A cross-institutional evaluation on breast cancer phenotyping NLP algorithms on electronic health records

Keyphrases

• Breast Cancer 100%
• Electronic Health Records 100%
• Cross-institutional 100%
• BioNLP 100%
• Conditional Random Fields 100%
• Institutional Evaluation 100%
• NLP Algorithm 100%
• Cancer Phenotyping 100%
• NLP Model 100%
• University of Minnesota 66%
• Mayo Clinic 66%
• Bidirectional Long Short-term Memory (BiLSTM) 66%
• Breast Cancer Patients 33%
• Learning Strategies 33%
• Clinical Text 33%
• Cancer Phenotype 33%
• Machine Learning Models 33%
• Recognition Task 33%
• Excellent Performance 33%
• Strategy Model 33%
• Transformer-based 33%
• Model Performance 33%
• Local Data 33%
• Learnability 33%
• Permutation Test 33%
• Clinical Domain 33%
• Model Transfer 33%
• Breast Cancer Phenotype 33%
• Named Entity Recognition 33%
• Traditional Machine Learning 33%
• Phenotype Extraction 33%
• Transformer-based Language Models 33%
• High Similarity 33%
• Clinical Documents 33%
• Model Development Processes 33%
• NLP Tasks 33%

Computer Science

• Generalizability 100%
• Conditional Random Field 75%
• Bidirectional Long Short-Term Memory Network 50%
• Model Development 25%
• Performance Model 25%
• Language Modeling 25%
• Learning Ability 25%
• Named Entity Recognition 25%
• Cancer Patient 25%
• Development Process 25%
• Machine Learning 25%
• Learning System 25%