2012/12/14 Olivier Grisel <[email protected]>:
> 2012/12/14 Shishir Pandey <[email protected]>:
>> Hi
>>
>> I am trying to trying to find the regression matrix \beta. In the following:
>>
>> Y = X \beta + \epsilon
>>
>> Y - m x n matrix. (sparse matrix)
>> X - m x k matrix. (sparse matrix)
>> \beta - k x n matrix
>>
>> I try to fit the SGDRegressor as fit(X,Y). It gives a value error:
>> ValueError: Buffer has wrong number of dimensions (expected 1, got 2)
>>
>> I see in the documentation Y can only be a vector of type [n_samples]. Is
>> there a way in which I can use SGDRegressor with Y being a matrix.
>
> Right now, the SGDRegressor linear model accepts a single output. You
> will have to slice the Y matrix into column with shape `(n_samples,)`
> and run `n_outputs` independent linear regression and then aggregate
> the `coef_` vectors to find you beta matrix.

Alternatively you can use sklearn.linear_model.Ridge that supports
multiple ouputs directly:

http://scikit-learn.sourceforge.net/stable/modules/generated/sklearn.linear_model.Ridge.html#sklearn.linear_model.Ridge.fit

(the website scikit-learn.org has a DNS issue, hence the direct link
to the sourceforge page).

--
Olivier
http://twitter.com/ogrisel - http://github.com/ogrisel

------------------------------------------------------------------------------
LogMeIn Rescue: Anywhere, Anytime Remote support for IT. Free Trial
Remotely access PCs and mobile devices and provide instant support
Improve your efficiency, and focus on delivering more value-add services
Discover what IT Professionals Know. Rescue delivers
http://p.sf.net/sfu/logmein_12329d2d
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Reply via email to