{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/medical-concept-representation-learning-from","title":"Medical Concept Representation Learning from Electronic Health Records and its Application on Heart Failure Prediction","arxiv_id":"1602.03686","date":"2016-02-11","proceeding":null,"authors":["Edward Choi","Andy Schuetz","Walter F. Stewart","Jimeng Sun"],"abstract":"Objective: To transform heterogeneous clinical data from electronic health\nrecords into clinically meaningful constructed features using data driven\nmethod that rely, in part, on temporal relations among data. Materials and\nMethods: The clinically meaningful representations of medical concepts and\npatients are the key for health analytic applications. Most of existing\napproaches directly construct features mapped to raw data (e.g., ICD or CPT\ncodes), or utilize some ontology mapping such as SNOMED codes. However, none of\nthe existing approaches leverage EHR data directly for learning such concept\nrepresentation. We propose a new way to represent heterogeneous medical\nconcepts (e.g., diagnoses, medications and procedures) based on co-occurrence\npatterns in longitudinal electronic health records. The intuition behind the\nmethod is to map medical concepts that are co-occuring closely in time to\nsimilar concept vectors so that their distance will be small. We also derive a\nsimple method to construct patient vectors from the related medical concept\nvectors. Results: For qualitative evaluation, we study similar medical concepts\nacross diagnosis, medication and procedure. In quantitative evaluation, our\nproposed representation significantly improves the predictive modeling\nperformance for onset of heart failure (HF), where classification methods (e.g.\nlogistic regression, neural network, support vector machine and K-nearest\nneighbors) achieve up to 23% improvement in area under the ROC curve (AUC)\nusing this proposed representation. Conclusion: We proposed an effective method\nfor patient and medical concept representation learning. The resulting\nrepresentation can map relevant concepts together and also improves predictive\nmodeling performance.","url_abs":"http://arxiv.org/abs/1602.03686v2","url_pdf":"http://arxiv.org/pdf/1602.03686v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"medical-concept-representation-learning-from","repo_url":"https://github.com/mp2893/retain","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok","spdx":"BSD-3-Clause"}}],"tasks":[{"task_slug":"representation-learning","task_name":"Representation Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1602.03686","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}