arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2302.12749cs.LG
收藏

SurvivalGAN:生成生存分析的至事件时间数据

SurvivalGAN: Generating Time-to-Event Data for Survival Analysis

University of Cambridge(剑桥大学);University of California, Los Angeles(加州大学洛杉矶分校);Alan Turing Institute(艾伦·图灵研究所)

Alexander Norcliffe, Bogdan Cebere, Fergus Imrie, Pietro Lio, Mihaela van der Schaar

摘要:

合成数据正成为一种越来越有前途的技术,成功的应用可以提高隐私、公平和数据民主化。虽然有许多方法可以生成合成表格数据,但对于特定场景来说,这项任务仍然是不平凡的,也是未经探索的。其中一个场景是生存数据。在这里,关键的困难是审查:在某些情况下,我们不知道事件发生的时间,或者甚至不知道是否发生。删失和时间范围的不平衡导致生成模型经历三种特定于生存分析的新失效模式:(1)生成太少的风险成员;(2)生成太多的风险成员;(3)过早删失。我们正式这些故障模式,并提供了三个新的生成指标来量化它们。在此之后,我们提出了SurvivalGAN,这是一种生成模型,它首先通过解决删失和事件范围的不平衡来处理生存数据,其次通过使用专用机制来近似事件/删失时间。我们通过对医学数据集的广泛实验来评估这种方法。SurvivalGAN在生成生存数据方面优于多个基线,特别是解决了通过新指标测量的故障模式,此外还提高了在合成数据上训练的生存模型的下游性能。

英文摘要:

Synthetic data is becoming an increasingly promising technology, and successful applications can improve privacy, fairness, and data democratization. While there are many methods for generating synthetic tabular data, the task remains non-trivial and unexplored for specific scenarios. One such scenario is survival data. Here, the key difficulty is censoring: for some instances, we are not aware of the time of event, or if one even occurred. Imbalances in censoring and time horizons cause generative models to experience three new failure modes specific to survival analysis: (1) generating too few at-risk members; (2) generating too many at-risk members; and (3) censoring too early. We formalize these failure modes and provide three new generative metrics to quantify them. Following this, we propose SurvivalGAN, a generative model that handles survival data firstly by addressing the imbalance in the censoring and event horizons, and secondly by using a dedicated mechanism for approximating time-to-event/censoring. We evaluate this method via extensive experiments on medical datasets. SurvivalGAN outperforms multiple baselines at generating survival data, and in particular addresses the failure modes as measured by the new metrics, in addition to improving downstream performance of survival models trained on the synthetic data.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑