http://qs1969.pair.com?node_id=600463


in reply to decomposing binary matrices

I was initially going to suggest principal components analysis as a solution, since the example decomposition you presented appears to be a clustering problem. I would then have pointed you to PDL and PDL::PCA. However, after reading some of the other comments my poor old neuron suspects that there is something more afoot that may invalidate PCA for the general case.

Replies are listed 'Best First'.
Re^2: decomposing binary matrices
by hv (Prior) on Feb 17, 2007 at 14:09 UTC

    I'm afraid I can't understand very much of the description of principal components analysis; as far as I can tell though, it requires an interpretation of what the source data represents quite unconnected with what my data represents. The fact that no mention is made of the special case of binary matrices reinforces that view in my mind.

    Hugo