Linked e-resources
Details
Table of Contents
Intro
Preface
Acknowledgements
Introduction
Contents
About the Authors
Acronyms
1 Theoretical Basis
1.1 Research Purposes
1.2 Future Directions
1.2.1 Trends in Research and Development
1.2.2 Long-Distance Dependencies
1.2.3 Location Information Awareness Capability
1.2.4 Dataset Noise Problem
1.3 Purpose and Significance of the NER
1.4 Current Status and Trends
1.4.1 Research Trends
1.4.2 Previous Research Work
References
2 Related Existed Models
2.1 Word Embedding
2.1.1 One-Hot Encoding
2.1.2 Word2Vec
2.1.3 Glove
2.1.4 CoVe
2.1.5 ELMo
2.2 Conditional Random Fields (CRF)
2.3 Deep Neural Networks
2.3.1 Long Short-Term Memory (LSTM)
2.3.2 Transformers
2.3.3 The Pretrained BERT Model
2.4 The Task Description
2.4.1 The Purpose of the Task
2.4.2 The Problems of Chinese Named Entity Recognition
2.4.3 The Characteristics of Chinese Medical NER
2.5 Evaluation Indexes
References
3 Medical Named Entity Recognition Models with the Attention Distraction Mechanism
3.1 General Framework
3.2 Research Targeted Problem
3.3 Improved Neural Network Models
3.3.1 Extended Input Units
3.3.2 Bi-SC-LSTM
3.3.3 Attention Distraction Mechanisms
3.3.4 Decoding Layer
3.4 Experiments and Simulation Results
3.4.1 Datasets
3.4.2 Experiments
3.4.3 Introduction to Comparison Models
3.4.4 Experimental Results
3.4.5 Experiment Analysis
3.5 Conclusions
References
4 Transformer Entity Automatic Extraction Models in Multi-layer Soft Location Matching Format
4.1 General Framework
4.2 Research Targeted Problem
4.3 Multilayer Soft Position Matching Format Transformer
4.3.1 WordPiece Word Segmentation
4.3.2 BERT
4.4 The Lattice Transformer
4.4.1 The General
4.4.2 The Word Lattice Structure Transformer
4.5 Multi-layer Soft Position Matching
4.6 Fuzzy CRF
4.7 Experimental Setup and Simulation Results
4.7.1 Dataset
4.7.2 Pretrained Embedding and Modelling
4.7.3 Experimental Results
4.7.4 Experiment Analysis
4.8 Conclusion
References
5 Medical Named Entity Recognition Modelling Based on Remote Monitoring and Denoising
5.1 A General Framework
5.2 Research Targeted Problems
5.3 Methods
5.3.1 Positive Sample and Unlabeled Learning Based on Category Risk Prediction
5.3.2 Negative Sampling Based on Positive/Negative Entity Probabilities
5.3.3 Encoding Modelling
5.4 Experimental Setup and Simulation Results
5.4.1 Datasets
5.4.2 Experimental Setup
5.4.3 Simulation/Experiments Results
5.5 Conclusions
References
Outlook
Appendix
Preface
Acknowledgements
Introduction
Contents
About the Authors
Acronyms
1 Theoretical Basis
1.1 Research Purposes
1.2 Future Directions
1.2.1 Trends in Research and Development
1.2.2 Long-Distance Dependencies
1.2.3 Location Information Awareness Capability
1.2.4 Dataset Noise Problem
1.3 Purpose and Significance of the NER
1.4 Current Status and Trends
1.4.1 Research Trends
1.4.2 Previous Research Work
References
2 Related Existed Models
2.1 Word Embedding
2.1.1 One-Hot Encoding
2.1.2 Word2Vec
2.1.3 Glove
2.1.4 CoVe
2.1.5 ELMo
2.2 Conditional Random Fields (CRF)
2.3 Deep Neural Networks
2.3.1 Long Short-Term Memory (LSTM)
2.3.2 Transformers
2.3.3 The Pretrained BERT Model
2.4 The Task Description
2.4.1 The Purpose of the Task
2.4.2 The Problems of Chinese Named Entity Recognition
2.4.3 The Characteristics of Chinese Medical NER
2.5 Evaluation Indexes
References
3 Medical Named Entity Recognition Models with the Attention Distraction Mechanism
3.1 General Framework
3.2 Research Targeted Problem
3.3 Improved Neural Network Models
3.3.1 Extended Input Units
3.3.2 Bi-SC-LSTM
3.3.3 Attention Distraction Mechanisms
3.3.4 Decoding Layer
3.4 Experiments and Simulation Results
3.4.1 Datasets
3.4.2 Experiments
3.4.3 Introduction to Comparison Models
3.4.4 Experimental Results
3.4.5 Experiment Analysis
3.5 Conclusions
References
4 Transformer Entity Automatic Extraction Models in Multi-layer Soft Location Matching Format
4.1 General Framework
4.2 Research Targeted Problem
4.3 Multilayer Soft Position Matching Format Transformer
4.3.1 WordPiece Word Segmentation
4.3.2 BERT
4.4 The Lattice Transformer
4.4.1 The General
4.4.2 The Word Lattice Structure Transformer
4.5 Multi-layer Soft Position Matching
4.6 Fuzzy CRF
4.7 Experimental Setup and Simulation Results
4.7.1 Dataset
4.7.2 Pretrained Embedding and Modelling
4.7.3 Experimental Results
4.7.4 Experiment Analysis
4.8 Conclusion
References
5 Medical Named Entity Recognition Modelling Based on Remote Monitoring and Denoising
5.1 A General Framework
5.2 Research Targeted Problems
5.3 Methods
5.3.1 Positive Sample and Unlabeled Learning Based on Category Risk Prediction
5.3.2 Negative Sampling Based on Positive/Negative Entity Probabilities
5.3.3 Encoding Modelling
5.4 Experimental Setup and Simulation Results
5.4.1 Datasets
5.4.2 Experimental Setup
5.4.3 Simulation/Experiments Results
5.5 Conclusions
References
Outlook
Appendix