11:47 < manish7294> Anyone here?

15:54 < manish7294> @rcurtin would you welcome some new clustering algorithms?

15:54 < rcurtin> manish7294: absolutely, what do you have in mind?

15:55 < manish7294> Will Affinity Propogation do?

15:58 < rcurtin> I'm not particularly familiar with it, but if you can provide a fast implementation of it that matches the other clustering APIs it should definitely be merged and included :)

15:58 < rcurtin> it shouldn't be too hard to make the implementation reasonably fast thanks to armadillo, and I can help with that during the review process if needed

15:59 < manish7294> Thanks, I would look more into it and will shortly start building it.

16:01 < manish7294> And I have opened a PR #1193 and will be adding some more tests into it.

16:02 < rcurtin> I saw that, can you comment when it is ready for review please?

16:02 < rcurtin> so that I don't review it before it is done :)

16:03 < manish7294> Definitely, I will notify you when done!

16:04 < manish7294> rcurtin: Is there any other clustering algorithm which you would like to get implemented?

16:20 < rcurtin> manish7294: I would have to think about it; one thing that would be nice would be to use the existing EMST code to produce an agglomerative min-linkage clustering

16:29 < manish7294> rcurtin: will check on that.

16:33 < rcurtin> I think there is an issue open for it somewhere, but I am not usre

16:34 < rcurtin> sure*

16:34 < rcurtin> in any case, you can do hierarchical single-linkage clustering using a minimum spanning tree results

16:34 < rcurtin> er, a minimum spanning tree result

16:35 < manish7294> okay sure, will than start working on it. So, should I open an empty PR stating above.

16:38 < zoq> Is there a problem with the masterblaster connection? 70% package loss.

16:38 < rcurtin> yeah I think that they must be doing something to the networking there right now

16:38 < zoq> okay

16:38 < rcurtin> I know they are shutting down the lab that masterblaster is hosted in, but they are not moving masterblaster or related systems for another few weeks

16:39 < rcurtin> if it keeps up I'll send an email and ask what is going on, but my guess is something is tripping over cables in a network closet

16:39 < rcurtin> something or someone

16:39 < zoq> that might be the issue :)

17:08 < dungeon_master> Greetings everyone!! I know basic object Oriented C++ and all the basics of Machine learning. Am I able enough to contribute to Mlpack?

17:10 < zoq> dungeon_mas: Hello there, yes definitely.

17:14 < dungeon_master> So how can I start? I've compiled mlpack from source!

17:15 < rcurtin> I sent a note to the folks in Springfield to see what is up, because the masterblaster connection is not doing any better...

17:15 < zoq> dungeon_mas: Have you seen: https://github.com/mlpack/mlpack/blob/master/CONTRIBUTING.md?

17:15 < zoq> rcurtin: Okay

17:15 < rcurtin> dungeon_master: you could take a look at http://www.mlpack.org/involved.html or http://www.mlpack.org/gsoc.html; or find an issue on Github that you might like to try and solve

17:16 < rcurtin> zoq: they say they aren't aware of any issues so I am now filing an issue on their tracking system...

17:17 < dungeon_master> Yup!! I've seen it. I was just scared to start. I was thinking will I be able to contribute or not since I'm still a noob

17:19 < dungeon_master> btw Thanks <zoq> I'll now try my best to contribute

17:20 < zoq> rcurtin: Oh okay - http://kurg.org/media/IMG_3069.png

17:21 < zoq> dungeon_mas: everybody has to start somewhere, sometimes it's a good idea to jump right into :)

17:21 < rcurtin> zoq: thanks, I'll include that in the report if you don't mind

17:21 < zoq> dungeon_mas: Let us know if you need any help or if you have any further question

17:21 < zoq> rcurtin: sure

17:23 < dungeon_master> zoq sure :)

17:24 < zoq> btw. nice nickname :)

17:25 < zoq> Reminds me of a game I liked: https://en.wikipedia.org/wiki/Dungeon_Keeper