The Design And Implementation Of The K-Anonymity Algorithm Based On Multi-attribute Generalization

Many valuable information resources are hidden in the medical data. As a result, the rational mining and utilization of it will play a significant role in promoting the development of all walks of life. Thus, the release of medical information received widespread attention. However, a large number of patients’privacy information is included in the medical information, such as the physiological and psychological state of the patient. So, if it’s leaked by mistake, the risk which it leaves for its owner is inestimable. Therefore, the privacy problems in the releasing of medical information attracted the attention of experts and scholars..At present, there are three kinds of means which are always used to protect the privacy in the information distribution system, they are deleting or removing the attribute、data encryption and Data disrupt or data randomization. However, in the releasing of the medical information, not only the usable of the released data but also the protection of patients’ privacy is needed. As a result, the three methods above are not suitable for the releasing of the medical information. Therefore, we selected the k- anonymity model which is proposed by Samarati P et al. in 1980 to achieve the releasing of medical information. However, K-anonymous algorithm could be improved, especially in the selection of generalization attribute and some details of its realization. The lack of careful consideration may cause an overgeneralization for the data, which will decline the useable and the privacy protection of its processed data.Thus, on the basis of Datafly, we proposed the k-anonymity algorithm based on multi-attribute generalization which can further improve the accuracy of the processed data. For this algorithm, we make improvements in two aspects, one is the selection of generalization attribute, the other is the situation that the most value attribute may not only when we select the generalization attribute. These make a great improvement on the accuracy of the processed data.In this paper, we give a brief introduction of research questions and related basic theories first. Next, we point out the shortage of the Datafly algorithm. And on the basis of this, we proposed the k-anonymity algorithm based on multi-attribute generalization For this algorithm, we make improvements in two aspects, one is the selection of generalization attribute, the other is the situation that the most value attribute may not only when we select the generalization attribute. These make a great improvement on the accuracy of the processed data. What’s more, the improvements make the algorithm be more suitable for the releasing of the medical information. Finally, on the basis of the k-anonymity algorithm based on multi-attribute generalization, we designed and implemented the medical information distribution model based on anonymity.