Rule and Tree Ensembles for Unrestricted Coreference Resolution
Cícero Nogueira dos Santos, Davi Israel Abtibol Carvalho · Conference on Computational Natural Language Learning · 2011
In this paper, we describe a machine learning system based on rule and tree ensembles for unrestricted coreference resolution. We use Entropy Guided Transformation Learning (ETL) and Decision Trees as the base learners, and, respectively, ETL Committee and Random Forest as ensemble algorithms. Our system is evaluated on the closed track of the CoNLL 2011 shared task: Modeling Unrestricted Coreference in OntoNotes. A preliminary version of our system achieves the 6th best score out of 21 competitors in the CoNLL 2011 shared task. Here, we depict the system architecture and our experimental results and findings.