Research On Privacy Protection Algorithm For Electronic Medical Record Data Publishing

With the progress of Internet technology,electronic medical records are gradually replacing the traditional medical records because of their advantages of portability and ease of storage,and digital medical data are also showing exponential growth.In the application of electronic medical records,users can easily access the electronic medical record information system to query related medical information;Medical institutions exchange medical data to improve the level of medical treatment through the electronic medical information system;In order to promote drug development and prevent the disease,scientific research institutions collect medical data to analyze and dig.However,due to the large number of patient data in electronic medical records,the voice of patients for personal privacy security is getting higher and higher.Therefore,it is imperative to restrict the user to access to electronic medical record systems and to protect sensitive information on electronic medical records.This paper analyzes the problem of privacy disclosure in the publication of electronic medical records under the application of electronic medical records,and studies the privacy protection technology of electronic medical records.The main contents of this paper are:(1)In this paper,we propose a scheme to construct the patient attribute classification tree in the process of anonymous medical records.It is possible to classify the similar electronic medical records and replace the original values with the lowest generalization value by constructing the attribute classification tree so that it can protect the privacy of the patients and reduce the loss of information caused by the generalization of attribute values.(2)In order to solve the problem of privacy leakage caused by K anonymous model and l diversity model,the paper presents a data privacy protection algorithm based on KD tree for electronic medical records.The algorithm uses the properties of the KD tree and the attribute classification tree to recalculate the maximum equivalence class until it can not be decomposed to enhance the validity of the electronic medical record.The equivalence class of the electronic medical records is used as an anonymous constraint to improve the privacy protection of electronic medical records after the decomposition.The results show that the risk of privacy disclosure of electronic medical records is reduced,while the overall quality of the electronic medical records is improved.(3)This paper analyzes the advantages and disadvantages of the K anonymous model and the differential privacy model applied to the electronic medical record data distribution,and proposes an algorithm based on K anonymity for the differential privacy electronic medical records.According to the actual electronic medical records,the algorithm obtains the maximum value of attribute equivalence class,and sorts the electronic medical records in the equivalence class to obtain the minimum value,the median value and the maximum value.Then the algorithm breaks it into two equivalence classes and ensures that the number of records has at least k in the equivalence classes to satisfy the anonymous request.The algorithm uses the centroid instead of the attribute values of the records in each equivalence class,and obtains an anonymous table of electronic medical records,and adds noise to each data in the anonymous table,and makes it possible to achieve differential privacy requirements.Thereby the algorithm enhances the ability of electronic medical records to withstand attacks.The results show that the algorithm not only improves the universality of the electronic medical records,but also improves the overall efficiency of the algorithm.