pyclustering
pyclustering copied to clipboard
correcting rock
@annoviko sorry to bother, it seems you are the only maintainer... I just realized there are also C++ files that need modification, unfortunately my C++ ability is quite low so I can't help with that. If you don't have time to correct the C++ functions maybe force the use of Python for the ROCK module ? I'm also planning to add support for boolean input (as someone else raised an issue about), but there again I can't code the C++ functions, only Python. edit: someone already raised an issue about categorical input.
Hello @JosephChataignon ,
Thank you for the pull request, I will take a look at the correction, run tests for python code (your PR has been checked by integration tests where C++ implementation has been used). If everything OK, I will try to implement C++ solution using your python changes.
I added the outlier removal method used by the authors of the original paper (remove clusters of size 1 once the number of clusters reaches 1/3 of the data size)