Research On Machine Learning Algorithm For Electronic Health Records Analysis And Processing

With the development of information technology,computer technology is increasingly applied to the medical field.Under the current situation of rapid spread of electronic medical systems,a large amount of medical-related information is stored in the form of electronic health records.Electronic health records contain great medical value.Applying machine learning and other methods for analysis and processing would bring incalculable value,which will have far-reaching effects on future medical research and medical development.The paper studies the machine learning algorithm based on electronic health records analysis and processing,and carries out a series of tasks and experiments.The main work includes:(1)Classification algorithms based on structured electronic health records: This experiment is designed to study the application of multiple machine learning classification algorithms on structured electronic medical record datasets.(2)Embedding methods based on electronic health records: This experiment mainly explores the embedding methods of electronic health records,including sentence embedding and text embedding.The embedding methods are analyzed,tested and compared through a variety of machine learning models and network structures.(3)Named Entity Recognition based on electronic health records: This experiment mainly conducts related research on Named Entity Recognition tasks based on electronic health records,and also compares and analyzes the results from multiple model structures.(4)Character vector training based on electronic health records: This experiment explores a character vector training method based on electronic health records.The training method is based on multi-task learning.There are also quantitative analysis and research through experiments.Based on the electronic health records,the thesis studies the application of various machine learning methods in various tasks,as well as making some improvements and innovations on some of the models,and verifies,compares and analyzes through a series of experiments.