Hi Pavan. I meant robust to outliers. But I guess that is encoded in the merging strategy. I didn't know about the book. Is it any good / recent?
Cheers, Andy On 03/04/2013 09:20 PM, Pavan Mallapragada wrote: > Thanks Andreas. > > I will be implementing them for my own work anyway, and will add the > necessary stuff required to pass the quality standards and send it around for > approval. I am not sure how long it will take, but was asking this to align > myself with the requirements while working on my stuff. > > In any case, if you think it is useful (with any necessary revisions), you > can pull it in. No worries otherwise. > > When you said robust did you mean it in the robust-statistics sense (say, > noise tolerant version) or just a definitive algorithm? The algorithm itself > is pretty well defined (say the classical one from Hartigain's book, or SLINK > with Sibson's improvements) , with similarity metrics and merging strategies > passed as the input, along with the data. > > Best, > Pavan > > On Mar 4, 2013, at 2:00 PM, Andreas Mueller <amuel...@ais.uni-bonn.de> wrote: > >> Hi Pavan. >> There are no hierarchical algorithms beside WARD. >> It would indeed be great to have single-link and complete link. >> Is there any robust version of single-link btw? What description would >> you go by? >> >> As always, the disclaimer: Getting a new algorithm into scikit-learn is >> a bit more than >> writing down the algorithm. You need to provide tests and documentation >> and need >> to respect our coding guidelines. So that usually involves a bit of work. >> >> On the other hand, having this classical algorithms in sklearn would >> definitely be great! >> >> Cheers, >> Andy >> >> >> On 03/04/2013 08:46 PM, Pavan Mallapragada wrote: >>> Hi, >>> >>> I am trying to find the single link / complete link algorithms in >>> scikit-learn. I see Ward's is the only hierarchical clustering algorithm >>> implemented (from the documentation). >>> >>> I did find other extensions of scipy implementing these, e.g. hcluster >>> (http://code.google.com/p/scipy-cluster/). >>> >>> I am willing to contribute by adding the implementations of other >>> hierarchical clustering algorithms. I am emailing to check if I am missing >>> something before trying to add my implementations. >>> >>> Thanks, >>> Pavan >>> ------------------------------------------------------------------------------ >>> Everyone hates slow websites. So do we. >>> Make your web apps faster with AppDynamics >>> Download AppDynamics Lite for free today: >>> http://p.sf.net/sfu/appdyn_d2d_feb >>> _______________________________________________ >>> Scikit-learn-general mailing list >>> Scikit-learn-general@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general >> >> ------------------------------------------------------------------------------ >> Everyone hates slow websites. So do we. >> Make your web apps faster with AppDynamics >> Download AppDynamics Lite for free today: >> http://p.sf.net/sfu/appdyn_d2d_feb >> _______________________________________________ >> Scikit-learn-general mailing list >> Scikit-learn-general@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general > > ------------------------------------------------------------------------------ > Everyone hates slow websites. So do we. > Make your web apps faster with AppDynamics > Download AppDynamics Lite for free today: > http://p.sf.net/sfu/appdyn_d2d_feb > _______________________________________________ > Scikit-learn-general mailing list > Scikit-learn-general@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general ------------------------------------------------------------------------------ Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_feb _______________________________________________ Scikit-learn-general mailing list Scikit-learn-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
