Adaptive Differential Privacy Interactive Publishing Model Based on Dynamic Feedback

Data publishing is very meaningful and necessary. However, there are much personal (especially sometimes sensitive) information in the datasets to be published. So, privacy preserving has become a more and more important problem what we must deal with in big data era. Because of the strong mathematics foundation, provable and quantized privacy properties, DP (differential privacy) attracts the most interests and is becoming one of the most prevalent privacy models. This paper, based on differential privacy preserving mechanism, engages in queries restriction problem in interactive privacy data publishing framework. One adaptive differential privacy interactive publishing model based on dynamic feedback model (ADP M-DF) is proposed. Then, its technological process is presented by the flow chart in detail. And, the dynamic feedback scheme is proposed with an iteration algorithm to generate new privacy budget parameter. Finally, some qualities are discussed. Analysis shows that the new model can run well with good practical meanings and provide better user query experience.