[pymvpa] getting simple regression working

[Yaroslav Halchenko]

the major "problem" was that dv was actually 2D with degenerate 2nd
dimension... so my version of your code is

from mvpa.suite import *
import numpy as n
iv=n.random.normal(0,1,(5,1))
dv=(2.0*iv).squeeze()
mydata=Dataset(samples=iv,labels=dv)
cc=GPR(regression=True,kernel=KernelLinear())
cc.train(mydata)
print cc.predict(mydata.samples)-dv


cc=SVM(svm_impl='NU_SVR', kernel_type='rbf',
       regression=True)
cc.train(mydata)
print cc.predict(mydata.samples) - dv

Mention that regression=True is needed also for SVM since by default in 0.4.x
they all are treated as classifiers (that was changed in 0.5, where regressions
are regressions ;))

P.S. Also your snippet managed to hit yet another bug which will get
fixed and pushed into maint/0.4 to be included in 0.4.5 bugfix release ;)

On Tue, 09 Feb 2010, kimberg at mail.med.upenn.edu wrote:
>  Incidentally, is there an easy way to retrieve the model parameters?

unfortunately we do not have some simple uniform way (yet?) but for
kernel GPR, model is described by original samples (._train_fv) and
._alpha vector 

-- 
                                  .-.
=------------------------------   /v\  ----------------------------=
Keep in touch                    // \\     (yoh@|www.)onerussian.com
Yaroslav Halchenko              /(   )\               ICQ#: 60653192
                   Linux User    ^^-^^    [175555]