Ferramentas de recuperación de textos para bibliotecas dixitais: lematización

Abstract

The ability to search documents by content, i. e., to look for documents dealing with a certain subject, is one of the most interesting services offered by a Digital Library. In order to offer these services, digital libraries need resources and text retrieval tools (such as corpora, electronic dictionaries, stemmers, or morphological analyzers), which must be developed for the language in which the documents of the library are written. The quantity and quality of the developed resources and tools depend on the used language. English has always had a great advantage in this field. On the contrary, in the Iberian Peninsula, Digital Libraries devoted to texts written in Galician have difficulties to develop content search services, since there are not enough tools and resources to do these implementations yet. This paper shows a Text Retrieval tool for the Galician language, built through a collaboration between Galician–Portuguese Philology and Computer Science researchers from the University of A Corunna. This tool is a stemmer that was first introduced in 2002, and it has been optimized, completed and tested during last years. We have used several different corpora to perform the tests, in order to accurately incorporate content search services in Digital Libraries.