A Review on Electronic Health Record Text-Mining for Biomedical Name Entity Recognition in Healthcare Domain

Area of Science:

• Biomedical informatics
• Artificial Intelligence in Healthcare

Background:

• Biomedical-named entity recognition (bNER) is crucial for extracting meaningful information from electronic health records (EHR).
• Traditional rule-based bNER systems struggle with the complexity of biomedical text.
• Deep learning (DL) has revolutionized bNER, enabling automated pattern learning for improved accuracy and efficiency.

Purpose of the Study:

• To review deep learning techniques for bNER in the healthcare domain.
• To explore the application of AI and DL in clinical records for treatment prediction.
• To systematically categorize bNER tools based on input, context, and tag architecture.

Main Methods:

• Review of deep learning-based biomedical-named entity recognition systems.
• Application of manual coding and multi-task learning for dataset creation.
• Systematic categorization of bNER tools (encoder/decoder models).

Main Results:

• Deep learning significantly improves the robustness and efficiency of bNER systems compared to traditional methods.
• Categorization of bNER tools provides insights into input, context, and tag distributions.
• The study highlights the potential of bNER for treatment prediction in clinical records.

Conclusions:

• Deep learning-based bNER is a powerful tool for advancing biomedical informatics and healthcare.
• Challenges and future directions for bNER in the healthcare field are discussed.
• bNER facilitates knowledge discovery and treatment prediction from complex clinical data.