Using Maximal Embedded Syntactic Subtrees for Textual Entailment Recognition
Sophia Katrenko, Pieter Adriaans · 2006
In this paper we address the textual entail-ment task by using tree mining and match-ing technique. Our results show that accu-racy can be improved when using a combi-nation of lexical entailment with syntactic matching. The best result we received by combining two components is the follow-ing: 70 % of recall and 57,5 % of precision on the test set. 1