Open tarang-jain opened 2 years ago
For computing the core distances, during training the (min_samples+1)-th neighbor is considered, but while building the PredictionData object, the (minsamples)-th neighbor is considered. The specific parts in the code that I am referring to are: https://github.com/scikit-learn-contrib/hdbscan/blob/379d523d4e6b059db30970c8f5a08f383d5f3a6f/hdbscan/hdbscan.py#L245
and
https://github.com/scikit-learn-contrib/hdbscan/blob/379d523d4e6b059db30970c8f5a08f383d5f3a6f/hdbscan/prediction.py#L103
Hey, I would be willing to work on this. Can this please be assigned to me?
For computing the core distances, during training the (min_samples+1)-th neighbor is considered, but while building the PredictionData object, the (minsamples)-th neighbor is considered. The specific parts in the code that I am referring to are: https://github.com/scikit-learn-contrib/hdbscan/blob/379d523d4e6b059db30970c8f5a08f383d5f3a6f/hdbscan/hdbscan.py#L245
and
https://github.com/scikit-learn-contrib/hdbscan/blob/379d523d4e6b059db30970c8f5a08f383d5f3a6f/hdbscan/prediction.py#L103