New Issue: BCS Informer
14 pages; PDF.

+ Articles include: The Falls and Rise of Collaborative Filtering
From the article:

CF encompasses a range of techniques that focus on various aspects of the user and their known preferences. The “classic” technique is to look at the user – item graph and identify similar users based on items that have been similarly ranked. Recommendations are then generated based on rankings from those similar users (for items that you don’t have). Another approach is to use a clustering algorithm to group users according to various interests and then recommend based on this identification of “like minded” peers. A third approach, popularised by Amazon, is to only look at the item level and build proximity lists based on items that have appeared together in shopping baskets in the past. Services noted include Last.FM, Live Plasma, and Library Thing.

Another service, a ResourceShelf favorite, and an early publicly available collaborative filtering tool (still going strong), MovieLens from the University of Minnesota, is not included.

+ “Searching for People in the Personal Workspace”

+ Book Review: “Intelligent Document Retrieval”
Reviewed by Andrew Neill

Source: BCS Information Retrieval Specialist Group (British Computer Society)

This entry was posted on Monday, November 6th, 2006 at 12:33 am and is filed under Info Management and Retrieval, Reviews. You can follow any responses to this entry through the RSS 2.0 feed. Responses are currently closed, but you can trackback from your own site.