Research On Access Control Model For Medical Big Data Based On Trust And Risk Adaptation

With the rapid development of big data collection,storage and other technologies,big data era has come quietly.At present,the country and society have realized the potential huge value of big data,big data and big data related technologies have been widely used in health care,e-commerce,logistics transportation and other fields,such as health care,e-commerce,logistics transportation and so on.In China,the study of health care big data has been formally incorporated into the national big data strategic layout.However,privacy leaks of medical data emerge one after another.How to effectively protect medical big data has become a hot issue of social concern.As one of the means of privacy protection,access control technology will play an important role in the privacy protection of medical big data.This paper systematically studies the privacy protection of medical big data from the perspective of role-based access control.Through the study of the relevant literatures at home and abroad,it is found that the role-based access control model only verifies the authenticity of the identity when assigning the role to the user.But ignore the subsequent access behavior may pose a risk.In another case,the role is assigned to the user in advance according to the identity,and the behavior of the user is monitored and controlled only in the process of the actual use of permissions,while the risks that may occur when assigning the role are ignored.In view of the above two situations,combined with the characteristics of medical big data,this paper develops and innovates the role-based access control model.On the one hand,trust control is added to the role assignment.On the other hand,it carries on the risk control to the access behavior.Both trust control and risk control monitor the whole process of user access.In trust control,we use the weighted sum of direct trust,recommendation trust and historical trust to obtain the comprehensive trust value.This paper analyzes the relationship between trust and risk,finds out the influence of risk on trust,and puts forward the influence of historical visit risk on historical trust.Finally,the effectiveness is verified by experiments.In risk control,this paper uses the user trust value related to the access behavior risk,selects the objective entropy and the access behavior entropy as the input of the BP neural network to predict the risk value of the access behavior.The risk adaptive access control is realized by the method of reducing quota.Finally,the effectiveness of the trust impact factor is verified by simulation experiments.After comparing the training effects of three kinds of neural network training functions,finally,we select the L-M method to train the samples,and select 20 representative test samples to test the performance of the trained neural network,and finally choose the Lam method to train the samples,and select 20 representative test samples to test the performance of trained neural networks.The results show that the access control model is feasible and accurate.In summary,on the basis of the related research on access control,this paper improves the role-based access control model,and also improves the trust value calculation method and the risk value calculation method.Finally,the availability and accuracy of the model are verified by simulation experiments.