Published on Mon Jul 01 2019

Time-to-Event Prediction with Neural Networks and Cox Regression

Håvard Kvamme, Ørnulf Borgan, Ida Scheel

New methods for time-to-event prediction are proposed by extending the Cox proportional hazards model with neural networks. The proposed loss function is verified to be a good approximation for the Cox partial partial log-likelihood.

0
0
0
Abstract

New methods for time-to-event prediction are proposed by extending the Cox proportional hazards model with neural networks. Building on methodology from nested case-control studies, we propose a loss function that scales well to large data sets, and enables fitting of both proportional and non-proportional extensions of the Cox model. Through simulation studies, the proposed loss function is verified to be a good approximation for the Cox partial log-likelihood. The proposed methodology is compared to existing methodologies on real-world data sets, and is found to be highly competitive, typically yielding the best performance in terms of Brier score and binomial log-likelihood. A python package for the proposed methods is available at https://github.com/havakv/pycox.

Fri Mar 26 2021
Machine Learning
Time-to-event regression using partially monotonic neural networks
0
0
0
Sun Apr 14 2019
Machine Learning
Analysis of overfitting in the regularized Cox model
The Cox proportional hazards model is ubiquitous in the analysis of time-to-event data. However, when the data dimension p is comparable to the sample size $N, maximum likelihood estimates for its regression parameters are biased. This prompted the introduction of the so-called regularized Cox model.
0
0
0
Fri Apr 03 2020
Machine Learning
Neural Conditional Event Time Models
Event time models predict occurrence times of an event of interest based on known features. No distinction is made between a) the probability of event occurrence and b) the predicted time of occurrence. This distinction is critical when predicting medical diagnoses, equipment defects, and social media posts.
0
0
0
Sat Jun 27 2020
Machine Learning
A General Machine Learning Framework for Survival Analysis
Time-to-event data, also known as survival analysis, requires specialized machine learning methods. We present a very general machine learning framework that uses a data augmentation strategy to reduce complex survival tasks to standard Poisson regression tasks. Any algorithm that can optimize a Poisson (log-)likelihood
1
0
1
Tue May 21 2019
Machine Learning
Survival Function Matching for Calibrated Time-to-Event Predictions
The proposed survival function estimator can be used in practice as a means of estimating and comparing conditional survival distributions. Extensive experiments show that the proposed model outperforms existing approaches, trained both with and without adversarial learning.
0
0
0
Sun Jul 26 2020
Machine Learning
DeepHazard: neural network for time-varying risks
DeepHazard is a neural network for time-varying risks. It is tailored for a wide range of continuous hazards forms, with the only restriction of being additive in time. A flexible implementation, allowing different optimization methods, is developed.
0
0
0