A Vector based implementation of the kmeans clustering algorithm
Pull request Compare This branch is 21 commits behind alpmestan:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
Math
tests
LICENSE
README.markdown
Setup.hs
kmeans-vector.cabal

README.markdown

k-means clustering algorithm in Haskell

kmeans-vector is a Haskell library to perform the k-means clustering algorithm. It is based on the very efficient 'vector' library.

Feel free to contribute, may it be features, performance improvements, etc.

Performances

kmeans-vector is much faster than the existing kmeans package. For example, the kmeans package performs k-means on 10000 2D points with k=5 in 21.099s while kmeans-vector does the same in 6.605s. For 50000 3D points, with k=5 still, kmeans-vector performs in 47.853s while I Ctrl+C'd the kmeans version after 6 minutes and a half.

Author

This library is written and maintained by Alp Mestanogullari, alpmestan@gmail.com