No. We don't. We used to use Lanczos, but that has improved. On Mon, Mar 25, 2013 at 4:43 PM, Abhijith CHandraprabhu <[email protected] > wrote:
> Sorry, I actually meant svds(sparse SVD). I think in mahout they use > Lanczos also. > > > On Mon, Mar 25, 2013 at 4:25 PM, Ted Dunning <[email protected]> > wrote: > > > Yes. But SSVD != Lanczos. Lanczos is vector at at time sequential like > > you said. SSVD does all the vectors in one go. That one go requires a > few > > steps, but does not require O(k) iterations. > > > > On Mon, Mar 25, 2013 at 12:16 PM, Sean Owen <[email protected]> wrote: > > > > > OK, the 'k iterations' happen inline in one job? I thought the Lanczos > > > algorithm found the k eigenvalues/vectors one after the other. Yeah I > > > suppose that doesn't literally mean k map/reduce jobs. Yes the broader > > > idea was whether or not you might get something useful out of ALS > > > earlier. > > > > > > On Mon, Mar 25, 2013 at 11:06 AM, Ted Dunning <[email protected]> > > > wrote: > > > > SVD need not be iterative at all. The SSVD code uses roughly 5 > > > map-reduces > > > > to give you a high quality SVD approximation. There is the option to > > add > > > > 0, 1 or more extra iterations, but it is rare to need more than 1. > > > > > > > > ALS could well be of use after less work. This is especially try for > > > > incremental solutions. > > > > > > > > > -- > Best regards, > Abhijith >
