UNED at PASCAL RTE-2 challenge

Abstract

This paper reports the description of the developed system and the results obtained in the participation of the UNED in the Second Recognizing Textual Entailment (RTE) Challenge. New techniques and tools have been added: enriched queries to WordNet, detection of numeric expressions and their entailment, and Support Vector Machine classi cation (SVM) are the more relevant. The accuracy performed is slightly higher than the one from the previous edition system.