Serviços Personalizados
Journal
Artigo
Indicadores
Citado por SciELO
Acessos
Links relacionados
Similares em SciELO
Compartilhar
Computación y Sistemas
versão On-line ISSN 2007-9737versão impressa ISSN 1405-5546
Resumo
SIKDAR, Utpal Kumar; EKBAL, Asif e SAHA, Sriparna. Entity Extraction in Biochemical Text using Multiobjective Optimization. Comp. y Sist. [online]. 2014, vol.18, n.3, pp.591-602. ISSN 2007-9737. https://doi.org/10.13053/CyS-18-3-2034.
In this paper we propose a multiobjective modified differential evolution based feature selection and classifier ensemble approach for biochemical entity extraction. The algorithm performs in two layers. The first layer concerns with determining an appropriate set of features for the task within the framework of a supervised statistical classifier, namely, Conditional Random Field (CRF). This produces a set of solutions, a subset of which is used to construct an ensemble in the second layer. The proposed approach is evaluated for entity extraction in chemical texts, which involves identification of IUPAC and IUPAC-like names and classification of them into some predefined categories. Experiments that were carried out on a benchmark dataset show the recall, precision and F-measure values of 86.15%, 91.29% and 88.64%, respectively.
Palavras-chave : Multiobjective modified differential evolution (MODE); feature selection; ensemble learning; conditional random field (CRF); named entity (NE).