Privacy Protection Based Retrieval on WBAN Big Data

Information privacy for Wireless body area network (WBAN) includes the user's physiological parameters. When a user of WBAN system intends to query medical data, the security issue is involved. Ciphertext retrieval technology is an effective way to ensure this confidentiality. However, efficiency and accuracy for data retrieval are challenges under the consideration of the aggregation of medical data is so large that it has to be stored in a cloud server. This paper proposes the similarity search tree structure to enhance the hit rate of multi-keyword ranking search. We also propose dynamic interval clustering algorithm DIK-MEDOIDS under the environment of cloud storage, while documents clustering initialization, the differences between the maximum and minimum document vector are divided into k slots, and the size of each slot equals to hypersphere diameter. Then the document vector which is closest to the middle value of the range is set as the hypersphere center. The size of each slot depends on the number of total documents amount. Therefore the clustering process dynamically changes with the increase of amount of documents, and the initialization time complexity is o(l). This algorithm works well for Big Data environment of ciphertext retrieval scenarios. The experimental results prove that the time consumption is linear function the document amount being at a low level. It shows that DIK-MEDOIDS algorithm has larger ascension than traditional DK-MEDOIDS algorithm in initialization.