February 20, 2007

Still a Few Kinks in the Algorithm

Why is it that Burge's "Content Preservation" comes in at 56% relevance on this JSTOR search?

(I assure you that I've read the paper and have a paper copy that I could probably find very quickly -- but I decided it would be nice to have a pdf on this computer.)

Posted by Matt Weiner at February 20, 2007 06:22 AM

For my non-JSTOR readers, the search was "burge content preservation."

Posted by: Matt Weiner at February 20, 2007 06:24 AM

That's probably a rhetorical question, but could it be that the search is case sensitive?

Posted by: Matt's mom at February 20, 2007 04:10 PM

No -- it's not just that the Burge paper only has 56% relevance, but that it's only in fourth or fifth place; another paper is at 100%. I suppose that could happen if relevance is graded on a curve, and the other paper mentions the words "Burge," "Content," and "preservation" more than the original paper, which I guess doesn't need to keep saying its name.

Posted by: Matt Weiner at February 20, 2007 08:36 PM

In my experience, JSTOR's search function doesn't work well at all.

Posted by: teofilo at February 22, 2007 08:29 PM