On 02/13/2011 02:40 PM, Nir Krakauer wrote:
> I found that the regress function runs out of memory when given
> moderately large problems, such as
>
> y = randn(1E5, 1);
> X = randn(1E5, 2);
> [b, bint, r, rint, stats] = regress(y, X);
I applied your changes. Thanks.
Arno
diff --suppress-common-lines -d --speed-large-files
--suppress-blank-empty
/usr/share/octave/packages/3.2/statistics-1.0.10/regress.m regress.m
1a2
> ## Nir Krakauer, 2011: Modified to avoid computing X * pinv(X), which will
> lead to running out of memory when the number of data points is aroun
søn, 13 02 2011 kl. 14:40 -0500, skrev Nir Krakauer:
> I found that the regress function runs out of memory when given
> moderately large problems, such as
>
> y = randn(1E5, 1);
> X = randn(1E5, 2);
> [b, bint, r, rint, stats] = regress(y, X);
>
> This is because a large (1E5*1E5) intermediate m
I found that the regress function runs out of memory when given
moderately large problems, such as
y = randn(1E5, 1);
X = randn(1E5, 2);
[b, bint, r, rint, stats] = regress(y, X);
This is because a large (1E5*1E5) intermediate matrix is computed.
Slight modification of the code, as given below, a
