There is a deep chasm between my awareness of The Netflix prize and my experience of never having very good Netflix recommendations.
-
-
Replying to @generativist
1) they didn't use the solution because it was too complicated 2) they learned that user-supplied ratings are not nearly as strong a predictor as user behavior 3) they learned customers trust recommendations with explanations or categories more than unexplained recommendations :)
3 replies 2 retweets 2 likes -
Replying to @JasonThalken @generativist
but yea, on top of all that, recommendations just about anywhere are bad. Those models are nothing to write home about. :) Spotify's "discover weekly" playlist is an exception, but they rely heavily on the result of human-curated playlists to guide their suggestions.
1 reply 0 retweets 1 like
Yep. Only one I like. And youtubes is very good if the loss function were salience, but it has...uhh...very pathological effects
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.