Re: [R] computing time for solve() in Matrix package

2012-06-28 Thread Paul Rathouz
Thank you for your response, Oliver. I think that the Matrix package has its own solve() functions that take advantage of the structure. If you make a block diagonal matrix, say, with kronecker() and do not impose any special structure on it. then, calling solve() will take very much longer

Re: [R] computing time for solve() in Matrix package

2012-06-27 Thread Oliver Ruebenacker
Hello, My guess would be that solve() does not take advantage of the special structure of the matrix and that you may want a sparse matrix representation. Take care Oliver On Tue, Jun 26, 2012 at 1:56 PM, Paul Rathouz wrote: > Hi -- I am wondering why the time to solve (invert)