The Filter
How It Works
If you're sitting comfortably, let the lesson begin. Please note, there will be questions at the end, so pay attention!
The Filter's recommendation and discovery engine is powered by a complex algorithm, which uses Bayesian mathematics to deliver content to a consumer for which there is the highest probability that they will like (NB. Recommended items will not necessarily by default be similar to the seed(s), but there's a high chance you will like them). It is based upon a unique implementation of artificial intelligence technology that means it adapts and learns organically to changing usage and changing consumer information provided. At its heart is an evidence-based transaction engine - the more evidence (e.g. music bought or listened to, TV or movies watched), the more The Filter learns - which, unlike more traditional correlation engines, improves over time. In short, the more stuff you rate, share or recommend, the more personalised will be your recommendations.
How Can I Get Better Recommendations?
Remember, the more we know about your tastes, the better the recommendations, so rate as many artists, songs, movies or actors as you can, and start sharing things you like with friends.
This all adds to the evidence The Filter has for you, and will result in even more personalised filtering. Also, don't forget that you can always dive back to the Taste Profiler to give us your entertainment preferences at any time. Finally, you can download The Filter application, as this can monitor your own music collection and learn what your play habits are (see the section on the right).
The Filter Application
What's this?
We have developed an application that you can download for Windows or for Mac. To find out more detail about this please go to the tools page.
