Natural language processing and machine learning to identify alcohol misuse from the electronic health record in trauma patients: Development and internal validation

Alcohol misuse is present in over a quarter of trauma patients. Information in the clinical notes of the electronic health record of trauma patients may be used for phenotyping tasks with natural language processing (NLP) and supervised machine learning. The objective of this study is to train and validate an NLP classifier for identifying patients with alcohol misuse.
Nội dung trích xuất từ tài liệu:
Natural language processing and machine learning to identify alcohol misuse from the electronic health record in trauma patients: Development and internal validation