Paper: Explicit and Implicit Syntactic Features for Text Classification

ACL ID P13-2150
Title Explicit and Implicit Syntactic Features for Text Classification
Venue Annual Meeting of the Association of Computational Linguistics
Session Short Paper
Year 2013
Authors

Syntactic features are useful for many text classification tasks. Among these, tree kernels (Collins and Duffy, 2001) have been perhaps the most robust and effective syntactic tool, appealing for their empirical success, but also be- cause they do not require an answer to the difficult question of which tree features to use for a given task. We compare tree kernels to different ex- plicit sets of tree features on five diverse tasks, and find that explicit features of- ten perform as well as tree kernels on accuracy and always in orders of mag- nitude less time, and with smaller mod- els. Since explicit features are easy to generate and use (with publicly avail- able tools), we suggest they should al- ways be included as baseline compar- isons in tree kernel method evaluations.