Biomedical named entity recognition using TCN approaches and bio tagging

Abstract

<p>Biomedical named entity recognition (BNER) is to identify instances in biomedical field such as chemical compounds, drugs, genes, RNA, DNA and proteins used in extracting information. It extracts relation between various drugs and their usage, profiles of similar and related drugs with help of machine learning approach. The efficiency in biomedical field is still in research for further improvement even many supervised methods are applied. The proposed method combines two algorithms and improve performance based on features used. It uses conditional random field (CRF) for entity identification and classification of temporal conventional network (TCN) to detect and recognize subtypes in BNER. Datasets such as GENIA and CHEMDNER corpus are used for evaluation with different entity types. Results shows that proposed methods performed better compared to other machine learning approach. The detailed study of TCN has been discussed. The classification of BNER is mapped with various classification methods to enhance result of high recognition.</p>