Effectiveness study of lexically mapping two thesauri
- Taboada, M. 1
- Lalín, R. 1
- Martínez, D. 1
- Tellado, S. 1
-
1
Universidade de Santiago de Compostela
info
- Borrajo Millán, Daniel (coord.)
- Castillo Vidal, Luis (coord.)
- Corchado Rodríguez, Juan Manuel (coord.)
Editorial: Universidad de Salamanca
ISBN: 978-84-611-8846-8, 978-84-611-8847-5
Ano de publicación: 2007
Volume: 1
Páxinas: 127-136
Congreso: Conferencia de la Asociación Española para la Inteligencia Artificial (12. 2007. Salamanca)
Tipo: Achega congreso
Resumo
Mapping thesauri is the task of identifying correspondences between entities in different thesauri. Discovering these matches is intrinsically problematic to automate. Earlier research has proposed solutions based on using lexical matching techniques and then, manually revising the resulting lexical mappings with the help of graphical user interfaces. Nevertheless, these solutions cannot guarantee the validity, accuracy and quality of the vocabulary mappings, as human capacity is limited. In this paper, we propose a method to automatically evaluate the quality of the results of a lexical technique. Our method combines structural constraints and annotations with part-of-speech tags to identifying error patterns from the results of lexical matches, differentiating between those leading to fall in precision and those producing decrease in recall.