Hi Andreas, Thank you for your reply!
The NFL belongs in the k-NN family of algorithms for classification / regression. In general, for a query point, NFL works like k-NN but instead of using the k feature points to determine the results, it generalized any two feature points of the same class by a feature line and then computes the distance of the query point from its projection on feature lines. It is supposed to improve the results of NN in some cases and especially on patter recognition tasks. You can have a look here http://www.scholarpedia.org/article/Nearest_feature_line http://www.dsp.toronto.edu/juwei/Publication/IEEE_NN.pdf I recently learnt about it while studying for a feature extraction course. I haven't implemented / tested it yet and I am new to sci-kit learn but I thought it would be fun (from my side) to implement it and have the opportunity to contribute too. I know that this is not how things work for you. To add a feature to the library it must be something that is going to be useful for many people so that there is a reason for maintaining it. Having a look at the issue page on github I think I can help with some minor contributions like adding more related projects in documentation (I will see If I can do that asap). Besides that my idea is to implement something from scratch so that I will be able to get familiar with the project step by step and if I end up with something that you will want to add to the library I will be very happy to contribute. Thank you for your time. I will continue watching the issue page and maybe help with something. Best Regards, Prokopis On Tue, Jul 28, 2015 at 8:43 PM, Andreas Mueller <t3k...@gmail.com> wrote: > Hi Gryllos. > > Before contributing a new feature (which is usually a major undertaking) > it us usually a good idea to get started working on known issues, have a > look at the issue tracker. > > I'm not familiar with the feature line approach. Can you elaborate and > provide a reference? > Please see the FAQ for our policy on algorithms we like to include: > > > http://scikit-learn.org/dev/faq.html#can-i-add-this-new-algorithm-that-i-or-someone-else-just-published > > > http://scikit-learn.org/dev/faq.html#can-i-add-this-classical-algorithm-from-the-80s > > Cheers, > Andy > > > On 07/23/2015 06:54 PM, Prokopis Gryllos wrote: > > Hi everyone, > > I would like to contribute code to the project and I was thinking of > implementing > a nearest feature line approach to the nearest neighbor class. > As it is suggested in the instruction set about contributing I thought it > would be best > to ask you first before I start working on it. > > Thank you in advance, I am waiting for your reply! > > Best Regards, > Gryllos Prokopis > > > ------------------------------------------------------------------------------ > > > > _______________________________________________ > Scikit-learn-general mailing > listScikit-learn-general@lists.sourceforge.nethttps://lists.sourceforge.net/lists/listinfo/scikit-learn-general > > > > > ------------------------------------------------------------------------------ > > _______________________________________________ > Scikit-learn-general mailing list > Scikit-learn-general@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general > >
------------------------------------------------------------------------------
_______________________________________________ Scikit-learn-general mailing list Scikit-learn-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
