[SciPy-User] scipy.interpolate.rbf sensitive to input noise ?

josef.pktd@gmai... josef.pktd@gmai...
Mon Feb 22 11:27:42 CST 2010


On Mon, Feb 22, 2010 at 11:45 AM,  <josef.pktd@gmail.com> wrote:
> On Mon, Feb 22, 2010 at 11:14 AM, denis <denis-bz-gg@t-online.de> wrote:
>> On Feb 22, 3:08 pm, josef.p...@gmail.com wrote:
>>> On Mon, Feb 22, 2010 at 7:35 AM, denis <denis-bz...@t-online.de> wrote:
>>> > On Feb 19, 5:41 pm, josef.p...@gmail.com wrote:
>>> >> On Fri, Feb 19, 2010 at 11:26 AM, denis <denis-bz...@t-online.de> wrote:
>>>
>>> >> > Use "smooth" ? rbf.py just does
>>> >> >    self.A = self._function(r) - eye(self.N)*self.smooth
>>> >> > and you don't know A .
>>>
>>> > That's a line from scipy/interpolate/rbf.py: it solves
>>> >    (A - smooth*I)x = b  instead of
>>> >    Ax = b
>>> > Looks to me like a hack for A singular, plus the caller doesn't know A
>>> > anyway.
>>
>>> It's not a hack it's a requirement, ill-posed inverse problems need
>>
>> OK, I must be wrong; but (sorry, I'm ignorant) how can (A - smooth)
>> penalize ?
>> For gauss the eigenvalues are >= 0, many 0, so we're shifting them
>> negative ??
>> Or is it a simple sign error, A + smooth ?
>
> ouch, standard Ridge is A + smooth * identity_matrix to make it
> positive definite.
>
> I don't know why there is a minus. When I checked the eigenvalues, I
> found it strange that there were some large *negative* eigenvalues of
> A, but I didn't have time to figure this out.
> Generically, A - smooth*eye would still make it invertible, although
> not positive definite
>
> I haven't looked at the sign convention in rbf, but if you figure out
> what's going on, I'm very interested in an answer.
>
> I just briefly ran an example with a negative smooth (-0.5 versus
> 0.5), rbf with gauss seems better, but multiquadric seems worse. If
> smooth is small 1e-6, then there is not much difference.
>
> Even with negative smooth, all except for gauss still have negative
> eigenvalues. I have no idea, I only looked at the theory for gaussian
> process and don't know how the other ones differ.

My intuition for gaussan might not be correct for the other rbfs,
gauss is decreasing in distance, all others are increasing in distance

>>> for func in Rbfuncs: print func, Rbf( x, y, function=func ,smooth=smooth)._function(np.arange(5))

gaussian [ 1.      0.9272  0.739   0.5063  0.2982]
linear [0 1 2 3 4]
thin-plate [  0.       0.       2.7726   9.8875  22.1807]
multiquadric [ 1.      1.0371  1.1413  1.2964  1.4866]
quintic [   0    1   32  243 1024]
cubic [ 0  1  8 27 64]

and the distance matrix itself has negative eigenvalues

Josef

>
>
>>
>>> penalization, this is just Ridge or Tychonov with a kernel matrix. A
>>> is (nobs,nobs) and the number of features is always the same as the
>>> number of observations that are used. (I was looking at "Kernel Ridge
>>> Regression" and "Gaussian Process" before I realized that rbf is
>>> essentially the same, at least for 'gauss')
>>> I don't know anything about thinplate.
>>>
>>> I still don't understand what you mean with "the caller doesn't know
>>> A".  A is the internally calculated kernel matrix (if I remember
>>> correctly.)
>>
>> Yes that's right; how can the caller of Rbf() give a reasonable value
>> of "smooth"
>> to solve (A - smoothI) inside Rbf, without knowing A ?  A is wildly
>> different for gauss, linear ... too.
>> Or do you just shut your eyes and try 1e-6  ?
>
> That's the usual problem of bandwidth selection for non-parametric
> estimation, visual inspection, cross-validation, plug-in, ... I don't
> know what's recommended for rbf.
>
> Cheers,
>
> Josef
>
>>
>> Thanks Josef,
>> cheers
>>  -- denis
>> _______________________________________________
>> SciPy-User mailing list
>> SciPy-User@scipy.org
>> http://mail.scipy.org/mailman/listinfo/scipy-user
>>
>


More information about the SciPy-User mailing list