Information Extraction for Biomedical Literature Using Artificial Intelligence: A Comparative Study

Growing biomedical literature necessitates efficient and effective knowledge extraction methodologies. Automated information extraction (IE) techniques offer promising solutions. This article presents a comprehensive comparative study that evaluates the latest research publications focusing on extracting information such as drug interactions and diseases from Biomedical Literature. The study analyses the selected studies, compares their findings, and discusses IE techniques’ attributes. This investigation examines a comprehensive range of IE methods, including deep learning, machine learning, rule-based methods, and hybrid approaches. The performance of specific methods is evaluated using discerning metrics such as F1-score, precision, recall, and accuracy. Evidence suggests that deep learning methods achieved significant improvements in accuracy, while hybrid approaches demonstrated flexibility and robustness. Also, domainspecific models and pre-trained language models are emphasized to enhance contextual understanding. Despite progress, challenges persist, including the handling of complex sentences, data availability, and generalization. In the context of the rapidly evolving biomedical field, IE methods play an increasingly critical role in the development of medical knowledge and patient care. This study highlights the critical need for continued research to develop and refine IE techniques for their broader application.