Name matching tool

Post date: Jul 22, 2008 1:30:05 PM

The company:patent matching utilized the attached Perl routine. Starting with files of names that had been cleaned using a series of Stata routines (a file of patent assignee names and a file of corporate names from Compustat), this uses a word frequency algorithm to identify exact matches and (scored) likely matches. The likely matches above a threshold score were then examined manually to identify matches.