vdra: Vertical Distributed Regression Analysis

Implements linear, logistic, and Cox regression on vertically partitioned data across several data partners. Data is not shared between data partners or the analysis center and the computations can be considered secure. Three different protocols are implemented. 2-Party: two data partners which communicate directly without an intermediate analysis center; 2T-Party: two data partners communicate indirectly via an analysis center, and KT-Party: two or more data partners plus an analysis center are all allowed to communicate directly. 2-Party and 2^T-Party use a form of secure multiplication as found in Karr, et. al. (2009) "Privacy-Preserving Analysis of Vertically Partitioned Data Using Secure Matrix Products" and Slavkovic et. al. (2007) "Secure Logistic Regression of Horizontally and Vertically Partitioned Distributed Databases" <doi:10.1109/ICDMW.2007.114>. Full details can be found in Samizo (In preparation).

Package details

AuthorThomas Kent [aut, cre], Yuji Samizo [aut]
MaintainerThomas Kent <kentedegrees@gmail.com>
Package repositoryView on CRAN
Installation Install the latest version of this package by entering the following in R:

Try the vdra package in your browser

Any scripts or data that you put into this service are public.

vdra documentation built on Sept. 9, 2021, 9:10 a.m.