Formação​ ​de​ ​gentílicos​ ​a​ ​partir​ ​de​ ​topônimos:​ ​descrição linguística​ ​e​ ​aprendizado​ ​automático

• Roger A. de M. R. Antunes UFSCar
• Thiago​ A. S. ​Pardo USP
• Gladis​ M. B. ​Almeida UFSCar

Abstract

This paper aims to describe the rules required in the transformation of toponyms into demonyms in order to identify regularities. From these regularities, we developed an algorithm that automatically generates demonyms for toponyms of interest. As a theoretical basis, the concepts of Derivational Morphology are considered, and, concerning the methodology, we used data about cities and demonyms provided by the Brazilian Institute of Geography and Statistics (IBGE) website, for which we produced procedures to make this data tractable. A complementary process of automatic learning was also carried out. As a result, a good accuracy was obtained in the prediction of demonyms, revealing new and relevant rules and features for the task.