High-dimensional mass data release privacy protection method

Abstract

The invention discloses a high-dimensional mass data release privacy protection method, which comprises the following steps of: reducing dimensions of data by introducing an attribute importance optimization PCA (Principal Component Analysis) algorithm so as to reduce time and space; a differential privacy protection strategy meeting individuation is designed in the dimension reduction process, and different protection degrees are designed for different sensitive attributes; a mutual information evaluation mechanism is introduced into a PCA algorithm for the first time, data generated by selecting different principal component numbers in the dimension reduction process is evaluated, and the optimal principal component number is determined.

Description

technical field [0001] The present invention relates to the technical field of information technology security, and more particularly, the present invention relates to a privacy protection method for publishing high-dimensional mass data. Background technique [0002] At present, many data collection agencies need to release the collected raw data (such as medical data, financial data, etc.) to facilitate data analysis and mining, and to generate more effective decision support from the released data. However, a large amount of personal sensitive information is involved in the released raw data, and the direct release of data will lead to serious leakage of personal privacy. Therefore, data publishers need to process private data through special protection technology before releasing the data. [0003] In the prior art, there have been a few research results on high-dimensional data publishing methods, but these methods all have some problems: [0004] The PriView algorith...

AI Technical Summary

Problems solved by technology

However, part of the privacy budget is wasted when constructing the noise covariance matrix, and the processing time of this method cannot meet the actual requirements when dealing with data with large attribute dimensions

[0010] At present, the research on data release mainly focuses on one-dimensional or low-dimensional data. However, these data release methods are not suitable for the release of high-dimensional data, and cannot solve the With the increase of domains, the publishing space formed grows exponentially, encountering the problem of "dimension disaster", introducing large noise, resulting in low availability of published data

Therefore, in the release of high-dimensional data, it is urgent to design a release method that can solve the problem of low data availability caused by the disaster of dimensionality and satisfy data privacy and security while providing data researchers with a large amount of effective information.

Images