Modeling the Dative Alternation with Automatically Extracted Features

Huayan Zhong, Amanda Stent, Mary Swift

We show that generation of contextually appropriate syntactic variation can be improved using a model based on automatically extracted features. We adapt a model for predicting dative alternation from (Bresnan et al. 2005); this model incorporates lexical, syntactic, semantic and pragmatic features. We evaluate the effect of using different types of feature on this classification task and show that the most predictive features for text differ from those for dialog. Finally, we show that modeling this type of syntactic variation improves the performance of surface realization for both text and dialog.

Subjects: 13.3 Syntax; 13.1 Discourse

Submitted: Jun 8, 2006


This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.