9/1/2011

There are three important desiderata for similarity metrics:

-Boolean out (avoid 0, 1), we need one that accepts/produces in-between values too.

-The importance of a word is measured on how common the word is in the documents. This can be seen as in how come a cow or dog is not in a zoo.

-The similarity level should stay equal, whenever the same proportion of words is duplicated/copied/concatenated in the same document.

--
Ivan Zhou

Graduate Student

Graduate Professional Student Association (GPSA) Assembly Member

School of Computing, Informatics and Decision Systems Engineering

Ira A. Fulton School of Engineering

Arizona State University