Yes... though this is almost identical to just retrieving recommendations from an external server in the first place!
If you download the user's neighborhood, including things those users like, you have effectively downloaded a list of all recommendable items, and info to rank them. This isn't enough info to recompute recommendations completely when new information comes in. Yes, you can use new clicks to figure how similarities within the neighborhood have changed, but this won't do much to the recommendations, but can't find new users that might be in the neighborhood without going back to the server. I do know of one company that loaded an entire, tiny recommender onto their mobile app. It was item-based and stored pre-computed item-item similarity across their "items", of which there were only 100 or so. That worked quite well. On Wed, Aug 17, 2011 at 4:35 AM, Lance Norskog <goks...@gmail.com> wrote: > Are there any recommender algorithms designed for micro-sharding the > data model? The use case would be a mobile app that stores only a data > model for the phone owner. > > It seems like a user-user recommender does not need data for all > users; nearby users plus some background noise should be enough to > achieve good quality recommendations. The entire algorithm could > create a global dataset, and then pull out a small amount for a given > user. > > -- > Lance Norskog > goks...@gmail.com >