Recognizing Textual Entailment Using Lexical Similarity

Abstract

We describe our participation in the PASCAL-2005 Recognizing Textual En-tailment Challenge. Our method is based on calculating " directed " sentence simi-larity: checking the directed " semantic " word overlap between the text and the hy-pothesis. We use frequency-based term weighting in combination with two differ-ent lexical similarity measures. Our best run shows 0.55 accuracy on the test data, although the difference between our two runs is not significant. We found remark-ably different optimal threshold values for the development and test data. We argue that, in addition to accuracy, precision and recall are valuable measures to consider for textual entailment.