2011/11/30 by Rob Hall, Hall, Rob, Yuval Nardi +4
Computer Science · Social Sciences · #Privacy-Preserving Technologies in Data #Internet Traffic Analysis and Secure E-voting #Survey Methodology and Nonresponse
paper · pdf · doi:10.48550/arxiv.1111.7277
Preserving the privacy of individual databases when carrying out statistical\ncalculations has a long history in statistics and had been the focus of much\nrecent attention in machine learning In this paper, we present a protocol for\ncomputing logistic regression when the data are held by separate parties\nwithout actually combining information sources by exploiting results from the\nliterature on multi-party secure computation. We provide only the final result\nof the calculation compared with other methods that share intermediate values\nand thus present an opportunity for compromise of values in the combined\ndatabase. Our paper has two themes: (1) the development of a secure protocol\nfor computing the logistic parameters, and a demonstration of its performances\nin practice, and (2) and amended protocol that speeds up the computation of the\nlogistic function. We illustrate the nature of the calculations and their\naccuracy using an extract of data from the Current Population Survey divided\nbetween two parties.\n