TY - CHAP

T1 - ETL ensembles for chunking, NER and SRL

AU - dos Santos, Cícero N.

AU - Milidiú, Ruy L.

AU - Crestana, Carlos E.M.

AU - Fernandes, Eraldo R.

N1 - Conference code: 11

PY - 2010

Y1 - 2010

N2 - We present a new ensemble method that uses Entropy Guided Transformation Learning (ETL) as the base learner. The proposed approach, ETL Committee, combines the main ideas of Bagging and Random Subspaces. We also propose a strategy to include redundancy in transformation-based models. To evaluate the effectiveness of the ensemble method, we apply it to three Natural Language Processing tasks: Text Chunking, Named Entity Recognition and Semantic Role Labeling. Our experimental findings indicate that ETL Committee significantly outperforms single ETL models, achieving state-of-the-art competitive results. Some positive characteristics of the proposed ensemble strategy areworth to mention. First, it improves the ETL effectiveness without any additional human effort. Second, it is particularly useful when dealing with very complex tasks that use large feature sets. And finally, the resulting training and classification processes are very easy to parallelize.

AB - We present a new ensemble method that uses Entropy Guided Transformation Learning (ETL) as the base learner. The proposed approach, ETL Committee, combines the main ideas of Bagging and Random Subspaces. We also propose a strategy to include redundancy in transformation-based models. To evaluate the effectiveness of the ensemble method, we apply it to three Natural Language Processing tasks: Text Chunking, Named Entity Recognition and Semantic Role Labeling. Our experimental findings indicate that ETL Committee significantly outperforms single ETL models, achieving state-of-the-art competitive results. Some positive characteristics of the proposed ensemble strategy areworth to mention. First, it improves the ETL effectiveness without any additional human effort. Second, it is particularly useful when dealing with very complex tasks that use large feature sets. And finally, the resulting training and classification processes are very easy to parallelize.

KW - Ensemble methods

KW - Entropy guided transformation learning

KW - Named entity recognition

KW - Semantic role labeling

KW - Text chunking

KW - Informatics

KW - Business informatics

UR - http://www.scopus.com/inward/record.url?scp=78650569670&partnerID=8YFLogxK

UR - https://d-nb.info/1000252183

U2 - 10.1007/978-3-642-12116-6_9

DO - 10.1007/978-3-642-12116-6_9

M3 - Article in conference proceedings

AN - SCOPUS:78650569670

SN - 3-642-12115-2

SN - 978-3-642-12115-9

T3 - Lecture Notes in Computer Science

SP - 100

EP - 112

BT - Computational Linguistics and Intelligent Text Processing

A2 - Gelbukh, Alexander

PB - Springer Verlag

CY - Berlin

T2 - 11th International Conference on Computational Linguistics and Intelligent Text Processing - CICLing 2010

Y2 - 21 March 2010 through 27 March 2010

ER -