Named Entity Recognition on Medical text by using Deep Nueral Networks

Technological advancements have caused widespread changes in the medical industry. The development of new tools has made it possible to gain useful information from massive amounts of unstructured data. The medical literature publications issued by researchers include a vast amount of information. There are a lot of groups working on deciphering the papers of literature to find the hidden information. Thanks to advancements in Natural Language Processing technology, identifying drug names, ailments, symptoms, routes of administration, species, and dosage forms inside a textual material is a breeze. In this work, we offer a novel hybrid-based method for extracting named entities from papers in the medical literature. A brand-new vocabulary was developed to annotate the entities in the medical records with information about the route of administration, dose forms, and symptoms. Spacy is a blank machine learning model that trains the annotated entities. The trained model outperforms the baseline model in terms of accuracy. The confusion matrix is determined by comparing the results of the hybrid model’s dictionary- and human-validation processes.