hdbscan error: TypeError: 'numpy.float64' object cannot be interpreted as an integer

Question

I ran hdbscan function code both on Linux and google colab and got the same error TypeError: 'numpy.float64' object cannot be interpreted as an integer

error seems to happen when applying data to the 'fit_predict' function

code comes from hdbscan and umap documentation so I did not expect an error. Is it my system or the package that makes the error occur?

Answer 1

I think there was a glitch introduced by Sklearn 1.3.0.

The solution that worked for me was downgrading to versions scikit-learn=1.2.2 and hdbscan=0.8.29. Once you uninstall your current packages:

conda install -c conda-forge scikit-learn=1.2.2 conda install -c conda-forge hdbscan=0.8.29