Evolution of AI enabled healthcare systems using textual data with a pretrained BERT deep learning model.

Area of Science:

• Medical Informatics
• Artificial Intelligence
• Health Services Research

Background:

• Artificial Intelligence (AI) is transforming healthcare, improving diagnostics and services.
• Traditional trend analysis relies on expert judgment, lacking efficiency and accuracy.
• Identifying emerging AI trends is crucial for future healthcare development.

Purpose of the Study:

• To uncover emerging trends and future directions of AI in healthcare.
• To apply advanced text mining and deep learning for trend analysis.
• To provide an empirical basis for AI commercialization in healthcare.

Main Methods:

• Utilized text mining on 1587 scientific papers (Web of Science) and 1314 patents (Derwent Innovations Index) from 2018-2022.
• Employed multiple deep learning algorithms, including a novel self-supervised approach leveraging BERT.
• Analyzed market trends, research hotspots, and technological advancements.

Main Results:

• Identified key market trends: Internet of Things (IoT), data security, and image processing in AI healthcare.
• Revealed current research hotspots and technological trajectories.
• Provided empirical data on AI commercialization potential in healthcare.

Conclusions:

• The study introduces a novel deep learning self-supervised model for AI in healthcare trend analysis, enhancing accuracy and efficiency.
• Findings offer valuable guidance for researchers, policymakers, and industry professionals.
• Highlights potential transformation directions for future healthcare services through AI commercialization.