dorianHe/wasserstein_distance

Full Implementation of this code

Rajatsharma07 opened this issue · 1 comments

Hi,
Firstly, this is not an issue. You mentioned in the code the number of clusters that you fixed while performing Clustering. May I know how you decided to choose these values? Do you have its full implementation?

Hi, sorry for this late reply.
Short answer: The number of clusters is an arbitrary value. It is full implmentation.

The code in the notebook is just a POC for my supervisor, showing the Wassersteindistance concept. The idea is based on this paper (reference 4). I looked through the paper again. Authors didn't mention how they chose the number of clusters either. They only mention that the number of clusters influences the algorithm efficiency.

I hope this could be helpful for you.