Hi Soren and all on octave-dev list, My sourceforge ID is tealev.
According to the non-lin conj gradient method: I am a bit confused with cg_min case since: http://www.nabble.com/Optimization-%3A-bfgs-and-cg_min-td9238614.html#a9238614 but the function still exists in the octave as /usr/local/share/octave/3.0.1/site/m/cg_min.m In fact I would need to replace that one which is not coming from a forge package. I guess it would be a nice thing to have way to search for a specific function in octave on web as we can do it on octave forge. Wouldn't it be possible to process the functions of octave and display them in octave-forge as well? I am confused about the state of octave and octave-forge. One more question: I wrote the header of my function as # CL = confidence_level( x, N, b, calculation_type ) # # Returns the confidence level of categorical data summed in bins which were sampled from N items and not as ## -*- texinfo -*- ## @deftypefn {Function File} {} cgs (@var{A}, @var{b}) What is the preferable style? Levente On Saturday 07 March 2009, you wrote: > Hi > > In the future, please remember to CC the list. That way you're not > depending on me to reply (I can be quite slow sometimes) > > lør, 07 03 2009 kl. 12:38 +0100, skrev Levente Torok: > > The first is the confidence level calculation. It should be put in the > > Statistical package I believe. > > Please find it enclosed. > > This looks nice. I'd be glad to add you as a developer. What is your > sourceforge ID? > > > On the other hand I am in the process of creating a machine learning > > package. > > In there I made an adaboost, and also adopted Exponential family PCA > > (http://www.cs.cmu.edu/afs/cs/project/jair/pub/volume23/roy05a-html/node4.html), > > and Fast ICA package ( > > http://www.cis.hut.fi/projects/ica/fastica/code/dlcode.shtml ) to version > > 3.0.x. > > I am in the testing process of the latter ones. > > Sounds nice. I've been thinking for a while if we should have a Machine > Learning (or Pattern Recognition) package or if we should extend the > statistics package. > > > Recently I made an extension to nonline conjugte gradient menthod of > > Etienne ( http://users.isr.ist.utl.pt/~etienne/octave/ ) > > to include Fletcher–Reeves and Hestenes-Stiefel methods too. My version > > uses analytic derivatives rather than numerical ones. > > However Etienne suggested me to use bfgsmin instead. Unfortunately as I see > > the latter one calles over too many number of times > > the evaluation of the function under optimization and dumps so many > > unstoppable message to cerr. > > Cool. It's great to see the optimisation tools getting some love. > > > I will send yuo code time to time if you prefer it so. > > You don't need to send all code to the list. Just send it when you'd > like opinions about the code. Ok. Thanks. Lev -- Blogger of http://fapuma.blogspot.com Pictures of my life at http://picasaweb.google.com/TorokLev In addition: http://www.youtube.com/watch?v=0CUDg3NPEXY&fmt=16 ------------------------------------------------------------------------------ Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA -OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise -Strategies to boost innovation and cut costs with open source participation -Receive a $600 discount off the registration fee with the source code: SFAD http://p.sf.net/sfu/XcvMzF8H _______________________________________________ Octave-dev mailing list Octave-dev@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/octave-dev
