Approximate String Matching for Geographic Names and Personal Names

Clodoveu A. Davis Jr.Emerson de Salles

The problem of matching strings allowing errors has recently gained importance, considering the increasing volume of online textual data. In geotechnologies, approximate string matching algorithms find many applications, such as gazetteers, address matching, and geographic information retrieval. This paper presents a novel method for approximate string matching, developed for the recognition of geographic and personal names. The method deals with abbreviations, name inversions, stopwords, and omission of parts. Three similarity measures and a method to match individual words considering accent marks and other multilingual aspects were developed. Test results show high precision-recall rates and good overall matching efficiency.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: