Community segmentation is going to become a growing need for large online communities where some parts of the community will not want to encounter other parts for various reasons, ether for content rating issues or because there is too much information from a certain part of the community that a member might not be interested in.
Are there automatic methods for achieving this: for clustering content to a community members behaviour and then dynamically filtering content/notifications to this grouping?
BJ‘s literature review discusses various nearest neighbour approaches to ‘recommendation systems’ (‘collaborative filtering’ is no longer trend but sounds more impressive ;). The group is well treaded, so perhaps what I’m asking is has any site actually used this sort of system to completely adapt itself.
What are the advances on the front end?