The computation of cooccurrence matrices is a more general task than for just recommendations and I of more interest to me than just recommendations. Cooccurrence counting is a key step in building a random index, for instance.
I will also second Jake's comment about how even large scale organizations could well use an efficient off-line/on-line recommendation system. Few of them have had the luxury of building the system they really wanted (certainly that was true at Veoh). By pooling resources, we can build something of use to all of us. On Sat, Dec 5, 2009 at 5:17 PM, Sean Owen <[email protected]> wrote: > I guess that's why I've been reluctant to engineer and complicate the > framework to fit in offline distributed recommendation -- because this > can become as complex as we like -- since I wonder at the 'market' for > it. But it seems inevitable that this must exist, even if just as a > nice clean simple reference implementation of the idea. Perhaps I > won't go overboard on designing something complex yet here at the > moment. > -- Ted Dunning, CTO DeepDyve
