Privacy-Preserving Multiple Linear Regression of Vertically Partitioned Real Medical Datasets
Hiroaki Kikuchi, Chika Hamanaga, Hideo Yasunaga, Hiroki Matsui, Hideki Hashimoto · 2017
This paper studies the feasibility of privacypreserving data mining in epidemiological study. As for the datamining algorithm, we focus to a linear multiple regression that can be used to identify the most significant factors among many possible variables, such as the history of many diseases. We try to identify the linear model to estimate a length of hospital stay from distributed dataset related to the patient and the disease information. In this paper, we have done experiment using the real medical dataset related to stroke and attempt to apply multiple regression with six predictors of age, sex, the medical scales, e.g., Japan Coma Scale, and the modified Rankin Scale. Our contributions of this paper include (1) to propose a practical privacy-preserving protocols for linear multiple regression with vertically partitioned datasets, and (2) to show the feasibility of the proposed system using the real medical dataset distributed into two parties, the hospital who knows the technical details of diseases during the patients are in the hospital, and the local government who knows the residence even after the patients left hospital. (3) to show the accuracy and the performance of the PPDM system which allows us to estimate the expected processing time with arbitrary number of predictors.