Privacy-Preserving Collaborative Data Collection and Analysis with Many Missing Values

Abstract

Privacy-preserving data mining techniques are useful for analyzing various information, such as Internet of Things data and COVID-19-related patient data. However, collecting a large amount of sensitive personal information is a challenging task. In addition, this information may have missing values, which are not considered in the existing methods for collecting personal information while ensuring data privacy. Failure to account for missing values reduces the accuracy of the data analysis. In this paper, we propose a method for privacy-preserving data collection that considers many missing values. The patient data are anonymized and sent to a data collection server. The data collection server creates a generative model and a contingency table suitable for multi-attribute analysis based on expectation-maximization and Gaussian copula methods. Using differential privacy (the de facto standard) as a privacy metric, we conduct experiments on synthetic and real data, including COVID-19-related data. The results are 50--80\% more accurate than those of existing methods that do not consider missing values.