On 7/30/2012 7:41 AM, Gael Varoquaux wrote: > Hi Jim, > > It is not possible for us to give a general advice: there is no universal > classifier working for all datasets (this is known as the "no free lunch > theorem). > > If you have a lot of training data, you can try gradient boosted trees, > or maybe random forests. If your training data is limited, I think that > I'd try support vector machines. If your features can be reasonnably > considered as independent, Naive Bayes could work well. > > You should really try the estimators on the data and see what they give. > > Hope this helps, > > Gael
Thanks, Gael. This gives me a starting point. We have the situation of very limited training data. Also, I have concerns the features can really be considered independent. We will try to better ascertain the validity of this assumption. -- jv > > On Fri, Jul 27, 2012 at 11:29:53AM -0600, Jim Vickroy wrote: >> Hi, >> I recently discovered scikit-learn and it looks very impressive! >> I have a project that may be able to make use of scikit-learn and help >> me dispense with allot of custom code. >> The task is to identify 8 categories of features on 1024x1024 Solar >> images captured in 6 channels (wavelengths). A new set of 6 images >> arrives every 2 minutes. >> The current implementation is a Bayesian algorithm (mostly Python with >> f2py-wrapped Fortran handling a few "hot" spots). >> Having browsed the site documentation, I'm wondering if there is a >> better (all Python, simpler, easier to train, faster) approach. I would >> appreciate your thoughts on this. >> By the way, I'm a complete novice in this area. >> Thanks for your time. >> -- jv >> ------------------------------------------------------------------------------ >> Live Security Virtual Conference >> Exclusive live event will cover all the ways today's security and >> threat landscape has changed and how IT managers can respond. Discussions >> will include endpoint security, mobile security and the latest in malware >> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ >> _______________________________________________ >> Scikit-learn-general mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Scikit-learn-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
