USE OF EXTERNAL INFORMATION IN ZIP CODE RECOGNITION by Jonathan J. Hull and Sargur

The recognltIon of zip codes in a postal address can utilize many sources of external information. City and state names are obvious examples that can be used in conjunction with a city-state-zip directory to provide evidence about digits in a zip code. This paper describes an extension of this methodology that uses information about legal street names and sUffixes to constrain the digits in a zip code. The design of such a technique is presented and some preliminary experiments with the ZIP+4 database are discussed. A method for extracting useful information from an address is discussed that does not require complete recognition of all the characters in words. Rather, a feature description is computed and used to retrieve a set of zip codes from a dictionary that correspond to words with the same description. A statistical simulation explores the effect of several alternative feature sets. It is shown that even a relatively simple description of two words in the street line of an address can significantly reduce the number of zip codes that could appear on a piece of mail. 361 USPS Advanced Technology Conference, Washington, D.C., Oct. 21-23, 1986