von Hippel: For a list of local scope, telephone directories tend to separate business from residential directories. A nicely hash-indexed dataset would support a very quick look-up to check for presence in a business directory.

The INDEX() and INDEXW() function in SAS will also search name strings for substrings that would distinguish many business names from personal names: 'Inc.', 'Ltd.', 'Co.', 'Company', 'Store', 'Shop', 'Cafe', 'Restaurant', 'Pharmacy', 'Clinic', and 'Warehouse' come to mind immediately. A lexicon with frequencies of a sample list of business names and another of personal names would help identify differences and limit the number of look-ups required.

At some stage program will likely have to throw out some names for review. A fuzzy measure of membership in the set of business names would help you make a decision on whether to review or not. Sig

-----Original Message----- From: owner-sas-l@listserv.uga.edu [mailto:owner-sas-l@listserv.uga.edu] On Behalf Of von-hippel.1@osu.edu Sent: Monday, February 27, 2006 2:50 PM To: sas-l@uga.edu Subject: Distinguishing business names from personal names

Has anyone developed code for distinguishing business names from personal names?