ANONYMIZING DEVICE AND METHOD USING INFORMATION THEORY APPROACH OF ELECTRONIC MEDICAL RECORDS

摘要

A de-identification apparatus, method and recording medium using an information theory of an electronic medical record are disclosed. The present invention includes: a storage part storing electronic medical record (EMR) data; an entropy calculation part calculating entropy on each of attributes of the stored data; an attribute combination part combining each of the attributes to generate attribute combinations; a joint entropy calculation part calculating joint entropy on the generated attribute combinations; a dependence calculation part calculating dependence of the calculated joint entropy; an attribute combination selecting part selecting the attribute combinations with information loss amount less than a preset amount among the attribute combinations of the joint entropy and the dependence; and a k-anonymity part de-identifying data using a k-anonymity method based on the selected attribute combinations.