A machine learning approach for Biomedical Named Entity Recognition

[摘要] This paper aims to develop a Named Entity Recognition (NER) that creates new tags that facilitate fast query processing, information retrieval and data preprocessing of Biomedical Domain. We have used a machine learning approach that uses domain specific knowledge to train the data and label the entities with appropriate tags. Conditional Random Fields (CRF) is implemented and used to train the input domain specific data that yields good performance. Experimental and evaluation result shows that the learned model yields a Biomedical domain specific NER recall of 82%, precision of 85%, accuracy of 82% and Fmeasure of 83%. Thus learned CRF model builds a domain specific NER for Biomedical Domain and tags the domain keywords with appropriate tags.

[学科分类] 基础医学

[关键词] Named Entity Recognition;Statistical Model;Machine Learning;Conditional R;om Field Classifier;Bio medical Data analytics