Deep survival algorithm based on nuclear norm. Issue 9 (13th June 2022)

• Record Type: Journal Article
• Authors: Tong, Jianyang; Zhao, Xuejing
• Is Part Of: Journal of statistical computation and simulation. Volume 92:Issue 9(2022)
• Page Start: 1964
• Page End: 1976
• Publication Date: 2022-06-13
• Subjects: Survival analysis -- deep learning -- missing data complement -- nuclear norm -- NN-DeepSurv -- concordance index; 62N01; Mathematical statistics -- Data processing -- Periodicals; Digital computer simulation -- Periodicals; 519.5028505
• DOI: 10.1080/00949655.2021.2015770
• Languages: English
• ISSNs: 0094-9655

Abstract

Abstract : This paper devotes to propose a nuclear-norm-based deep survival algorithm (NN-DeepSurv), to study the regression problem of survival data with right censoring. The nuclear norm method is used to impute missing covariates, and it's combined with DeepSurv algorithm to train the regression model. We compare our algorithm with other state-of-the-art methods: Cox proportional hazards regression model (Coxph), Cox proportional hazards model with lasso regression (Cox-lasso), random survival forests (RSF), DeepSurv, and Xgboost algorithm, on 2 simulated datasets and 6 clinical datasets, to show the superiority of the performance of our algorithm.