Predicting times to event based on vine copula models

In statistics, time-to-event analysis methods traditionally focus on the estimation of hazards. In recent years, machine learning methods have been proposed to directly predict the event times. A method based on vine copula models is proposed to make point and interval predictions for a right-censor...

Full description

Saved in:
Bibliographic Details
Published in:Computational statistics & data analysis 2022-11, Vol.175, p.107546, Article 107546
Main Authors: Pan, Shenyi, Joe, Harry
Format: Article
Language:English
Subjects:
Conditional quantiles
Copula regression
Prediction interval
Survival analysis
Time-to-event analysis
Vine copula
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In statistics, time-to-event analysis methods traditionally focus on the estimation of hazards. In recent years, machine learning methods have been proposed to directly predict the event times. A method based on vine copula models is proposed to make point and interval predictions for a right-censored response variable given mixed discrete-continuous explanatory variables. Extensive experiments on simulated and real datasets show that the proposed vine copula approach provides a decent approximation to other time-to-event analysis models including proportional hazards and Weibull Accelerate Failure Time models. When the proportional hazards or Weibull Accelerate Failure Time assumptions do not hold, predictions based on vine copulas can significantly outperform other models, depending on the shape of the conditional quantile functions. This shows the flexibility of the proposed vine copula approach for general time-to-event datasets.
ISSN:0167-9473
1872-7352
DOI:10.1016/j.csda.2022.107546