Discovering privacy

Abstract

I like to say that I backed into data mining. I come from the database community, although with a dissertation in database support for hypertext (while at Princeton University) I have never really been a core database researcher. With years (Bachelor's and Master's) at M.I.T., I had also had more machine learning than most. This leads to my first real data-mining work (although it was not called that at the time): applying machine learning techniques to a problem in heterogeneous databases. Schema integration, or more specifically schema element matching, has all of the hallmarks we see in good knowledge discovery problems: