Recurrent neural networks with specialized word embeddings for health-domain named-entity recognition.
Background: Previous state-of-the-art systems on Drug Name Recognition (DNR) and Clinical Concept Extraction (CCE) have focused on a combination of text "feature engineering" and conventional machine learning algorithms such as conditional random fields and support vector machines. However, developi...
| Publicado en: | Journal of Biomedical Informatics Vol. 76; pp. 102 - 110 |
|---|---|
| Autores principales: | , , , , |
| Formato: | Journal Article |
| Publicado: |
Academic Press Inc.
Dec2017
|
| Acceso en línea: | Ver este registro en EBSCOhost |
| fields | @attributes: recordID: 1 pdfLink: plink: https://search.ebscohost.com/login.aspx?direct=true&db=ccm&AN=127986491&site=ehost-live header: @attributes: shortDbName: ccm uiTerm: 127986491 longDbName: CINAHL Complete uiTag: AN controlInfo: bkinfo: dissinfo: jinfo: jid: 15320464 OMB jtl: Journal of Biomedical Informatics issn: 15320464 maglogo: N pubinfo: dt: Dec2017 vid: 76 pid: 735 pub: Academic Press Inc. place: Burlington, Massachusetts artinfo: ui: 127986491 127986491 NLM29146561 10.1016/j.jbi.2017.11.007 NLM29146561 127986491 ppf: 102 ppct: 8 formats: tig: atl: Recurrent neural networks with specialized word embeddings for health-domain named-entity recognition. aug: au: Jauregi Unanue, Iñigo Zare Borzeshi, Ehsan Piccardi, Massimo Unanue, Iñigo Jauregi Borzeshi, Ehsan Zare affil: University of Technology Sydney (UTS), Australia sug: subj: Resource Databases Neural Networks (Computer) Algorithms Scales ab: Background: Previous state-of-the-art systems on Drug Name Recognition (DNR) and Clinical Concept Extraction (CCE) have focused on a combination of text "feature engineering" and conventional machine learning algorithms such as conditional random fields and support vector machines. However, developing good features is inherently heavily time-consuming. Conversely, more modern machine learning approaches such as recurrent neural networks (RNNs) have proved capable of automatically learning effective features from either random assignments or automated word "embeddings".Objectives: (i) To create a highly accurate DNR and CCE system that avoids conventional, time-consuming feature engineering. (ii) To create richer, more specialized word embeddings by using health domain datasets such as MIMIC-III. (iii) To evaluate our systems over three contemporary datasets.Methods: Two deep learning methods, namely the Bidirectional LSTM and the Bidirectional LSTM-CRF, are evaluated. A CRF model is set as the baseline to compare the deep learning systems to a traditional machine learning approach. The same features are used for all the models.Results: We have obtained the best results with the Bidirectional LSTM-CRF model, which has outperformed all previously proposed systems. The specialized embeddings have helped to cover unusual words in DrugBank and MedLine, but not in the i2b2/VA dataset.Conclusions: We present a state-of-the-art system for DNR and CCE. Automated word embeddings has allowed us to avoid costly feature engineering and achieve higher accuracy. Nevertheless, the embeddings need to be retrained over datasets that are adequate for the domain, in order to adequately cover the domain-specific vocabulary. pubtype: Academic Journal doctype: Journal Article ougenre: Article language: English refInfo: holdings: @attributes: islocal: N |
|---|