Hi Renzo,

Thank you for the comment.

As I understand you always get the same results as plotting but different results as indexes, but it happens sometimes, and for the same dataset. if so, it is really weird because there is no randomness in the implementation.

Could you please follow theses steps to clarify the problem and share the results with me:

- Find indexes multiple times usiong the same dataset and compare the results (without genereating the plot)

- Regenerate plots and see is there any difference between re-generated plots.

In the mean time, I also recommend you to use `sckitlearn` implementation to validate the distance scores. https://docs.scipy.org/doc/scipy/reference/generated/scipy.spatial.distance.mahalanobis.html

Best

Sergen