Possibility of setting weight coefficients in OPTICS algorithm.

[sergeirubinsky]

We are very interested in your efforts on development and enhancement of OPTICS program sklearn implementation. It would be very important in our job (github.com/alashkov83/hydrocluster) to have possibility of setting weights, as it has been made in DBSCAN. Please inform us about your further plans.
A.Laskov, S.Rubinsky

[jnothman]

This sounds useful. I wonder if, like in DBSCAN, the only thing that really needs to change is the determination of core points (or perhaps core distances too).

(I'm also chuffed that something I wanted for my own work at the time I proposed and implemented - sample_weight in DBSCAN - is being so appreciated!)

[thoo]

@sergeirubinsky If you are not working on it, I would like to give a try. Thanks!!!

[adrinjalali]

@thoo there are some major changes in the pipeline to OPTICS in progress. You probably should wait for those to get in first.

[thoo]

@adrinjalali Thanks. I wiil.

[jnothman]

But you are welcome to investigate the algorithm and what would be needed, and write notes here.

[sergeirubinsky]

@sergeirubinsky If you are not working on it, I would like to give a try. Thanks!!!

We ourselves work on bioinformatics and would be glad you to try the problem.

[qinhanmin2014]

A note to everyone here: Our OPTICS is experimental now. There're some known bugs and we're likely to change the API in the future.
But yes, this seems to be an useful feature.

[adrinjalali]

@thoo you can work on this one if you want now.

[trabbart]

There was no comments or PR's for few months so I have started working on this.

[MitalPattani]

@trabbart Are you still on this, I would like to start on this otherwise.

[trabbart]

@MitalPattani I have stoped working on this and I do not have enough time to finish it in the future.

[MitalPattani]

Okay, I am working on this.

[OmarManzoor]

@adrinjalali Can I try working on this?

[adrinjalali]

I remember some major challenges here, but it'd be nice to try and see what comes out @OmarManzoor

[OmarManzoor]

I remember some major challenges here, but it'd be nice to try and see what comes out @OmarManzoor

After looking at this as well as at the closed PR linked with this issue, I think one of the challenges might be to come up with an efficient implementation of the code.

[adrinjalali]

So in #25855 we decided it's a wontfix issue. I think it applies here as well. Closing as won't fix, happy to re-open if others think otherwise.