolation step properly and find faster, yet sufficiently accurate, algorithms/data structures for necessary speed boosts. This will be updated around the clock and this document will reflect the activity.
Update (Oct 29, 2015) : Added the PDF (brun2005fast.pdf) that outlines/informs this implementation.
Update (Nov 10, 2015) : I have made some major revisions to all the code in the source directory. I can say with confidence that I have optimized the code to the best of my ability. Just be mindful that the algorithm does have a certain limitation; to even get decent results, you have to feed the algorithm a very large input (~5,000 points) and you have to ensure that the graph is connected. This is unfortunately very costly.
This should catch everything Cairo related (guessing based on memory, didn't have time to test in Docker, please update if this is wrong)
sudo apt-get install libghc-gtk-dev
stack build