Steven's comments are correct. Weka has a larger collection of algorithms. Mahout is specialized around scalable algorithms and scalable implementations.
Both packages support supervised and unsupervised algorithms. Due to scalability concerns, Mahout does not have much in the way of agglomerative algorithms. The highlights of Mahout right now are: - very large scale SVD - very large scale clustering - scalable item-set detection - the beginnings of very strong supervised classifiers for large features sets x large training data - decent underlying math library - command line or API focus. This is better than GUI focus for production work. On Wed, Sep 15, 2010 at 12:16 PM, Steven Bourke <sbou...@gmail.com> wrote: > Hi Qf, > > It's worth considering that weka does make available an SDK so you can > extend or develop any algorithms. I've only used weka a little, I got the > impression it had more algorithms implemented. I could be wrong. > > On Wed, Sep 15, 2010 at 8:09 PM, First Qaxy <qa...@yahoo.ca> wrote: > > > Hi all, > > I've been asked and I'm trying to figure out what are the major > differences > > between Mahout and WEKA concerning classification, clustering and assoc > > rules(PFPGrowth). My understanding so far is that:- Mahout supports > > unsupervised algorithms; WEKA also supports supervised through its UI.- > > Mahout scales much better while WEKA is memory bound.- Mahout is > targeting > > developers directly; WEKA mainly data mining analysts.- WEKA supports > > automating detection of classification algorithms. does Mahout have > > something similar?- Anything important that I've missed? > > If anyone can provide any insight that would be great. Thanks. > > /qf > > > > >
