Address Matcher

Cross reference by address in Latin America. This poses a major challenge as addresses in Latin America are very unstructured and descriptive. A typical address will have the name of a house or building, a street name, a neighborhood, and a city, with all kind of spelling errors, abbreviations, and even descriptions such as “at the end of xyz street??, or “between such and such a street?? Cross referencing by addresses would allow us to identify households and therefore provide very rich data. The goal is a program matches any database against a master file (currently our master file in Venezuela has 16 million names) by addresses even if the addresses in the database is just similar (with spelling errors, abbreviations, incomplete, etc.) to the one in the master file. Let me summarize it in layman’s terms the key steps that we believe need to be accomplished: ·The addresses need to be normalized to eliminate abbreviations, spelling errors, etc. ·The addresses need to be parsed following several rules so as to determine, for example, a name street vs. building name ·Using a Google like search process addresses with key words need to be matched against addresses with the same key words ·A match identifying each letter in addresses key words against the closest number of the same letters in a similar word in a corresponding address. ·Matching partial addresses (i.e; street name and city) against a file with all street names and neighborhoods in a given city, to fill in gaps in the address. ·All these elements and processes need to be combined to select the best possible match ## Deliverables 1) Complete and fully-functional working program(s) in executable form as well as complete source code of all work done. 2) Deliverables must be in ready-to-run condition, as follows: a) Must be installed by the Seller in ready-to-run condition in the Buyer's environment. 3) All deliverables will be considered "work made for hire" under U.S. Copyright law. Buyer will receive exclusive and complete copyrights to all work purchased. (No GPL, GNU, 3rd party components, etc. unless all copyright ramifications are explained AND AGREED TO by the buyer on the site per the coder's Seller Legal Agreement) 4) Program pass all testing as determined by the buyer (document to be provided) 5) Program will upload two files, the first one considered the "master file", the second one considered the "matching file", will compare both files and identify similar records by address. Similar records will be considered those corresponding to the same address even if written differently. 6) Results expected are 80% match or better. ## Platform Software platforms: -Windows 2003 Server (.NET) currently, migrating to Suse Linux 9.3 -Borland Delphi 2005 and Borland Jbuilder 2005 (Java-J2EE JDK 5.0) -Borland Interbase 7.5, testing with MS SQL Server 2000, MySQL 4.1 and PostgreSQL Hardware platforms: -Intel Based Server, Poweredge 2850 with Dual 3ghz XEON Processors, 4Gig RAM, 3 SCSI 320 10K RPM 73Gb Hard Drives on a Raid 5 Array.