Hi Federico, Scikit-Learn does not handle missing values. Imputing missing values is quite a hard topic in itself and I don't think estimators should also handle them. Instead, I would rather advise you to try fill in these missing values with simple strategies (e.g., for each column, replace NaN values with the mean (known) value for that column) and then to train any of our estimators on the completed data. It might be good enough depending on your application.
Best, Gilles On 29 May 2013 18:03, federico vaggi <vaggi.feder...@gmail.com> wrote: > Hi everyone, > > I'm tackling a classification problem with a lot of missing values for > several features. A quick google search suggested that GradientBoosting and > RandomForest would both be able to handle NaN values - but even in the > bleeding edge repo, both classifiers complain and throw an error. > > Is this just a case of Google deceiving me? And does anyone have any advice > for systematically dealing with missing values? I have very few 'complete' > rows, so I'd have to throw away a lot of data to get a clean dataset with no > NaN. > > Federico > > ------------------------------------------------------------------------------ > Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET > Get 100% visibility into your production application - at no cost. > Code-level diagnostics for performance bottlenecks with <2% overhead > Download for free and get started troubleshooting in minutes. > http://p.sf.net/sfu/appdyn_d2d_ap1 > _______________________________________________ > Scikit-learn-general mailing list > Scikit-learn-general@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general > ------------------------------------------------------------------------------ Introducing AppDynamics Lite, a free troubleshooting tool for Java/.NET Get 100% visibility into your production application - at no cost. Code-level diagnostics for performance bottlenecks with <2% overhead Download for free and get started troubleshooting in minutes. http://p.sf.net/sfu/appdyn_d2d_ap1 _______________________________________________ Scikit-learn-general mailing list Scikit-learn-general@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
