Complex biological event extraction from full text using signatures of linguistic and semantic features

Authors:
Liam R. McGrath;Kelly Domico;Courtney D. Corley;Bobbie-Jo Webb-Robertson
Affiliations:
Pacific Northwest National Laboratory, Battelle BLVD, Richland, WA;Pacific Northwest National Laboratory, Battelle BLVD, Richland, WA;Pacific Northwest National Laboratory, Battelle BLVD, Richland, WA;Pacific Northwest National Laboratory, Battelle BLVD, Richland, WA
Venue:
BioNLP Shared Task '11 Proceedings of the BioNLP Shared Task 2011 Workshop
Year:
2011

Citing 11
Cited 1

The nature of statistical learning theory

The nature of statistical learning theory
Making large-scale support vector machine learning practical

Advances in kernel methods
Automatic labeling of semantic roles

Computational Linguistics
Accurate unlexicalized parsing

ACL '03 Proceedings of the 41st Annual Meeting on Association for Computational Linguistics - Volume 1
Overview of BioNLP'09 shared task on event extraction

BioNLP '09 Proceedings of the Workshop on Current Trends in Biomedical Natural Language Processing: Shared Task
Biomedical event annotation with CRFs and precision grammars

BioNLP '09 Proceedings of the Workshop on Current Trends in Biomedical Natural Language Processing: Shared Task
Complex event extraction at PubMed scale

Bioinformatics
Towards event extraction from full texts on infectious diseases

BioNLP '10 Proceedings of the 2010 Workshop on Biomedical Natural Language Processing
Overview of BioNLP Shared Task 2011

BioNLP Shared Task '11 Proceedings of the BioNLP Shared Task 2011 Workshop
Overview of the infectious diseases (ID) task of BioNLP Shared Task 2011

BioNLP Shared Task '11 Proceedings of the BioNLP Shared Task 2011 Workshop
BioNLP Shared Task 2011: supporting resources

BioNLP Shared Task '11 Proceedings of the BioNLP Shared Task 2011 Workshop

Overview of the infectious diseases (ID) task of BioNLP Shared Task 2011

BioNLP Shared Task '11 Proceedings of the BioNLP Shared Task 2011 Workshop

Quantified Score

Hi-index	0.00

Visualization

Abstract

Building on technical advances from the BioNLP 2009 Shared Task Challenge, the 2011 challenge sets forth to generalize techniques to other complex biological event extraction tasks. In this paper, we present the implementation and evaluation of a signature-based machine-learning technique to predict events from full texts of infectious disease documents. Specifically, our approach uses novel signatures composed of traditional linguistic features and semantic knowledge to predict event triggers and their candidate arguments. Using a leave-one out analysis, we report the contribution of linguistic and shallow semantic features in the trigger prediction and candidate argument extraction. Lastly, we examine evaluations and posit causes for errors in our complex biological event extraction.