The Privacy Protection Study Against Incremental Datasets

With the development of network technology and the expanding of database applications field,government,enterprises and ivdividuals released their own statistics according to need on the Internet,while some research studied their published data.The data included a variety of information of the community and some data involved sensivive personal information.With the rapid development of information technology,the amount of global information grew as exponential.People extensively used data mining technology in order to obtain useful information.But with the continuous development of data mining,data privacy was brought no small trouble.The privacy information of personal,for example,medical data of hospital included sensitive information involved individuals suffered from a disease.The phenomenon of data privacy leaked is become more and more seriously so that people's lives suffered from trouble at a great degree.Currently,the queries of data privacy protection technology vave become a large number of theoretical results and system prototypes.The main job mostly focused on protecting the anonymity of the static database.The early anonymity protecting method is static,while the data in the database are always in change and are dynamic.How to handle the inference attack caused by the repeated release is the key point of study.The study of dynamic data includes a m-invariance,supporting insert and deleting the anonymous strategy.Maintaining the incremental k-anonymity only supports inserted anonymous trategy.The paper proposes a new method which sovles all the privacy caused by a number of published leaked against the lack of above research .First of all,the possibility of causing the privacy leakage has been analyzed.In the anonymous process,the new method not only ensures the k-anonymous of anonymity data but also ensures the l-diversity of anonymity data.At the same time,it can guarantee k-anonymity and l-diversity of many published anonymity data talbe in the attacker's reasoning talbe.With the increase of dimension,multi-dimensional data space index whose running time will invrease an the geometric exponential and whose the efficiency of operation time will be very low.On this basis,the paper uses the space-filling curve approach to make multi-dimensional data into one-dimensional data,which would greatly reduce the running time and increase the efficiency of the data index.