Uncorrelated patient similarity learning

Abstract

Patient similarity learning aims to derive a clinically mean-ingful similarity metric to measure the similarity between a pair of patients according to their historical clinical infor-mation, which could help to predict the clinical outcomes of the patient of interest. However, the patient clinical data are usually complex, and contain much irrelevant and redun-dant information, which makes it dificult to learn the sim-ilarity metric with high accuracy. Although some methods have been proposed to address the complex nature of patient data, they overemphasize sparsity-based relevant feature se-lection and fail to take into consideration the redundant fea-tures that are highly correlated with each other, and this heavily degrades the accuracy of the learned results. To ad-dress the above challenges, we propose a novel uncorrelated patient similarity learning approach, which can not only se-lect the most relevant features for the learning task, but also guarantee that the selected features have low correla-tions with each other. Additionally, to address the scenarios where the patient data are distributed across different sites, we extend the proposed approach and design a distributed mechanism, based on which the similarity metric can be ac-curately learned without directly accessing the raw patient data at each site. The desirable performance of the pro-posed methods are verified through extensive experiments conducted on both real-world and synthetic datasets.