tslearn-team/tslearn

Support for silhouette_samples() functionalities like sklearn.

srimantacse opened this issue · 5 comments

srimantacse commented

Is there any support for silhouette_samples() functionalities as sklearn has.

rtavenar commented

Hi,

We don't have support for silhouette_samples at the moment.

In terms of implementation, given the current implementation of silhouette_score (https://github.com/tslearn-team/tslearn/blob/main/tslearn/clustering/utils.py#L66) that heavily relies on sklearn's implem, it should be pretty much the same for one willing to adapt silhouette_samples:

  1. compute cross-similarity metrics between points
  2. rely on sklearn's implementation using metric="precomputed"

Anyone is welcome to open a PR on the topic.

srimantacse commented

@rtavenar thanks for the help. Able to get the sample score as you mentioned the steps. Thanks again.

rtavenar commented

@srimantacse would you be willing to contribute your code to tslearn ? If so, we would be happy to help you along the process.

srimantacse commented

@rtavenar please share me the process, I will do it.
Request to give me some time considering my schedule.
Share me the process.

rtavenar commented

You can find info there: https://github.com/tslearn-team/tslearn/blob/main/CONTRIBUTING.md#more-details-on-pull-requests

No problem if it takes some time, of course, it's always great to welcome new contributors