jottings http://blog.lib.umn.edu/ranga013/notes/ en Copyright 2009 Tue, 28 Jul 2009 12:54:53 -0600 http://www.sixapart.com/movabletype/?v=4.25 http://blogs.law.harvard.edu/tech/rss Recommendation systems
Netflix has had an open challenge for development any recommendation systems that does 10% better than what they are doing currently. The reward to take was a million dollars. Recently that challenge was met by two teams, "The Ensemble" and "Belkor's Pragmatic Chaos".

The challenge is still open for others to try.

A group at the U is also working on its own movie recommendation system - do check it out!

]]> http://blog.lib.umn.edu/ranga013/notes/2009/07/recommendation_systems.html http://blog.lib.umn.edu/ranga013/notes/2009/07/recommendation_systems.html Tue, 28 Jul 2009 12:54:53 -0600 Literature Reviews Here are some tools to help with literature reviews.

If you are looking to find the most cited articles on a topic or in a particular area; if you have an authoritative but out-dated article and you want to find all the published works that referenced that article you can use ISI Knowledge: Web of Science (If you are logged into UofM library use this link) is a great resource (use 'cited reference search'). UofM library has a comprehensive list of indexes classified by subject category.

ISI also allows you to electronically search through the references in a particular article, and list other related articles by matching articles in reference. It also allows for saving references from a session as a bibliographic text file or importing the references into a online reference manager like RefWorks (UofM now has group subscription for RefWorks).

Google scholar is also good for citation search. Google also lists unpublished articles from online repositories. On the other hand Google is likely to be more up to date.

]]> http://blog.lib.umn.edu/ranga013/notes/2007/11/literature_reviews.html http://blog.lib.umn.edu/ranga013/notes/2007/11/literature_reviews.html Thu, 01 Nov 2007 12:00:02 -0600