Multi-source data fusion privacy protection method for multi-privacy policy combination optimization

Abstract

The invention relates to the field of data release, in particular to a multi-source data fusion privacy protection method for multi-privacy policy combination optimization. A multi-party data fusion framework based on the over-anonymity is provided, and the situation of privacy disclosure of fused data is prevented. Furthermore, the practical significance of data fusion is to provide a more comprehensive data basis for the user, so that extensive knowledge mining can be carried out on the basis. Therefore, a multi-privacy protection strategy combination optimization scheme is designed, and the availability of fused data is improved to the greatest extent while privacy constraints of all parties are met. According to the strategy, multi-source multi-privacy constraint data are fused and mapped into a hypergraph, hyperedges are selected, solved and eliminated on the hypergraph one by one by using a heuristic rule, the hyperedges are eliminated, the privacy constraints are realized one by one, and a data fusion scheme is formulated.

Description

technical field [0001] The invention relates to the field of data release, and specifically relates to a multi-source data fusion privacy protection method optimized by combining multiple privacy policies. Background technique [0002] Multi-source, cross-platform, and cross-domain data applications are the most prominent features of big data. In the era of big data, due to the explosive growth of data in different application fields, a single type of data (such as location data, social data, cookie logs, shopping website flow, etc.) It is difficult to meet people's needs for upper-layer complex application services. For example, Bob needs the App to search for nearby friends who like to play basketball. The realization of this requirement requires the organic integration of location data and social data. Not only do individuals have a need for cross-domain data fusion, but there are also practical needs for cross-domain data fusion between different departments within an e...

AI Technical Summary

Problems solved by technology

However, the risks faced by big data publishing are reflected in the dynamic nature of the publishing process and the characteristics of multi-source cross-platform publishing. This requires preventing attackers from performing correlation analysis on multi-source fused data, thereby destroying the anonymity of data. sex

[0005] In terms of privacy protection of data fusion, H Patel et al. proposed a bottom-up approach to achieve secure fusion of two-party data, but the premise of this model is that there is a trusted third party to fuse all data to form a complete The original data table, and then anonymize the data table, and in most cases, there is no trusted third party, so the method of this document is of little value; Jiang et al. proposed a semi-honest model Realize the DkA security fusion model of two-party data. This algorithm uses the exchangeable encryption strategy to hide the original information in the communication process, and realizes the privacy protection of the data fusion process by constructing a complete anonymity table to judge whether the anonymity threshold k is satisfied. The resource consumption of the method is too large, and it is not suitable for the fusion of large data sets; Clifton et al. developed a multi-party data integration tool for four typical operations of relational data counting, union, intersection, and Cartesian product; Yeom et al. studied The indirect privacy leakage caused by the insufficient generalization ability of the model. Then, Mohammed et al. used the data generalization technology based on the classification tree structure to realize the data privacy protection of all parties in the data integration, but the information loss of the integrated data is high, and the specific information The degree of loss is related to the data set

The above schemes all assume that multiple parties involved in data fusion adopt the same privacy protection strategy. However, in the face of different privacy protection requirements of big data, different platforms may adopt personalized privacy protection according to their own application requirements before big data fusion strategies, existing solutions are difficult to apply

Images