Research of Privacy Preservation Method Based on Data Coloring

In the cloud computing environment, outsourcing service mode of data storage causes the security problem, the reliability of data cannot be guaranteed, and the privacy preservation problem has aroused wide concern. In order to solve the problem of inefficiency and high-complexity caused by traditional privacy preservation methods such as data encryption and access control technology, a privacy preservation method based on data coloring is proposed. The data coloring model is established and the coloring mechanism is adopted to deal with the sensitive data of numerical attributes, and the cloud model similarity measurement based on arithmetic average least-approximability is adopted to authenticate the ownership of privacy data. On the premise of high availability of data, the method strengthens the security of the privacy information. Then, the performance, validity and the parameter errors of the algorithm are quantitatively analyzed by the experiments using the UCI dataset. Under the same conditions of privacy preservation requirements, the proposed method can track privacy leakage efficiently and reduce privacy leakage risks. Compared with the k-anonymity approach, the proposed method enhances the computational time efficiency by 18.5%.

In the cloud computing envi- ronment, outsourcing service mode of data storage causes the security problem, the reli- ability of data cannot be guaranteed, and the privacy preservation problem has aroused wide concem. In order to solve the problem of inefficiency and high-complexity caused by traditional privacy preservation methods such as data encryption and access control technol- ogy, a privacy preservation method based on data coloring is proposed. The data coloring model is established and the coloring mecha- nism is adopted to deal with the sensitive data of numerical attributes, and the cloud model similarity measurement based on arithmetic average least-approximability is adopted to authenticate the ownership of privacy data. On the premise of high availability of data, the method strengthens the security of the privacy information. Then, the performance, validity and the parameter errors of the algorithm are quantitatively analyzed by the experiments using the UCI dataset. Under the same condi- tions of privacy preservation requirements, the proposed method can track privacy leakage efficiently and reduce privacy leakage risks. Compared with the k-anonymity approach, the proposed method enhances the computational time efficiency by 18.5%.