4. Claremont Colleges Library January 8, 2010 Item : recommend things based on the item itself Personalized : recommend things based on the individual's past behavior Social : recommend things based on the past behavior of similar users Text: The Art, Science and Business of Recommendation Engines / Alex Iskold / 2007 Image: “Social Network” SaltLux Recommendation Systems
10. Claremont Colleges Library January 8, 2010 Days Requests Matches %% Click throughs %% 252 Total 94,033 45,842 49% 11,967 26% mean 743 362 49% 95 26% median 562.5 237 42% 42.5 18% mode 233 62 26% n/a n/a @Claremont: bX Results
11. Claremont Colleges Library January 8, 2010 “ I noticed the recommendations on the journal window.” “ Very cool, I hadn’t thought to have recommendations there, but I found things I needed!” Cecilia Conrad Dean of Faculty & Professor of Economics Pomona College
Hinweis der Redaktion
Story of me in boston for SFX training – just a humorous anecdote about how researchers view what we do as libraries.
More about the boston story – basically after describing what SFX was, my buddy Jeff, a medical researcher, said, “Hmm, glad you guys figured *that* out finally.” Sort of the user view of what we do in libraries – not a big deal but important for us to “finally figure it out”. Pertains to ‘figuring out recommendations’
About recommender systems – very tertiary glance at some ways to build them.
The evolution of the recommender system (SFX-style)
What my friend Jeff would say about the development of bX