학술논문

Mining electronic medical records for patient care patterns
Document Type
Conference
Source
2009 IEEE Symposium on Computational Intelligence and Data Mining Computational Intelligence and Data Mining, 2009. CIDM '09. IEEE Symposium on. :146-153 Mar, 2009
Subject
Computing and Processing
Diseases
Data mining
Medical control systems
System testing
Surveillance
Control systems
Social network services
Electronic equipment testing
Medical tests
Sensitivity and specificity
Language
Abstract
A novel approach for generating full Electronic Medical Records of synthetic victims is described. Special emphasis is put on the data mining steps that build patient care models and perform clustering of this highly dimensional data set. A methodology for cluster validation is proposed. Results for a large data set with Staphylococcus aureus and Methicillin-Resistant Staphylococcus aureus infections are presented.