Abstract—Biomedical Named Entity Recognition (BioNER) is a critical task in extracting valuable information from a biomedical corpus. There are different medical terminologies available, therefore it is difficult for scientists and researchers to find the important ones. An efficient integrated deep learning method that combines the advantages of Convolution Neural Network (CNN), Bidirectional-Long Short-Term Memory (Bi-LSTM) with Conditional Random Field (CRF) for precise and effective named entity recognition in the biomedical field is proposed. The proposed integrated model leverages the strengths of the deep learning models to effectively capture contextual information, learn label dependencies, and improve the accuracy of entity recognition. The MIMIC III free-text Electronic Health Record (EHR) dataset is processed using the SpaCy pipeline as a base. Different Entities are trained and tested. The CNN component allows for multi-level feature extraction by capturing local patterns and compositional features, including character-level representations and word embeddings. This allows the model to extract relevant and important features from the input sequence. The Bi-LSTM component further enhances the model’s performance by modeling contextual dependencies in both forward and backward directions, enabling a comprehensive understanding of the input sequence. By considering long-range dependencies, the Bi-LSTM component captures intricate relationships between words and the accuracy is improved. To incorporate label dependencies, a CRF layer is used on top of the Bi-LSTM layer. The CRF layer models the global structure of named entity labels and encourages consistent predictions, leading to more coherent and accurate entity recognition. Different performance parameters were considered to be compared with Named Entity Recognition (NER) systems. Comprehensive experimental findings demonstrate the model’s improved performance.
 
Keywords—Named Entity Recognition (NER), Bidirectional-Long Short-Term Memory (Bi-LSTM), Conditional Random Field (CRF), Convolutional Neural Networks (CNN), attention model


Copyright © 2024 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.