Toponym Identification in Epidemiology Articles – A Deep Learning Approach

When analyzing the spread of viruses, epidemiologists often need to identify location infected hosts. This information can be found in public databases, such as GenBank [3], however, provided these databases are usually limited country or state level. More fine-grained localization requires phylogeographers manually read relevant scientific articles. In this work we propose an approach automate process place name identification from medical (epidemiology) The focus paper is a deep learning based model for toponym detection and experiment with use external linguistic features domain specific information. was evaluated using collection 105 epidemiology articles PubMed Central [33] by recent SemEval task 12 [28]. Our best achieves F1 score $$80.13\%$$ , significant improvement compared art $$69.84\%$$ . These results underline importance embedding well journals.