Paper: UdS at CoNLL 2013 Shared Task

ACL ID W13-3612
Title UdS at CoNLL 2013 Shared Task
Venue International Conference on Computational Natural Language Learning
Session shared task
Year 2013

This paper describes our submission for the CoNLL 2013 Shared Task, which aims to to improve the detection and correction of the five most common grammatical error types in English text written by non-native speakers. Our system concentrates only on two of them; it employs machine learning classifiers for the ArtOrDet-, and a fully deterministic rule based workflow for the SVA error type.