Assumption-Lean Aalen Regression
Event details
| Date | 21.02.2025 |
| Hour | 15:15 › 16:15 |
| Speaker | Niels Richard Hansen, University of Copenhagen |
| Location | |
| Category | Conferences - Seminars |
| Event Language | English |
In this talk I will introduce a new model-free measure of conditional association between a time-to-event and an exposure given additional covariates. The measure is called the Aalen Covariance Measure (ACM). It reduces to the cumulative effect in the partially additive time-to-event regression model, which encompasses the classical additive Aalen regression model. I will describe how to estimate ACM, which is a function of time, leveraging double/debiased machine learning techniques. This allows us to derive the estimator's asymptotic distribution, which can then be used for inference. The target, ACM, and the estimator will be compared to alternative parametric, semiparametric and assumption-lean measures of conditional association and their estimators.
Practical information
- Informed public
- Free
Organizer
- Myrto Limnios
Contact
- Maroussia Schaffner