3 ms·
And how does it perform compared to the solution that won the Netflix prize, years ago?
by amelius 5y ago
And how does it perform compared to the solution that won the Netflix prize, years ago?
- kqr 5y agoNetflix gets away with mainly training on dense data from high-frequency users that binge on content. The techniques they use don't adapt very well to most other organisations.
- whiplash451 5y agoNetflix _designs_ movies/series so that the majority of people will like them. And so Netflix' recommendation has become largely irrelevant now that the company is pushing its own content agenda. Their recommendation algorithm is operating under massive override by marketing rules.