Re: bringing back the fp-growth code in mahout
What is the best way to tell if Apache code is being maintained, in particular the fp-growth algorithm in Spark's MLlib? My original intent (5 months ago) was to replace the map reduce portion of the fp-growth code with an alternate, though I wasn't sure what that alternate should be. My motivation for wanting frequent itemsets is that they are closed with respect to intersections, so they form simplicial complexes. I've written software for mining simplicial complexes for their geometry. Actually, for their 2-dimensional persistent homology. It means I can look at how the geometry changes as both the support and confidence parameters vary. I'm hoping to take at least some of the guesswork out of making the right choices for these parameters, which seems to be sort of an open question. So for now I'll see if Spark's implementation generates usable frequent item sets, and have some fun learning Scala, and see about maybe getting fp-growth running on top of Flink. On 04/27/2015 07:59 AM, Ted Dunning wrote: Ray, Is the Spark implementation usable? Is it maintained? If not, there is a decent reason to move forward. I don't think that we want to revive the old map-reduce implementation. On Mon, Apr 27, 2015 at 5:48 AM, ray mailto:rtmel...@gmail.com>> wrote: I had it in mind to volunteer to maintain the fp-growth code in Mahout, but I see that Spark has an fp-growth implementation. So now that I have the time to work on this, I'm wondering if there is any point, or if there is still any interest in the Mahout community. If not, so be it. If so, I volunteer. Regards, Ray.
bringing back the fp-growth code in mahout
I had it in mind to volunteer to maintain the fp-growth code in Mahout, but I see that Spark has an fp-growth implementation. So now that I have the time to work on this, I'm wondering if there is any point, or if there is still any interest in the Mahout community. If not, so be it. If so, I volunteer. Regards, Ray.
I would like to contribute to the Mahout library
Mahout Developers: I would like to make some kind of contribution to the Mahout library. To this end, here are a couple of ideas: 1) Sign up to maintain the fpgrowth code, with the thought of adding some alternative to the Hadoop MapReduce portion of the implementation. 2) Is there still interest in a deep autoencoder for time series? I do have a background in neural networks, and a very strong background in math. I lean towards working on an autoencoder thing. Then again, I'm open to suggestions. Regards, Ray Melton.
Scala code in Mahout
Just to verify: Is Mahout now accepting code comprised entirely of Scala? Ray
Re: 0xdata interested in contributing
On 03/12/2014 05:44 PM, Ted Dunning wrote: I have been working with a company named 0xdata to help them contribute ... *Required Additional Work* Sparse matrices Linear algebra bindings Class-file magic to allow off-the-cuff function definitions What kind of additional work is required in terms of sparse matrices? By that, I mean did you have a sparse matrix package that just didn't working out quite so well? Or do you have a sparse matrix package that could be a much better sparse matrix package? Ray.
Re: Fwd: Neural Network and Restricted Boltzman Machine in Mahout
Ted, It's very interesting that you put it this way, "a replication of this architecture". It's really compelling to pursue this. I'm working through one of the deep learning papers from the quoted link now to get an idea of what it would take to make it happen. For me right now though, it seems like a little much. I mean, as a mathematician I find the mathematics being used in Mahout refreshing. Some absolutely beautiful applications. Real life math on the hoof. On the other hand, as a relative newcomer to Java, I'm still coming up to speed. I'll focus for now on contributing to documentation, possibly some patches. See how the contribution process works, gain a little confidence there first. (I do have a background in neural networks.) Ray. On 03/14/2013 11:44 AM, Ted Dunning wrote: Yeah we have had little pull on these techniques beyond the simplest case of logistic regression. Would you guys be willing to sign up for maintaining the code that might result? The thing that might move the needle would be a replication of this architecture: http://deeplearning.net/2012/12/13/googles-large-scale-deep-learning-experiments/ On Thu, Mar 14, 2013 at 8:33 AM, Danny Busch wrote: I asked years ago whether someone would be interested in starting implementations of neural network algorithms in mahout, but apparently t here was no big interest / priority and feedback that time. Maybe this has changed, as Neural Networks yet again become more popular. ~ danny Ying Liao hat am 14. März 2013 um 16:24 geschrieben: Can anyone in dev/commits shed some lights on this? Thanks, Ying -- Forwarded message -- From: Ying Liao Date: Wed, Mar 13, 2013 at 6:04 PM Subject: Neural Network and Restricted Boltzman Machine in Mahout To: u...@mahout.apache.org Hi, I am interested using Neural Network type of tools to do classification in Mahout. I wonder if these methods are finished in Mahout 0.8. On Mahout website, it still says open. Thanks, Ying