Paper: High-Order Sequence Modeling for Language Learner Error Detection

ACL ID W11-1422
Title High-Order Sequence Modeling for Language Learner Error Detection
Venue Innovative Use of NLP for Building Educational Applications
Session
Year 2011
Authors

We address the problem of detecting Eng- lish language learner errors by using a dis- criminative high-order sequence model. Unlike most work in error-detection, this method is agnostic as to specific error types, thus potentially allowing for higher recall across different error types. The ap- proach integrates features from many sources into the error-detection model, ranging from language model-based fea- tures to linguistic analysis features. Evalua- tion results on a large annotated corpus of learner writing indicate the feasibility of our approach on a realistic, noisy and in- herently skewed set of data. High-order models consistently outperform low-order models in our experiments. Error analysis on the output shows that the calculation of precision on the test set rep...