Regularizing the Deepsurv Network Using Projection Loss for Medical Risk Assessment
Regularizing the Deepsurv Network Using Projection Loss for Medical Risk Assessment
复制标题
DOI：
10.1109/access.2022.3142032
复制
发表时间：
2022
期刊：
影响因子：
3.9
通讯作者：
Phawis Thammasorn;S. Schaub;D. Hippe;M. Spraker;J. Peeken;L. Wootton;Paul Kinahan;S. Combs;W. Chaovalitwongse;Matthew Nyflot
中科院分区：
文献类型：
--
作者：
Phawis Thammasorn;S. Schaub;D. Hippe;M. Spraker;J. Peeken;L. Wootton;Paul Kinahan;S. Combs;W. Chaovalitwongse;Matthew Nyflot
最先进的深度生存预测方法扩展了网络参数，以适应输出时间精细离散化的性能。对于数据有限的医疗应用，基于回归的Deepsurv方法更具优势，因为其连续输出设计限制了不必要的网络参数。尽管有实际的优势，但典型的网络缺乏对特征分布的控制，导致网络更容易产生噪声信息，并且偶尔预测性能不佳。我们提出了一种新的投影损失作为正则化目标来改进事件间隔时间Deepsurv模型。损失公式最大化了网络特征和期望危险值之间的多重相关系数的下界。从理论上讲，减少损失还会降低不协调对的可能性上限，并改善C指数的表现。在五个公共医学数据集和两个交叉队列验证任务的实验中，我们观察到正则化的Deepsurv在许多最先进的方法上具有优越的性能和稳健性。
State-of-the-art deep survival prediction approaches expand network parameters to accommodate performance over a fine discretization of output time. For medical applications where data are limited, the regression-based Deepsurv approach is more advantageous because its continuous output design limits unnecessary network parameters. Despite the practical advantage, the typical network lacks control over the feature distribution causing the network to be more prone to noisy information and occasional poor prediction performance. We propose a novel projection loss as a regularizing objective to improve the time-to-event Deepsurv model. The loss formulation maximizes the lower bound of the multiple-correlation coefficient between the network’s features and the desired hazard value. Reducing the loss also theoretically lowers the upper bound on the likelihood of discordant pair and improves C-index performance. We observe superior performances and robustness of regularized Deepsurv over many state-of-the-art approaches in our experiments with five public medical datasets and two cross-cohort validation tasks.