Hi
I've commited some early work on creating a class representing the
Kronecker product of two matrices to the 'linear-algebra' package. These
seem to appear more and more often in Machine Learning, but I guess they
also appear frequently in other fields. So, it would be nice to have
efficient ways of dealing with them.
The current code allows you to compute the inverse of a Kronecker
product, solve linear systems, compute eigenvalues and SVD's.
Specifically, the implements the following functions
columns.m full.m issquare.m numel.m sparse.m
uplus.m
ctranspose.m inv.m kronprod.m plus.m svd.m
det.m iscomplex.m minus.m rank.m times.m
display.m ismatrix.m mldivide.m rdivide.m trace.m
disp.m isreal.m mpower.m rows.m transpose.m
eig.m issparse.m mtimes.m size.m uminus.m
There are still some things that aren't working perfectly ('inv' and
'eig' doesn't work for Kronecker products of non-square matrices, 'eig'
doesn't provide the results in the right order, and some other stuff).
If you want to help out, there are a bunch of XXX's in the code, and I'm
also pretty sure much can be improved and that bugs exists. I'm
releasing this early. We also need to implement stuff like 'chol' and
'schur' as they are also pretty simple.
Anyway, let me know if this stuff interests you,
Søren
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Octave-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/octave-dev
