Biomedical Named Entity Recognition with BiLSTM-EDA: A Deep Learning Approach

In the medical field, dealing with a large volume of transcription is a difficult task, and it may take some time to read every line. Named Entity Recognition (NER) technology demonstrates the ability to scan entire documents and identify people, groups, and places. Identification of medical named entities from an unstructured natural language text is one of the most crucial topics in the field of Natural Language Processing (NLP), such as a medicine, disease, or treatment, using the medical NER model. This research presents a named entity recognition for a recurrent neural network structure based on the RNN version known as LSTM (Bi-LSTM), which makes use of Exploratory Data Analysis (EDA) to better comprehend the dataset. The experiments show the result of harmonic mean of 93.6%F1 score, Precision 95%, Recall 91%, and Accuracy of 92.89%.