Article: Improving named entity recognition accuracy for gene and protein in biomedical text literature Journal: International Journal of Data Mining and Bioinformatics (IJDMB) 2014 Vol.10 No.3 pp.239 - 268 Abstract: The task of recognising biomedical named entities in natural language documents called biomedical Named Entity Recognition (NER) is the focus of many researchers due to complex nature of such texts. This complexity includes the issues of character-level, word-level and word order variations. In this study, an approach for recognising gene and protein names that handles the above issues is proposed. Similar to the previous related works, our approach is based on the assumption that a named entity occurs within a noun group. The strength of our proposed approach lies on a Statistical Character-based Syntax Similarity (SCSS) algorithm which measures similarity between the extracted candidates and the well-known biomedical named entities from the GENIA V3.0 corpus. The proposed approach is evaluated and results are satisfied. For recognitions of both gene and protein names, we achieved 97.2% for precision (P), 95.2% for recall (R), and 96.1 for F-measure. While for protein names recognition we gained 98.1% for P, 97.5% for R and 97.7 for F-measure. Inderscience Publishers - linking academia, business and industry through research

Title: Improving named entity recognition accuracy for gene and protein in biomedical text literature

Authors: Hossein Tohidi; Hamidah Ibrahim; Masrah Azrifah Azmi Murad

Addresses: Department of Computer Science, Faculty of Computer Science and Information Technology, Universiti Putra Malaysia, 43400, Serdang, Selangor D.E., Malaysia ' Department of Computer Science, Faculty of Computer Science and Information Technology, Universiti Putra Malaysia, 43400, Serdang, Selangor D.E., Malaysia ' Department of Information System, Faculty of Computer Science and Information Technology, Universiti Putra Malaysia, 43400, Serdang, Selangor D.E., Malaysia

Abstract: The task of recognising biomedical named entities in natural language documents called biomedical Named Entity Recognition (NER) is the focus of many researchers due to complex nature of such texts. This complexity includes the issues of character-level, word-level and word order variations. In this study, an approach for recognising gene and protein names that handles the above issues is proposed. Similar to the previous related works, our approach is based on the assumption that a named entity occurs within a noun group. The strength of our proposed approach lies on a Statistical Character-based Syntax Similarity (SCSS) algorithm which measures similarity between the extracted candidates and the well-known biomedical named entities from the GENIA V3.0 corpus. The proposed approach is evaluated and results are satisfied. For recognitions of both gene and protein names, we achieved 97.2% for precision (P), 95.2% for recall (R), and 96.1 for F-measure. While for protein names recognition we gained 98.1% for P, 97.5% for R and 97.7 for F-measure.

Keywords: natural language processing; NLP; information extraction; NER; named entity recognition; biomedical literature; genes; proteins; biomedical texts; gene names; protein names; similarity measures; biomedical named entities; name recognition.

DOI: 10.1504/IJDMB.2014.064523

International Journal of Data Mining and Bioinformatics, 2014 Vol.10 No.3, pp.239 - 268

Received: 16 Jun 2011
Accepted: 02 May 2012
Published online: 21 Oct 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Improving named entity recognition accuracy for gene and protein in biomedical text literature

Keep up-to-date