Survey: Strategies for Loss-Based Discrete-Time Hazard and Survival Function Estimation

Citations

SCOPUS

1

초록

Survival analysis (time-to-event analysis) addresses a fundamental problem of understanding the relationship between input and the survival outcomes and predicting the survival outcomes. Survival analysis has been widely used in many areas - including economics, finance, engineering, and medicine - and recently has gained substantial interest from the view of non-parametric machine learning and deep learning frameworks. In this paper, we review strategies for non-parametric, especially deep learning based, estimation of the dynamics underlying discrete-time event processes. More specifically, we focus on loss functions that can be used for implementation and then briefly mention specific instantiation of such approaches from related work.

키워드

deep learning; non-parametric; survival analysis; time-to-event analysis
제목
Survey: Strategies for Loss-Based Discrete-Time Hazard and Survival Function Estimation
저자
Lee, Changhee; Lee, C.; Ha, T.; Cho, Sungrae
DOI
10.1109/ICTC55196.2022.9952504
발행일
2022-10
유형
Conference Paper
저널명
International Conference on ICT Convergence
권
2022-October
페이지
844 ~ 846